# dbt Labs

Learn more about the data analytics industry, dbt Cloud and dbt Core, as well as company news and updates.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## What's new in dbt: November 2026

DevFeed: [What's new in dbt: November 2026](<https://devfeed.tech/articles/what-s-new-in-dbt-november-2026-140183.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/whats-new-in-dbt-november-2026>)

Author: Corinne Hallander

Published: 2026-11-17T16:00:00Z

Content type: release

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [Test-driven development](<https://devfeed.tech/topics/tdd.md>), [ChatGPT](<https://devfeed.tech/topics/chatgpt.md>)

Tags: [1-7](<https://devfeed.tech/tags/1-7.md>), [2026](<https://devfeed.tech/tags/2026.md>), [about](<https://devfeed.tech/tags/about.md>), [beta](<https://devfeed.tech/tags/beta.md>), [charts](<https://devfeed.tech/tags/charts.md>), [chatgpt](<https://devfeed.tech/tags/chatgpt.md>), [claude](<https://devfeed.tech/tags/claude.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [desktop](<https://devfeed.tech/tags/desktop.md>), [free](<https://devfeed.tech/tags/free.md>), [local](<https://devfeed.tech/tags/local.md>), [november-2026](<https://devfeed.tech/tags/november-2026.md>), [plan](<https://devfeed.tech/tags/plan.md>), [product](<https://devfeed.tech/tags/product.md>), [public](<https://devfeed.tech/tags/public.md>), [public-beta](<https://devfeed.tech/tags/public-beta.md>), [state](<https://devfeed.tech/tags/state.md>), [tests](<https://devfeed.tech/tags/tests.md>), [unit-tests](<https://devfeed.tech/tags/unit-tests.md>), [what-s-new](<https://devfeed.tech/tags/what-s-new.md>), [wizard](<https://devfeed.tech/tags/wizard.md>)

### AI overview

The November 2026 updates focus on faster testing, lower-cost model reuse, and AI-assisted analytics. In dbt v2, local unit tests (Alpha) translate SQL to DuckDB and run without a warehouse round trip; batch testing (GA) combines `unique` and `not_null` checks, with internal testing reporting 40% lower runtime and 65% fewer warehouse queries. dbt State now has a free monthly allowance of 1,000 or 2,000 DATT depending on hosting and plan; after it is used, models build as they would without State until the next month. Wizard Desktop is in Public Beta across macOS, Windows, and Linux, isolating parallel tasks in git and the warehouse and offering visual DAG diffs for review. Beta dbt Charts integrations let users query boards and warehouses in Claude or ChatGPT through an MCP connector and their own permissions. Other updates include dbt MCP marketplace listings, model query history for Redshift and Databricks, and self-serve PrivateLink setup in Public Beta for specified platforms.

### Source excerpt

Local unit tests, a free dbt State plan, Wizard Desktop in Public Beta, and dbt Charts in Claude and ChatGPT.

## Making the economic case for warehouse-first AI context

DevFeed: [Making the economic case for warehouse-first AI context](<https://devfeed.tech/articles/making-the-economic-case-for-warehouse-first-ai-context-129183.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/economic-case-warehouse-first-ai-context>)

Author: Daniel Poppy

Published: 2026-11-11T13:00:00Z

Content type: article

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [LLM optimization](<https://devfeed.tech/topics/llm-optimization.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [Finance](<https://devfeed.tech/topics/finance.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [Claude](<https://devfeed.tech/topics/claude.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [2025](<https://devfeed.tech/tags/2025.md>), [2026](<https://devfeed.tech/tags/2026.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-context](<https://devfeed.tech/tags/ai-context.md>), [but](<https://devfeed.tech/tags/but.md>), [can](<https://devfeed.tech/tags/can.md>), [case](<https://devfeed.tech/tags/case.md>), [context](<https://devfeed.tech/tags/context.md>), [costs](<https://devfeed.tech/tags/costs.md>), [first](<https://devfeed.tech/tags/first.md>), [here](<https://devfeed.tech/tags/here.md>), [heres](<https://devfeed.tech/tags/heres.md>), [insights](<https://devfeed.tech/tags/insights.md>), [making](<https://devfeed.tech/tags/making.md>), [take](<https://devfeed.tech/tags/take.md>), [token](<https://devfeed.tech/tags/token.md>), [token-costs](<https://devfeed.tech/tags/token-costs.md>), [warehouse](<https://devfeed.tech/tags/warehouse.md>)

### AI overview

The article argues that storing prepared AI context in a data warehouse can reduce recurring agent token costs while projects take time to deliver ROI. Its team reports cutting token costs by more than 50% after replacing repeated retrieval across 20+ MCP servers with a governed warehouse layer exposed through dbt MCP. It identifies three sources of MCP overhead: loading tool definitions on calls, passing large raw payloads through the model, and repeatedly fetching the same context. The proposed approach processes data once, stores reusable context, and sends only relevant passages--for example, chunking and classifying transcripts while retaining full versions in the warehouse. The article cites an Anthropic example reducing upfront token use by 85% and another workflow falling from 150,000 to 2,000 tokens; these are reported examples, not a guarantee for every deployment. It recommends baselining token and connection costs, finding overlapping MCP sources, starting with expensive context-rich data, processing it into warehouse chunks, and measuring costs before and after.

### Source excerpt

AI agents can take years to pay off, but token bills arrive every month. Here's how we dramatically reduced token costs.

## A letter from our founders

DevFeed: [A letter from our founders](<https://devfeed.tech/articles/a-letter-from-our-founders-129184.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/founders-letter>)

Author: George Fraser, Tristan Handy, Taylor Brown

Published: 2026-11-11T00:00:00Z

Content type: opinion

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>), [AI for customer service automation](<https://devfeed.tech/topics/ai-for-customer-service-automation.md>), [Tool](<https://devfeed.tech/topics/tool.md>), [ai-governance](<https://devfeed.tech/topics/ai-governance.md>), [Digital Public Good](<https://devfeed.tech/topics/digital-public-goods.md>), [digital sovereignty](<https://devfeed.tech/topics/digital-sovereignty.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [apis](<https://devfeed.tech/tags/apis.md>), [automate](<https://devfeed.tech/tags/automate.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [autonomous-agents](<https://devfeed.tech/tags/autonomous-agents.md>), [company](<https://devfeed.tech/tags/company.md>), [customer](<https://devfeed.tech/tags/customer.md>), [customer-support](<https://devfeed.tech/tags/customer-support.md>), [data](<https://devfeed.tech/tags/data.md>), [data-infrastructure](<https://devfeed.tech/tags/data-infrastructure.md>), [founders](<https://devfeed.tech/tags/founders.md>), [future](<https://devfeed.tech/tags/future.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [intelligence](<https://devfeed.tech/tags/intelligence.md>), [open](<https://devfeed.tech/tags/open.md>), [open-data](<https://devfeed.tech/tags/open-data.md>)

### AI overview

The founders argue that useful agents combine intelligence with tools and data, and that their effectiveness depends on access to business context. They distinguish data-heavy analysis, which may need structured data plus OKRs, notes, and initiatives, from tasks such as contract review that require a more limited set of materials. Agents could expand access to analysis by being available on demand and able to investigate for longer, but the authors say this requires infrastructure such as evaluations and shared memory. They favor connecting agents to operational systems and the definitions and checks that explain the data, rather than relying only on documents and messages. The letter also describes a shift from AI augmenting workers to agents owning processes end to end, and names four responsibilities for agent engineers: process and systems engineering, evaluation and operations, and change management. It advocates open standards, portability, data ownership, and governance to avoid dependence on one vendor, while emphasizing continued investment in human practitioners.

### Source excerpt

Our founders on the age of intelligence: agents, open data infrastructure, and the humans building the agentic future.

## How organizations progress from trusted data to trusted AI and agents

DevFeed: [How organizations progress from trusted data to trusted AI and agents](<https://devfeed.tech/articles/how-organizations-progress-from-trusted-data-to-trusted-ai-and-agents-126364.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/ai-maturity-trusted-data-to-trusted-ai-agents>)

Author: Daniel Poppy

Published: 2026-11-10T13:00:00Z

Content type: article

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>), [Data Management](<https://devfeed.tech/topics/data-management.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [Data Space](<https://devfeed.tech/topics/data-space.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [accessible](<https://devfeed.tech/tags/accessible.md>), [accountability](<https://devfeed.tech/tags/accountability.md>), [act](<https://devfeed.tech/tags/act.md>), [advanced](<https://devfeed.tech/tags/advanced.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-maturity](<https://devfeed.tech/tags/ai-maturity.md>), [autonomy](<https://devfeed.tech/tags/autonomy.md>), [build](<https://devfeed.tech/tags/build.md>), [context](<https://devfeed.tech/tags/context.md>), [control](<https://devfeed.tech/tags/control.md>), [data](<https://devfeed.tech/tags/data.md>), [governance](<https://devfeed.tech/tags/governance.md>), [insights](<https://devfeed.tech/tags/insights.md>), [maturity](<https://devfeed.tech/tags/maturity.md>), [organizations](<https://devfeed.tech/tags/organizations.md>), [progress](<https://devfeed.tech/tags/progress.md>), [trusted-ai](<https://devfeed.tech/tags/trusted-ai.md>), [trusted-data](<https://devfeed.tech/tags/trusted-data.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

AI maturity is measured by whether an organization's data and controls can support the work assigned to AI, not simply by using more advanced models or agents. The excerpt describes building from consistent, tested, documented data toward shared governance, clear ownership, business definitions, machine-readable context, lineage, and oversight that work across teams and systems. These foundations help AI interpret data consistently and make greater autonomy possible while preserving accountability and control; when they are weak, more human oversight is needed. The excerpt reports that 41% of organizations use agentic AI in production while 15% feel fully prepared with their data foundation, and that 86% of data leaders consider extensibility and interoperability important or critical. It presents interoperable infrastructure as a way to preserve context and governance as tools evolve. Autonomy is framed as an outcome of maturity, not the starting point.

### Source excerpt

Autonomy is an outcome of AI maturity. Build the trusted data, governance, and context that let AI and agents act with control.

## How to assess your organization's AI readiness

DevFeed: [How to assess your organization's AI readiness](<https://devfeed.tech/articles/how-to-assess-your-organization-s-ai-readiness-121083.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/how-to-assess-your-ai-readiness>)

Author: Daniel Poppy

Published: 2026-11-05T21:49:00Z

Content type: article

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>), [data-architecture](<https://devfeed.tech/topics/data-architecture.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [2026](<https://devfeed.tech/tags/2026.md>), [3](<https://devfeed.tech/tags/3.md>), [accessible](<https://devfeed.tech/tags/accessible.md>), [accountability](<https://devfeed.tech/tags/accountability.md>), [act](<https://devfeed.tech/tags/act.md>), [adoption](<https://devfeed.tech/tags/adoption.md>), [advanced](<https://devfeed.tech/tags/advanced.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-adoption](<https://devfeed.tech/tags/ai-adoption.md>), [ai-readiness](<https://devfeed.tech/tags/ai-readiness.md>), [data](<https://devfeed.tech/tags/data.md>), [dimensions](<https://devfeed.tech/tags/dimensions.md>), [foundation](<https://devfeed.tech/tags/foundation.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [insights](<https://devfeed.tech/tags/insights.md>), [organization](<https://devfeed.tech/tags/organization.md>), [takes](<https://devfeed.tech/tags/takes.md>)

### AI overview

AI adoption alone does not show whether an organization can scale AI reliably; readiness depends on the data and organizational capabilities supporting it. The Fivetran + dbt Labs model assesses six dimensions: data quality, governance, ownership, cost and operational efficiency, context, and data architecture and interoperability. The excerpt recommends assessing each separately to expose uneven maturity, then bringing together stakeholders across data, technology, governance, and AI to compare perspectives and prioritize capabilities based on the organization's next use cases. Readiness needs reassessment as use cases, scale, and autonomy change. The text reports that 42% of organizations cited data quality and lineage issues as a barrier to agentic AI goals, while 86% of data leaders considered platform extensibility and interoperability important or critical. These findings are attributed to Fivetran's 2026 survey of 400 data professionals.

### Source excerpt

AI adoption moves fast. Readiness takes longer. Assess 6 dimensions to see where your data foundation limits AI.

## How to put generative AI directly into a dbt pipeline with plain SQL

DevFeed: [How to put generative AI directly into a dbt pipeline with plain SQL](<https://devfeed.tech/articles/how-to-put-generative-ai-directly-into-a-dbt-pipeline-with-plain-sql-102638.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/bigquery-generative-ai-dbt-sql-functions>)

Author: Daniel Poppy

Published: 2026-10-27T19:06:00Z

Content type: article

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>), [Generative AI](<https://devfeed.tech/topics/generative-ai.md>), [BigQuery](<https://devfeed.tech/topics/bigquery.md>), [SQL](<https://devfeed.tech/topics/sql.md>), [Analytics Pipeline](<https://devfeed.tech/topics/analytics-pipeline.md>), [JOIN](<https://devfeed.tech/topics/join.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-experiments](<https://devfeed.tech/tags/ai-experiments.md>), [ai-logic](<https://devfeed.tech/tags/ai-logic.md>), [api](<https://devfeed.tech/tags/api.md>), [applied-ai](<https://devfeed.tech/tags/applied-ai.md>), [bigquery](<https://devfeed.tech/tags/bigquery.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [functions](<https://devfeed.tech/tags/functions.md>), [generative](<https://devfeed.tech/tags/generative.md>), [generative-ai](<https://devfeed.tech/tags/generative-ai.md>), [google](<https://devfeed.tech/tags/google.md>), [google-cloud](<https://devfeed.tech/tags/google-cloud.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [json](<https://devfeed.tech/tags/json.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [learn](<https://devfeed.tech/tags/learn.md>), [llm](<https://devfeed.tech/tags/llm.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [sql](<https://devfeed.tech/tags/sql.md>)

### AI overview

BigQuery's AI SQL functions let dbt models apply LLM-based analysis directly to table rows, without a separate Python service. The excerpt describes AI.GENERATE for custom prompts and typed JSON via output_schema, plus managed functions such as AI.IF, AI.CLASSIFY, AI.SCORE, and AI.AGG for common analytical tasks; these are generally cheaper than custom prompts. Optimized AI.IF and AI.CLASSIFY use a sample-labeled distilled model to process the remaining data. For production, the speaker recommends incremental models with AI logic inside is_incremental() to avoid resending all rows on full refresh, token estimates and daily caps to manage spend, and max_error_ratio to stop queries when failures exceed a threshold. Pairing output_schema with dbt tests such as not_null and accepted_values can catch invalid results before they reach marts. The described demo scores review sentiment and summarizes it by product. This overview reflects the supplied excerpt.

### Source excerpt

Google Cloud on wiring BigQuery's AI SQL functions into dbt pipelines, from sentiment scoring to LLM-generated product briefings.

## How Sigma runs a 2,883-model semantic layer on dbt and Dagster

DevFeed: [How Sigma runs a 2,883-model semantic layer on dbt and Dagster](<https://devfeed.tech/articles/how-sigma-runs-a-2-883-model-semantic-layer-on-dbt-and-dagster-102652.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/sigma-dbt-dagster-semantic-layer>)

Author: Daniel Poppy

Published: 2026-10-27T18:51:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [version-control](<https://devfeed.tech/topics/version-control.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-applications](<https://devfeed.tech/tags/ai-applications.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [anthropic](<https://devfeed.tech/tags/anthropic.md>), [api](<https://devfeed.tech/tags/api.md>), [applications](<https://devfeed.tech/tags/applications.md>), [building](<https://devfeed.tech/tags/building.md>), [business](<https://devfeed.tech/tags/business.md>), [column](<https://devfeed.tech/tags/column.md>), [company](<https://devfeed.tech/tags/company.md>), [dagster](<https://devfeed.tech/tags/dagster.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [learn](<https://devfeed.tech/tags/learn.md>), [model](<https://devfeed.tech/tags/model.md>), [models](<https://devfeed.tech/tags/models.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [semantic-layer](<https://devfeed.tech/tags/semantic-layer.md>), [sigma](<https://devfeed.tech/tags/sigma.md>), [version](<https://devfeed.tech/tags/version.md>)

### AI overview

Sigma manages its semantic layer as version-controlled SQL and Jinja in a dbt project, using explicit column-level exposure controls and models organized by business function. Three dbt packages compile definitions into Sigma data models, Snowflake semantic views, and Cortex agents. Dagster coordinates 63 jobs across 2,883 dbt models and six regions: jobs download the dbt manifest from S3, run a state-aware dbt build, sync Sigma data models through Sigma's API, and upload the manifest for the next run. Sigma frames the layer as a way to apply model-level AI guardrails, keep metric definitions consistent across AI surfaces, and provide agents with context, synonyms, and verified queries. Senick also previewed Open Semantic Interchange ("Ossie"), intended to enable conversion among Snowflake semantic views, dbt metrics, and Sigma data models.

### Source excerpt

Sigma shows how it manages its semantic layer as version-controlled code across thousands of dbt models.

## ClickHouse + dbt: Building a GitOps workflow at Twilio

DevFeed: [ClickHouse + dbt: Building a GitOps workflow at Twilio](<https://devfeed.tech/articles/clickhouse-dbt-building-a-gitops-workflow-at-twilio-102655.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/twilio-clickhouse-dbt-gitops-workflow>)

Author: Daniel Poppy

Published: 2026-10-27T18:47:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [GitOps](<https://devfeed.tech/topics/gitops.md>), [data-engineering](<https://devfeed.tech/topics/data-engineering.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Apache Kafka](<https://devfeed.tech/topics/apache-kafka.md>), [GitHub](<https://devfeed.tech/topics/github.md>), [elasticsearch](<https://devfeed.tech/topics/elasticsearch.md>), [Service fabric](<https://devfeed.tech/topics/service-fabric.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [clickhouse](<https://devfeed.tech/tags/clickhouse.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [elasticsearch](<https://devfeed.tech/tags/elasticsearch.md>), [gitops](<https://devfeed.tech/tags/gitops.md>), [kafka](<https://devfeed.tech/tags/kafka.md>), [learn](<https://devfeed.tech/tags/learn.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [spark](<https://devfeed.tech/tags/spark.md>), [twilio](<https://devfeed.tech/tags/twilio.md>)

### AI overview

Twilio consolidated a Voice Insights pipeline that had used Kafka for ingestion, Spark Streaming for processing, and Elasticsearch for serving into ClickHouse, using materialized views for streaming transformations. Its workflow progressed from shared DDL documents to Git version control and a migration tool, then to dbt with the open-source dbt-clickhouse adapter for transformations, tests, and lineage. Twilio split its monolithic dbt project into product-area projects with shared logic, orchestrated by Argo Workflows. Engineers use real ClickHouse containers for local and CI integration tests and can deploy branch-specific previews against a slice of production data. The change replaced manual production edits with audited, version-controlled schemas and enabled consistent redeployment across regions. The session also reported that the dbt v1 adapter remained supported, v2 was in public beta and nearing feature parity, and ClickHouse support in the dbt platform was in private beta.

### Source excerpt

Twilio replaced Kafka, Spark, and Elasticsearch with one ClickHouse pipeline governed by dbt and GitOps.

## How KOHO built a Claude-powered self-serve analytics tool on years of dbt governance

DevFeed: [How KOHO built a Claude-powered self-serve analytics tool on years of dbt governance](<https://devfeed.tech/articles/how-koho-built-a-claude-powered-self-serve-analytics-tool-on-years-of-dbt-governance-102640.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/claude-self-serve-analytics-dbt-governance>)

Author: Daniel Poppy

Published: 2026-10-27T18:44:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [analytics stack](<https://devfeed.tech/topics/analytics-stack.md>), [metric-standardization](<https://devfeed.tech/topics/metric-standardization.md>), [Claude](<https://devfeed.tech/topics/claude.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [Claude Code](<https://devfeed.tech/topics/claude-code.md>), [DataOps](<https://devfeed.tech/topics/dataops.md>), [Amazon Redshift](<https://devfeed.tech/topics/amazon-redshift.md>), [Analytics Pipeline](<https://devfeed.tech/topics/analytics-pipeline.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [bedrock](<https://devfeed.tech/tags/bedrock.md>), [claude](<https://devfeed.tech/tags/claude.md>), [claude-code](<https://devfeed.tech/tags/claude-code.md>), [context](<https://devfeed.tech/tags/context.md>), [context-libraries](<https://devfeed.tech/tags/context-libraries.md>), [culture](<https://devfeed.tech/tags/culture.md>), [data](<https://devfeed.tech/tags/data.md>), [data-team](<https://devfeed.tech/tags/data-team.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [fintech](<https://devfeed.tech/tags/fintech.md>), [foundations](<https://devfeed.tech/tags/foundations.md>), [governance](<https://devfeed.tech/tags/governance.md>), [learn](<https://devfeed.tech/tags/learn.md>), [self-serve-analytics](<https://devfeed.tech/tags/self-serve-analytics.md>)

### AI overview

KOHO's KoDa lets employees ask analytics questions in Claude and receive answers from certified dbt marts in Redshift, using Bedrock and Claude. KOHO says the tool depended on years of data culture and governance: its dbt foundation has about 1,400 tested and documented models, with PII removed. KoDa is limited to analytics, uses no raw data, is read-only, and applies shared company-wide metric definitions. It connects a markdown-based "KoDa skill" and pre- and post-hook MCP calls through Claude Code to a Redshift MCP server. The team's substantial work went into evaluation: a library of about 375 real analyst questions supports offline quality checks and weekly human review of logged questions. KOHO still reviews self-serve results intended for the board; the speakers identify company-wide data literacy and knowing when to trust an answer as ongoing challenges.

### Source excerpt

KOHO explains why AI self-serve analytics needs a strong data culture and a governed dbt foundation first.

## Why an AI agent's revenue query was $3.5 million wrong, and how AWS fixed it with dbt Semantic Layer

DevFeed: [Why an AI agent's revenue query was $3.5 million wrong, and how AWS fixed it with dbt Semantic Layer](<https://devfeed.tech/articles/why-an-ai-agent-s-revenue-query-was-3-5-million-wrong-and-how-aws-fixed-it-with-dbt-semantic-layer-102636.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/aws-dbt-semantic-layer-ai-agent-revenue-accuracy>)

Author: Daniel Poppy

Published: 2026-10-27T18:41:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>), [Amazon S3 Tables](<https://devfeed.tech/topics/amazon-s3-tables.md>), [Amazon Web Services](<https://devfeed.tech/topics/aws.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-revenue](<https://devfeed.tech/tags/ai-revenue.md>), [aws](<https://devfeed.tech/tags/aws.md>), [business](<https://devfeed.tech/tags/business.md>), [business-context](<https://devfeed.tech/tags/business-context.md>), [context](<https://devfeed.tech/tags/context.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [dbt-semantic-layer](<https://devfeed.tech/tags/dbt-semantic-layer.md>), [errors](<https://devfeed.tech/tags/errors.md>), [layer](<https://devfeed.tech/tags/layer.md>), [learn](<https://devfeed.tech/tags/learn.md>), [prevent](<https://devfeed.tech/tags/prevent.md>), [query](<https://devfeed.tech/tags/query.md>), [revenue](<https://devfeed.tech/tags/revenue.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [semantic-layer](<https://devfeed.tech/tags/semantic-layer.md>)

### AI overview

A raw SQL query overstated revenue by 30.8%: it summed pre-discount, pre-refund amounts and included refunds incorrectly, cancelled and pending orders, and internal test accounts. The AWS presenters' proposed fix separates metric governance from query execution. The dbt Semantic Layer resolves questions against certified metrics and dimensions, while MetricFlow compiles the SQL; the Redshift MCP Server executes it as read-only, single-statement queries, with IAM and database permissions adding safeguards. Responses include the metric, SQL, source model, dbt test results, and Redshift query ID. The presenters said the same metric definitions worked across Redshift, a Redshift and Iceberg query, and dbt's Athena and DuckDB adapters without Redshift. Developers could review SQL before execution, while business users received cited plain-language answers.

### Source excerpt

A plausible AI revenue query was $3.5 million wrong. AWS explains how business context helps prevent those errors.

## How DoorDash scaled merchant-facing analytics to thousands of users in weeks, on dbt and ThoughtSpot

DevFeed: [How DoorDash scaled merchant-facing analytics to thousands of users in weeks, on dbt and ThoughtSpot](<https://devfeed.tech/articles/how-doordash-scaled-merchant-facing-analytics-to-thousands-of-users-in-weeks-on-dbt-and-thoughtspot-102646.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/doordash-merchant-analytics-thoughtspot-self-serve>)

Author: Daniel Poppy

Published: 2026-10-27T18:38:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [metric-standardization](<https://devfeed.tech/topics/metric-standardization.md>), [analytics stack](<https://devfeed.tech/topics/analytics-stack.md>)

Tags: [analytics](<https://devfeed.tech/tags/analytics.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [doordash](<https://devfeed.tech/tags/doordash.md>), [layer](<https://devfeed.tech/tags/layer.md>), [learn](<https://devfeed.tech/tags/learn.md>), [pilot](<https://devfeed.tech/tags/pilot.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [semantic-layer](<https://devfeed.tech/tags/semantic-layer.md>), [thoughtspot](<https://devfeed.tech/tags/thoughtspot.md>), [users](<https://devfeed.tech/tags/users.md>)

### AI overview

DoorDash scaled an embedded merchant analytics use case to thousands of users within a few weeks by separating its governed semantic layer from the tools that consume it. dbt defines shared metrics over data in Iceberg tables and Snowflake; ThoughtSpot connects to those definitions as one of several consumption layers, avoiding a separate set of metrics in each tool. Internally, DoorDash uses the setup for finance and revenue reporting, pairing governed metrics with ThoughtSpot compliance and audit reporting. For merchants, ThoughtSpot is fully white-labeled and role-level access controls restrict each merchant to its own data. Harsha Reddy advises treating semantic models as version-controlled, reviewed code and making the governed route as easy as ad hoc querying. DoorDash also grounds a global agent in the same dbt layer; the article says it covers roughly 70-80% of data domains.

### Source excerpt

DoorDash used a dbt-defined semantic layer to scale analytics from pilot to thousands of merchant users in weeks.

## How Ecolab used an AI migration agent to shrink 84,000 lines of legacy SQL by 6x

DevFeed: [How Ecolab used an AI migration agent to shrink 84,000 lines of legacy SQL by 6x](<https://devfeed.tech/articles/how-ecolab-used-an-ai-migration-agent-to-shrink-84-000-lines-of-legacy-sql-by-6x-102647.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/ecolab-ai-agent-legacy-sql-migration>)

Author: Daniel Poppy

Published: 2026-10-27T18:20:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [data migrations](<https://devfeed.tech/topics/data-migrations.md>), [Legacy Modernization](<https://devfeed.tech/topics/legacy-modernization.md>), [legacy systems](<https://devfeed.tech/topics/legacy-systems.md>), [SQL](<https://devfeed.tech/topics/sql.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Test-driven development](<https://devfeed.tech/topics/tdd.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agile](<https://devfeed.tech/tags/agile.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-migration](<https://devfeed.tech/tags/ai-migration.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [automated-testing](<https://devfeed.tech/tags/automated-testing.md>), [compliance](<https://devfeed.tech/tags/compliance.md>), [context](<https://devfeed.tech/tags/context.md>), [data-quality](<https://devfeed.tech/tags/data-quality.md>), [data-quality-issues](<https://devfeed.tech/tags/data-quality-issues.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [institutional-knowledge](<https://devfeed.tech/tags/institutional-knowledge.md>), [javascript](<https://devfeed.tech/tags/javascript.md>), [learn](<https://devfeed.tech/tags/learn.md>), [legacy](<https://devfeed.tech/tags/legacy.md>), [migration](<https://devfeed.tech/tags/migration.md>), [models](<https://devfeed.tech/tags/models.md>), [rebuild](<https://devfeed.tech/tags/rebuild.md>), [snowflake](<https://devfeed.tech/tags/snowflake.md>), [sql](<https://devfeed.tech/tags/sql.md>)

### AI overview

Ecolab and Datafold used a migration agent to reorganize a complex Snowflake system into more than 40 dbt projects with clearer ownership. The agent mapped lineage and code, incorporated context from wikis and meeting recordings, translated the logic, and checked each model against production data with automated diffs; unexplained differences were sent back for resolution. Across the migration, Ecolab reports roughly six times less code and four times fewer data connections. One flagship dimension fell from 409 models to 170, with about 70,000 lines of code and 1,000 connections removed. Visual row-level diffs helped data owners review changes, and roughly 90 tables were approved in the first four weeks. The team found that migrations benefit from iterative review, while cosmetic changes such as column renaming are costly for the agent and are better batched until data correctness is verified. Cutover remains out of scope until all dimensions and facts are reviewed and approved.

### Source excerpt

Ecolab and Datafold used an AI migration agent to rebuild a decade of tangled Snowflake logic as clean, documented dbt models.

## How Checkr cut an hours-long metrics investigation to minutes by engineering AI context on dbt and Omni

DevFeed: [How Checkr cut an hours-long metrics investigation to minutes by engineering AI context on dbt and Omni](<https://devfeed.tech/articles/how-checkr-cut-an-hours-long-metrics-investigation-to-minutes-by-engineering-ai-context-on-dbt-and-omni-102639.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/checkr-ai-context-metrics-investigation>)

Author: Daniel Poppy

Published: 2026-10-27T18:16:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [metric-standardization](<https://devfeed.tech/topics/metric-standardization.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [data-engineering](<https://devfeed.tech/topics/data-engineering.md>), [context-architecture](<https://devfeed.tech/topics/context-architecture.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [advice](<https://devfeed.tech/tags/advice.md>), [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-assistant](<https://devfeed.tech/tags/ai-assistant.md>), [ai-context](<https://devfeed.tech/tags/ai-context.md>), [ai-readiness](<https://devfeed.tech/tags/ai-readiness.md>), [ai-ready-data](<https://devfeed.tech/tags/ai-ready-data.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [checkr](<https://devfeed.tech/tags/checkr.md>), [context](<https://devfeed.tech/tags/context.md>), [context-layer](<https://devfeed.tech/tags/context-layer.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [investigation](<https://devfeed.tech/tags/investigation.md>), [learn](<https://devfeed.tech/tags/learn.md>), [metrics](<https://devfeed.tech/tags/metrics.md>), [structured](<https://devfeed.tech/tags/structured.md>), [support](<https://devfeed.tech/tags/support.md>)

### AI overview

Checkr found that connecting an AI agent directly to a Snowflake and dbt environment with more than 6,000 models produced inconsistent answers because the agent lacked business context. Its approach adds three layers: warehouse data, a semantic layer defining valid joins and metrics, and AI context in Omni--synonyms, sample values, field descriptions, behavioral guidance, and sample queries. Checkr builds that context from existing company material and improves it through evals: failures are clustered, relevant knowledge is retrieved to draft fixes, and failing evals are rerun before a subject-matter expert reviews a merge request. In one FRT investigation, the agent identified international queues as the source of underperformance, a task an operations lead said took minutes instead of hours. The excerpt also reports that broad, unfocused context additions can slow answers and roughly double token costs without improving accuracy, so the speakers recommend addressing measured failures and pruning unused fields.

### Source excerpt

Checkr gave its AI agent a structured dbt context layer, cutting a support-metric investigation from hours to minutes.

## How Guitar Center rebuilt its data platform on trust

DevFeed: [How Guitar Center rebuilt its data platform on trust](<https://devfeed.tech/articles/how-guitar-center-rebuilt-its-data-platform-on-trust-102648.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/guitar-center-data-pltform-trust-costs>)

Author: Daniel Poppy

Published: 2026-10-27T18:13:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [trust](<https://devfeed.tech/topics/trust.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>), [database-optimization](<https://devfeed.tech/topics/database-optimization.md>)

Tags: [access-control](<https://devfeed.tech/tags/access-control.md>), [ai](<https://devfeed.tech/tags/ai.md>), [airflow](<https://devfeed.tech/tags/airflow.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [analytics-stack](<https://devfeed.tech/tags/analytics-stack.md>), [ci-cd](<https://devfeed.tech/tags/ci-cd.md>), [containerized](<https://devfeed.tech/tags/containerized.md>), [costs](<https://devfeed.tech/tags/costs.md>), [data-infrastructure](<https://devfeed.tech/tags/data-infrastructure.md>), [data-platform](<https://devfeed.tech/tags/data-platform.md>), [data-warehouse](<https://devfeed.tech/tags/data-warehouse.md>), [days](<https://devfeed.tech/tags/days.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [delivery](<https://devfeed.tech/tags/delivery.md>), [design](<https://devfeed.tech/tags/design.md>), [learn](<https://devfeed.tech/tags/learn.md>), [snowflake](<https://devfeed.tech/tags/snowflake.md>), [speed](<https://devfeed.tech/tags/speed.md>), [stack](<https://devfeed.tech/tags/stack.md>), [trust](<https://devfeed.tech/tags/trust.md>)

### AI overview

Guitar Center rebuilt its analytics platform after a failed Netezza-to-Redshift migration left over 2,000 lines of stored procedures, no CI/CD, and little observability. Using managed dbt Core with DataCoves, the team focused on cost control, delivery flow, and data confidence. It ingests data once, scopes builds with selectors.yml, and separates work across small Snowflake warehouses; optimizing the hourly sales object cut Snowflake run rate by 40%, while reducing Airflow minutes cut consumption by 50%. Identical development and production configurations, a SourceWrapper pattern, and SQL feature flags helped shorten pipelines that took months to ship to days; the replacement platform was working in under a year. Doll emphasized that reliable pipelines can still be wrong, so legacy bug fixes and stakeholder validation accompanied the rebuild. He said AI amplifies the quality of an organization's data foundation rather than fixing unreliable data. The team also designed masking and role-based access in from the start, and Doll advised using consultants without outsourcing platform ownership.

### Source excerpt

Guitar Center rebuilt its data stack around control, speed, and trust, cutting Snowflake spend 40% along the way.

## How Sweetwater catches broken dbt pipelines before anyone notices, with Datadog

DevFeed: [How Sweetwater catches broken dbt pipelines before anyone notices, with Datadog](<https://devfeed.tech/articles/how-sweetwater-catches-broken-dbt-pipelines-before-anyone-notices-with-datadog-102654.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/sweetwater-dbt-pipeline-monitoring-datadog>)

Author: Daniel Poppy

Published: 2026-10-27T18:08:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [data observability](<https://devfeed.tech/topics/data-observability.md>), [datadog](<https://devfeed.tech/topics/datadog.md>), [Data pipelines](<https://devfeed.tech/topics/data-pipelines.md>), [Apache Kafka](<https://devfeed.tech/topics/apache-kafka.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [alert](<https://devfeed.tech/tags/alert.md>), [alerting](<https://devfeed.tech/tags/alerting.md>), [ci-cd](<https://devfeed.tech/tags/ci-cd.md>), [cli](<https://devfeed.tech/tags/cli.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [dag](<https://devfeed.tech/tags/dag.md>), [dashboard](<https://devfeed.tech/tags/dashboard.md>), [dashboards](<https://devfeed.tech/tags/dashboards.md>), [data-infrastructure](<https://devfeed.tech/tags/data-infrastructure.md>), [data-observability](<https://devfeed.tech/tags/data-observability.md>), [data-platform](<https://devfeed.tech/tags/data-platform.md>), [data-warehouse](<https://devfeed.tech/tags/data-warehouse.md>), [datadog](<https://devfeed.tech/tags/datadog.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [learn](<https://devfeed.tech/tags/learn.md>), [missing-data](<https://devfeed.tech/tags/missing-data.md>), [pipelines](<https://devfeed.tech/tags/pipelines.md>)

### AI overview

Sweetwater extended Datadog monitoring into its dbt pipelines as its data platform grew, building on the tool already used by IT to monitor source systems. In a recurring sales dashboard issue, data appeared fresh even though invoices lacked sales channels. Kafka insert dashboards helped the team trace a transaction spike to a websql source and an incorrectly scheduled legacy script; they rescheduled it and began considering a move to the cloud warehouse. Datadog's Google Cloud integrations and prebuilt metrics also helped the team manage timeout risk as DAGs and batch demands grew. Its setup includes schema- and database-wide monitors, table-adjusted row count checks, freshness checks that allow planned nightly downtime, team-specific alert and recovery routing, and machine-learning thresholds with manual annotations. The team is expanding into lineage and query monitoring, terminal and AI-assistant access through PupCLI and the Datadog MCP server, and CI/CD impact analysis and data profiling for changed models.

### Source excerpt

Sweetwater uses Datadog to catch broken dbt pipelines before dashboard users notice missing data.

## How LangChain grew self-serve analytics questions 40x by giving its AI agent layered context

DevFeed: [How LangChain grew self-serve analytics questions 40x by giving its AI agent layered context](<https://devfeed.tech/articles/how-langchain-grew-self-serve-analytics-questions-40x-by-giving-its-ai-agent-layered-context-102650.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/langchain-self-serve-analytics-ai-agent-context>)

Author: Daniel Poppy

Published: 2026-10-27T18:07:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [context](<https://devfeed.tech/topics/context.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [Documentation](<https://devfeed.tech/topics/documentation.md>), [BigQuery](<https://devfeed.tech/topics/bigquery.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [about](<https://devfeed.tech/tags/about.md>), [account](<https://devfeed.tech/tags/account.md>), [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-analytics](<https://devfeed.tech/tags/ai-analytics.md>), [also](<https://devfeed.tech/tags/also.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [bigquery](<https://devfeed.tech/tags/bigquery.md>), [building](<https://devfeed.tech/tags/building.md>), [built-in](<https://devfeed.tech/tags/built-in.md>), [business](<https://devfeed.tech/tags/business.md>), [context](<https://devfeed.tech/tags/context.md>), [giving](<https://devfeed.tech/tags/giving.md>), [hex](<https://devfeed.tech/tags/hex.md>), [langchain](<https://devfeed.tech/tags/langchain.md>), [learn](<https://devfeed.tech/tags/learn.md>), [questions](<https://devfeed.tech/tags/questions.md>), [self-serve-analytics](<https://devfeed.tech/tags/self-serve-analytics.md>)

### AI overview

LangChain reports a 40x increase in self-serve analytics questions after giving its Hex agent layered context. It syncs dbt table and column descriptions into BigQuery and gives the agent access to the dbt repo and DAG, letting it inspect model logic and dependencies. In Hex, version-controlled workspace guides explain business terms and metrics, a semantic layer defines key metrics and joins consistently, and selective endorsements identify trusted models and dashboards; changes to endorsed dashboards require data team review before publishing. The data team shifted toward maintaining context so other departments can query the agent directly. LangChain says this works despite documentation lagging business changes, provided context is updated continuously, users learn how to use the agent, and the team measures impact to find gaps. Evals and further context enrichment through Hex's CLI are on its roadmap.

### Source excerpt

LangChain gave its Hex agent layered context, turning self-serve analytics questions from occasional to routine.

## How a one-person data team at Keep built real-time banking infrastructure on dbt

DevFeed: [How a one-person data team at Keep built real-time banking infrastructure on dbt](<https://devfeed.tech/articles/how-a-one-person-data-team-at-keep-built-real-time-banking-infrastructure-on-dbt-102649.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/keep-real-time-banking-dbt-data-platform>)

Author: Daniel Poppy

Published: 2026-10-27T18:03:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [Snowflake streaming data architecture](<https://devfeed.tech/topics/snowflake-streaming-data-architecture.md>), [streaming data analytics](<https://devfeed.tech/topics/streaming-data-analytics.md>), [Event-Streaming](<https://devfeed.tech/topics/event-streaming.md>), [stream financial data to Snowflake](<https://devfeed.tech/topics/stream-financial-data-to-snowflake.md>), [Finance](<https://devfeed.tech/topics/finance.md>), [build vs buy database solutions](<https://devfeed.tech/topics/build-vs-buy-database-solutions.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>)

Tags: [analytics](<https://devfeed.tech/tags/analytics.md>), [banking](<https://devfeed.tech/tags/banking.md>), [banking-infrastructure](<https://devfeed.tech/tags/banking-infrastructure.md>), [data](<https://devfeed.tech/tags/data.md>), [data-team](<https://devfeed.tech/tags/data-team.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [keep](<https://devfeed.tech/tags/keep.md>), [learn](<https://devfeed.tech/tags/learn.md>), [real-time](<https://devfeed.tech/tags/real-time.md>), [snowflake](<https://devfeed.tech/tags/snowflake.md>)

### AI overview

Keep's one-person data team replaced spreadsheet-based underwriting that took a business day with Snowflake-triggered processing that completes in under a minute after bank data arrives. Artie copies Postgres and MongoDB changes into Snowflake at sub-second latency; Snowflake streams and tasks flatten incoming records, and dbt governs models, including transaction deduplication and risk features. The team chose this stack to avoid maintaining custom streaming infrastructure, accepting reduced observability in the real-time components. Custom freshness tests help catch failures, though some can be silent. The under-one-minute figure excludes Flinks' bank connection, which can take up to 30 minutes (under five minutes 95% of the time). Each underwriting result is stored with its producing code's git hash, supporting traceability and portfolio-wide reprocessing.

### Source excerpt

Keep's one-person data team used dbt, Artie, and Snowflake to cut credit underwriting from one business day to under a minute.

## Ship AI agents like tables: dbt materializations for Snowflake CoWork

DevFeed: [Ship AI agents like tables: dbt materializations for Snowflake CoWork](<https://devfeed.tech/articles/ship-ai-agents-like-tables-dbt-materializations-for-snowflake-cowork-102653.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/snowflake-cowork-ai-agents-dbt-materializations>)

Author: Daniel Poppy

Published: 2026-10-27T18:01:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [CI/CD](<https://devfeed.tech/topics/cicd.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ci](<https://devfeed.tech/tags/ci.md>), [cowork](<https://devfeed.tech/tags/cowork.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [learn](<https://devfeed.tech/tags/learn.md>), [models](<https://devfeed.tech/tags/models.md>), [process](<https://devfeed.tech/tags/process.md>), [snowflake](<https://devfeed.tech/tags/snowflake.md>)

### AI overview

Under Armour brought Snowflake AI agents into its established dbt deployment process by defining custom `agent` and `cortex_search_service` materializations, alongside a Snowflake Labs semantic-view materialization. Each object becomes a dbt model in the dependency graph, so teams can inspect downstream impacts and build modified objects in CI. A post-hook applies environment-specific grants, fails when a required role is missing, and registers production agents in Snowflake CoWork. Agents are altered in place to preserve conversation history and usage analytics. Five months after launch, more than 30 Cortex objects across three domains and environments had upstream and downstream lineage and a named reviewer for every change. The approach replaces direct production editing and manual recovery with pull requests and git reverts. Analyst-built objects still need manual migration from the Snowflake UI, and user access requests remain a manual handoff; CAB approvals and fixed release schedules also remain deliberate steps.

### Source excerpt

Under Armour turned Snowflake AI agents into dbt models, giving every agent the same review and CI process as every table.

## Charts worth stealing: the design calls BI should ship

DevFeed: [Charts worth stealing: the design calls BI should ship](<https://devfeed.tech/articles/charts-worth-stealing-the-design-calls-bi-should-ship-102637.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/bi-dashboard-chart-design-inspiration>)

Author: Daniel Poppy

Published: 2026-10-27T17:58:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [Data visualization](<https://devfeed.tech/topics/data-visualization.md>), [User interface design](<https://devfeed.tech/topics/ui-design.md>), [inclusive design](<https://devfeed.tech/topics/inclusive-design.md>), [Dark Mode](<https://devfeed.tech/topics/dark-mode.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [aeronautics](<https://devfeed.tech/tags/aeronautics.md>), [agent](<https://devfeed.tech/tags/agent.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [background](<https://devfeed.tech/tags/background.md>), [bi](<https://devfeed.tech/tags/bi.md>), [bi-tools](<https://devfeed.tech/tags/bi-tools.md>), [black](<https://devfeed.tech/tags/black.md>), [built-in](<https://devfeed.tech/tags/built-in.md>), [charts](<https://devfeed.tech/tags/charts.md>), [colors](<https://devfeed.tech/tags/colors.md>), [dashboard](<https://devfeed.tech/tags/dashboard.md>), [dashboards](<https://devfeed.tech/tags/dashboards.md>), [data](<https://devfeed.tech/tags/data.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [demand](<https://devfeed.tech/tags/demand.md>), [design](<https://devfeed.tech/tags/design.md>), [digital](<https://devfeed.tech/tags/digital.md>), [display](<https://devfeed.tech/tags/display.md>), [learn](<https://devfeed.tech/tags/learn.md>)

### AI overview

dbt Charts pairs flexible chart forms with defaults aimed at readability and consistency. At a dbt Summit 2026 session, RJ Andrews showed how a bar chart can become an area chart, line graph or scatterplot by changing a few lines of YAML. The tool adapts date-label density, keeps category colors consistent across dashboard charts, and uses harmonious palettes checked against several types of color vision deficiency. Its three themes--Clarity, Neon and Paper--adjust typography as well as color. Other features include aligned table decimals without fabricated precision, tooltips that highlight stacked segments and totals, shareable dashboard filters encoded in URLs, and review warnings for overly dense data. Pattern fills are built but not yet publicly available. The excerpt says dbt Charts and its dashboards launched as open source that week; it does not establish broader adoption or performance.

### Source excerpt

RJ Andrews explains the design defaults behind dbt Charts, from consistent colors to the small choices that make charts work.

## dbt Charts: a new open-source language for dashboards that live inside your dbt repo

DevFeed: [dbt Charts: a new open-source language for dashboards that live inside your dbt repo](<https://devfeed.tech/articles/dbt-charts-a-new-open-source-language-for-dashboards-that-live-inside-your-dbt-repo-102643.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/dbt-charts-open-source-dashboard-yaml>)

Author: Daniel Poppy

Published: 2026-10-27T17:54:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [Data visualization](<https://devfeed.tech/topics/data-visualization.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-tools](<https://devfeed.tech/tags/ai-tools.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [apache-2-0](<https://devfeed.tech/tags/apache-2-0.md>), [branch](<https://devfeed.tech/tags/branch.md>), [built-in](<https://devfeed.tech/tags/built-in.md>), [charts](<https://devfeed.tech/tags/charts.md>), [conversational-ai](<https://devfeed.tech/tags/conversational-ai.md>), [css](<https://devfeed.tech/tags/css.md>), [dashboards](<https://devfeed.tech/tags/dashboards.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [deploy](<https://devfeed.tech/tags/deploy.md>), [language](<https://devfeed.tech/tags/language.md>), [learn](<https://devfeed.tech/tags/learn.md>), [models](<https://devfeed.tech/tags/models.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [repo](<https://devfeed.tech/tags/repo.md>), [version](<https://devfeed.tech/tags/version.md>), [yaml](<https://devfeed.tech/tags/yaml.md>)

### AI overview

dbt Charts is an open-source YAML language for defining dashboards alongside dbt models. Its structure uses variables, queries, charts, and layout rows; a charts folder sits next to the models folder so changes can be versioned and deployed in the same Git branch and pass through the same tests. The language is Apache 2.0 licensed and can run locally with the dct CLI. A hosted service adds permissioning, sharing, a visual editor, version history, and conversational AI for building and editing dashboards. Its defaults include direct labels instead of legends, decimal-aligned support tables, and axis labels positioned to keep attention on recent data. The excerpt contrasts this code-centered approach with AI-generated dashboards that can vary in their technology and be difficult to trace to queries, and UI-first BI tools whose AI is described as limited in what it can change.

### Source excerpt

dbt Charts is a plain YAML language for building dashboards that version and deploy alongside your dbt models.

## What a year of building ADBC drivers taught Columnar about zero-copy's non-zero problems

DevFeed: [What a year of building ADBC drivers taught Columnar about zero-copy's non-zero problems](<https://devfeed.tech/articles/what-a-year-of-building-adbc-drivers-taught-columnar-about-zero-copy-s-non-zero-problems-102641.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/columnar-adbc-drivers-zero-copy-lessons>)

Author: Daniel Poppy

Published: 2026-10-27T17:50:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [apache-arrow](<https://devfeed.tech/topics/apache-arrow.md>), [copy](<https://devfeed.tech/topics/copy.md>), [Test-driven development](<https://devfeed.tech/topics/tdd.md>)

Tags: [dbt](<https://devfeed.tech/tags/dbt.md>), [drivers](<https://devfeed.tech/tags/drivers.md>), [learn](<https://devfeed.tech/tags/learn.md>), [production](<https://devfeed.tech/tags/production.md>), [type-system](<https://devfeed.tech/tags/type-system.md>)

### AI overview

ADBC provides a shared API that returns data in Apache Arrow's columnar format, avoiding serialization when databases return Arrow natively. In production, that benefit varies: many systems still return row-oriented or JSON data that drivers must convert, and SQL types differ sharply across vendors. The article details mismatches in numeric precision, string and binary handling, timestamp time zones, and list nulls; drivers must choose Arrow mappings before seeing actual data. ADBC eases setup through TOML driver manifests and connection profiles, but the excerpt notes gaps in synchronous APIs, type support, and metadata handling. Topol recommends testing schemas early, pinning driver versions, and tracking which version is in use.

### Source excerpt

Columnar co-founder Matt Topol shares the type-system challenges of building production ADBC drivers, now used by dbt v2.

## Why a semantic layer is the difference between a 96% and an 80% accurate AI analyst

DevFeed: [Why a semantic layer is the difference between a 96% and an 80% accurate AI analyst](<https://devfeed.tech/articles/why-a-semantic-layer-is-the-difference-between-a-96-and-an-80-accurate-ai-analyst-102651.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/semantic-layer-ai-analyst-accuracy>)

Author: Daniel Poppy

Published: 2026-10-27T17:46:00Z

Content type: article

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [ai](<https://devfeed.tech/tags/ai.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [anomaly-detection](<https://devfeed.tech/tags/anomaly-detection.md>), [context-engineering](<https://devfeed.tech/tags/context-engineering.md>), [context-layer](<https://devfeed.tech/tags/context-layer.md>), [data](<https://devfeed.tech/tags/data.md>), [data-modeling](<https://devfeed.tech/tags/data-modeling.md>), [data-quality](<https://devfeed.tech/tags/data-quality.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [documentation](<https://devfeed.tech/tags/documentation.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [fivetran](<https://devfeed.tech/tags/fivetran.md>), [grounding](<https://devfeed.tech/tags/grounding.md>), [learn](<https://devfeed.tech/tags/learn.md>), [llm-as-judge](<https://devfeed.tech/tags/llm-as-judge.md>), [mcp](<https://devfeed.tech/tags/mcp.md>), [mcp-server](<https://devfeed.tech/tags/mcp-server.md>), [semantic-layer](<https://devfeed.tech/tags/semantic-layer.md>)

### AI overview

In dbt Labs' team benchmark, the same model answered with 96-100% accuracy through a semantic layer, versus 80-91% when writing raw SQL against a database schema. Alex Noonan argues that language models can interpret questions, but reliable answers still depend on grounding business meaning. The talk describes three approaches: predefined metrics in a semantic layer, ontology-style entities and relationships, and an agent context layer that assembles relevant information at query time. Noonan highlights data tests, freshness checks, anomaly detection, machine-readable documentation, and consistent modeling as foundations for durable analytics. He distinguishes stable "just in time" context from search infrastructure for high-volume unstructured sources. For trust, he recommends visible correction in a public channel such as Slack and cautions that an LLM judge may share the evaluated model's blind spots. These are claims and recommendations reported in the supplied source text.

### Source excerpt

Why conversational analytics stalls and the data fundamentals that make it reliable.

## With great context comes great autonomy

DevFeed: [With great context comes great autonomy](<https://devfeed.tech/articles/with-great-context-comes-great-autonomy-102642.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/context-engineering-ai-agents-governed-autonomy>)

Author: Daniel Poppy

Published: 2026-10-27T17:43:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [context](<https://devfeed.tech/topics/context.md>), [knowledge-engineering](<https://devfeed.tech/topics/knowledge-engineering.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [MSP MCP](<https://devfeed.tech/topics/msp-mcp.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [api](<https://devfeed.tech/tags/api.md>), [autonomy](<https://devfeed.tech/tags/autonomy.md>), [cli](<https://devfeed.tech/tags/cli.md>), [concepts](<https://devfeed.tech/tags/concepts.md>), [confidence](<https://devfeed.tech/tags/confidence.md>), [confluence](<https://devfeed.tech/tags/confluence.md>), [context](<https://devfeed.tech/tags/context.md>), [context-layer](<https://devfeed.tech/tags/context-layer.md>), [data-infrastructure](<https://devfeed.tech/tags/data-infrastructure.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [documentation](<https://devfeed.tech/tags/documentation.md>), [fivetran](<https://devfeed.tech/tags/fivetran.md>), [glossary](<https://devfeed.tech/tags/glossary.md>), [hubspot](<https://devfeed.tech/tags/hubspot.md>), [learn](<https://devfeed.tech/tags/learn.md>)

### AI overview

Trustworthy AI agents depend on governed business context, not simply more data. Ben Moser of dbt Labs argues that teams should build context progressively: dbt models, tests, documentation, and lineage provide a governed base, while unmodeled sources such as support tickets, Slack threads, and institutional knowledge remain harder to use reliably. Fivetran Context Layer connects knowledge, operational, and analytics sources, indexes them for semantic search, and infers links between related concepts. It writes the resulting context as open Agents Schema tables in the warehouse, including a root table describing the schema and a concept table for company-specific terms; AI clients can query them over MCP. In a demo, an agent used these tables to find open opportunities with unresolved support tickets. A traces feature highlights low confidence or missing context so teams can curate it. The speakers caution that inaccurate context can compound errors and token costs, and recommend against trying to load everything at once.

### Source excerpt

AI agents need governed context, not just more data, to be trusted.

## dbt without the warehouse (or the bill): DuckDB end to end

DevFeed: [dbt without the warehouse (or the bill): DuckDB end to end](<https://devfeed.tech/articles/dbt-without-the-warehouse-or-the-bill-duckdb-end-to-end-102644.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/dbt-duckdb-warehouse-free-data-transformation>)

Author: Daniel Poppy

Published: 2026-10-27T17:36:00Z

Content type: news

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [DuckDB](<https://devfeed.tech/topics/duckdb.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [about](<https://devfeed.tech/tags/about.md>), [adapter](<https://devfeed.tech/tags/adapter.md>), [agent](<https://devfeed.tech/tags/agent.md>), [aggregation](<https://devfeed.tech/tags/aggregation.md>), [all](<https://devfeed.tech/tags/all.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [bill](<https://devfeed.tech/tags/bill.md>), [built-in](<https://devfeed.tech/tags/built-in.md>), [change](<https://devfeed.tech/tags/change.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [duckdb](<https://devfeed.tech/tags/duckdb.md>), [learn](<https://devfeed.tech/tags/learn.md>), [offline](<https://devfeed.tech/tags/offline.md>), [profile](<https://devfeed.tech/tags/profile.md>), [small](<https://devfeed.tech/tags/small.md>), [v2](<https://devfeed.tech/tags/v2.md>), [warehouse](<https://devfeed.tech/tags/warehouse.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

dbt v2 includes a built-in DuckDB adapter, and connecting it requires setting the profile output to duckdb and specifying a file path; omitting the path creates a temporary in-memory database. In a live demonstration, Hannes Mühleisen ran dbt fully offline over about 10 GB of compressed Wikipedia page-view CSVs stored locally. A build with window-function calculations across millions of article titles and an aggregation over roughly 300 million rows finished in about three seconds. DuckDB can query CSV, Parquet, and Iceberg files directly without importing them, and includes SQL features such as GROUP BY ALL, UNION BY NAME, and PIVOT. The excerpt presents DuckDB as a single-machine option for many datasets, while previewing DuckDB 2.0 features expected around October 2026.

### Source excerpt

dbt v2 ships with a built-in DuckDB adapter, so connecting dbt to a local, offline warehouse takes only a small profile change.

[Next page](<https://devfeed.tech/sources/dbt-labs.md?cursor=WyIyMDI2LTEwLTI3VDE3OjM2OjAwKzAwOjAwIiwgIjgyNzQ1OGQyLTE3MTQtNDZhOC05YzI2LWQ0ZDkzMjM4ZDRkZCJd>)