# Inferencing Published articles for Inferencing. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Something New Has to Sit Between Agents and Databases DevFeed: [Something New Has to Sit Between Agents and Databases]() Original publisher: [Read original article]() Author: flashdba Published: 2026-06-15T12:46:59Z Content type: opinion Language: en Sources: [flashdba]() Topics: [Architecture & Design](), [Enterprise Architecture](), [Artificial Intelligence](), [Databases]() Tags: [agentic](), [agentic-ai](), [ai](), [architecture](), [cloud](), [credentials](), [databases](), [databases-and-agentic-ai](), [enterprise-architecture](), [inferencing](), [performance]() ### AI overview This article argues that agentic AI removes human dependencies that enterprise architectures and systems of record have implicitly relied on. It proposes a deliberate intermediary layer between agents and databases to provide identity, intent recovery, rate control, behavioural reasoning, and accountability requirements that existing infrastructure does not address. ### Source excerpt Enterprise architecture depended on humans in ways nobody designed - and agentic AI is removing those dependencies. Something new has to sit between agents and databases, and it needs to be built deliberately. ## Enterprise Architecture Quietly Depended on Humans in Far More Places Than We Realised DevFeed: [Enterprise Architecture Quietly Depended on Humans in Far More Places Than We Realised]() Original publisher: [Read original article]() Author: flashdba Published: 2026-06-08T14:19:07Z Content type: opinion Language: en Sources: [flashdba]() Topics: [Enterprise Architecture](), [Architecture & Design](), [Artificial Intelligence](), [Actor](), [properties]() Tags: [actor](), [agentic](), [agentic-ai](), [ai](), [architecture](), [cloud](), [databases](), [databases-and-agentic-ai](), [enterprise-architecture](), [inferencing](), [performance](), [properties]() ### AI overview This article argues that enterprise architecture depended on implicit properties supplied by human participants, including predictable workload, intent, accountability, feedback, and limited capacity. It concludes that agentic AI exposes pressure across multiple technical and organisational domains by removing those assumptions simultaneously. ### Source excerpt Enterprise architecture was built around a human actor whose presence provided properties nobody designed and nobody priced. Agentic AI is the first technology to remove those properties simultaneously - and the absence is now visible across every domain. ## Why Agentic AI Challenges Traditional Enterprise Change Management DevFeed: [Why Agentic AI Challenges Traditional Enterprise Change Management]() Original publisher: [Read original article]() Author: flashdba Published: 2026-06-02T12:56:04Z Content type: opinion Language: en Sources: [flashdba]() Topics: [Artificial Intelligence](), [reproducibility](), [Testing](), [Deployment]() Tags: [agentic](), [agentic-ai](), [ai](), [cloud](), [databases](), [databases-and-agentic-ai](), [deployment](), [inferencing](), [performance](), [release](), [reproducibility](), [testing]() ### AI overview The article argues that enterprise change management depends on software behaving reproducibly. Agentic AI can produce different outputs from the same inputs, weakening the assumptions behind testing, UAT sign-off, approval gates, and rollback procedures. ### Source excerpt Enterprise change management was built around software that behaves reproducibly - and that assumption is load-bearing in ways that only become visible when agentic AI removes it. ## Agentic AI Exposes Gaps Between Deployment Teams and System Owners DevFeed: [Agentic AI Exposes Gaps Between Deployment Teams and System Owners]() Original publisher: [Read original article]() Author: flashdba Published: 2026-05-25T10:47:43Z Content type: opinion Language: en Sources: [flashdba]() Topics: [deploying agentic AI](), [Enterprise Architecture](), [Architecture & Design](), [audit]() Tags: [agentic](), [agentic-ai](), [agents](), [ai](), [ai-agents](), [architecture](), [audit](), [cloud](), [databases](), [databases-and-agentic-ai](), [enterprise-architecture](), [governance](), [inferencing](), [infrastructure](), [performance](), [systems]() ### AI overview The article argues that agentic AI is eroding the technical boundary between application and infrastructure teams while leaving organizational accountability and governance structures inadequate. It highlights impacts on workloads, identities, audit trails, and capacity costs. ### Source excerpt The people deploying AI agents are not the people responsible for the systems they stress - and the governance structures that might close that gap don't yet exist. ## You Won't See Failure First. You'll See Cost DevFeed: [You Won't See Failure First. You'll See Cost]() Original publisher: [Read original article]() Author: flashdba Published: 2026-05-18T20:59:25Z Content type: opinion Language: en Sources: [flashdba]() Topics: [Artificial Intelligence](), [Cloud](), [Monitoring]() Tags: [agentic](), [agentic-ai](), [ai](), [cloud](), [cost](), [databases](), [databases-and-agentic-ai](), [inferencing](), [monitoring](), [performance]() ### AI overview This commentary argues that agentic AI can fail without obvious alerts or immediate system breakdowns. Instead, the first warning may be unexpected cloud spending, shifting operational risk from visible incidents to costs that accumulate quietly. ### Source excerpt Agentic AI introduces a failure mode that doesn't announce itself. The first signal isn't a red dashboard or a paged engineer - it's a line in the cloud bill. ## How Agentic AI Can Bypass the Application Layer's Database Protections DevFeed: [How Agentic AI Can Bypass the Application Layer's Database Protections]() Original publisher: [Read original article]() Author: flashdba Published: 2026-05-11T11:06:56Z Content type: opinion Language: en Sources: [flashdba]() Topics: [Architecture & Design](), [Artificial Intelligence](), [Database]() Tags: [agentic](), [agentic-ai](), [ai](), [ai-agents](), [architecture](), [business-logic](), [cloud](), [database](), [databases](), [databases-and-agentic-ai](), [inferencing](), [performance](), [rate-limiting]() ### AI overview The article argues that application layers unintentionally protect databases through query abstraction, input validation, rate limiting, and human-oriented traffic patterns. Direct database access by agent frameworks can bypass these protections, allowing unconstrained queries and exposing the database to unfamiliar load and access patterns. ### Source excerpt The application layer was never designed as a database security boundary - but it acted as one. Agentic AI removes that protection, via bypass or overwhelm, and the database is left exposed. ## How AI Agents Challenge Enterprise Database Capacity Assumptions DevFeed: [How AI Agents Challenge Enterprise Database Capacity Assumptions]() Original publisher: [Read original article]() Author: flashdba Published: 2026-05-04T09:17:00Z Content type: opinion Language: en Sources: [flashdba]() Topics: [Database](), [Artificial Intelligence](), [Cloud]() Tags: [agentic](), [ai](), [ai-agents](), [autoscaling](), [cloud](), [concurrent](), [cost](), [database](), [databases](), [databases-and-agentic-ai](), [inferencing](), [parallel](), [performance](), [real-time](), [scale]() ### AI overview This commentary explains how AI agents can invalidate traditional enterprise database capacity assumptions. Unlike human-driven workloads, agents can issue recursive, parallel requests at machine speed, potentially changing concurrency, autoscaling behavior, and cloud cost forecasts. ### Source excerpt Every enterprise database capacity model rested on assumptions about human behaviour. Agents remove those assumptions - and in the cloud, that gap renews every month. ## How Agentic AI Creates Gaps in Audit-Trail Accountability DevFeed: [How Agentic AI Creates Gaps in Audit-Trail Accountability]() Original publisher: [Read original article]() Author: flashdba Published: 2026-04-27T13:43:04Z Content type: opinion Language: en Sources: [flashdba]() Topics: [AI Agent](), [audit trail](), [audit]() Tags: [account](), [agentic](), [agentic-ai](), [ai](), [ai-agent](), [audit](), [audit-trail](), [changes](), [cloud](), [commit](), [compliance](), [credentials](), [databases](), [databases-and-agentic-ai](), [identity](), [inferencing](), [model](), [performance](), [regulatory](), [service]() ### AI overview The article argues that conventional audit trails can record database changes accurately while failing to explain an AI agent's reasoning or establish accountable human ownership. It identifies unstable agent identity and the absence of externally documented reasoning as the central gaps. ### Source excerpt The audit trail still runs. Every commit is recorded. But agentic AI has broken the two assumptions it was built on - and the incompleteness is invisible until you need it. ## Agentic AI removes human safeguards from enterprise data feedback loops DevFeed: [Agentic AI removes human safeguards from enterprise data feedback loops]() Original publisher: [Read original article]() Author: flashdba Published: 2026-04-20T22:04:35Z Content type: opinion Language: en Sources: [flashdba]() Topics: [Artificial Intelligence](), [data-architecture](), [data-engineering](), [rate-limiting](), [audit trail]() Tags: [agentic](), [agentic-ai](), [ai](), [audit-trail](), [cloud](), [data-architecture](), [data-engineering](), [databases](), [databases-and-agentic-ai](), [inferencing](), [performance](), [rate-limiting]() ### AI overview This opinion article argues that enterprise data architectures traditionally relied on human involvement as an implicit safeguard. Delays, approvals and handoffs helped absorb errors, create audit points and limit the speed of changes. Agentic AI can close the feedback loop at machine speed, potentially removing those protections along with the human decision point. ### Source excerpt Classic enterprise data architecture had an implicit safeguard built into it. The human in the loop provided error absorption, audit accretion and natural rate-limiting - none of which were ever specified. Agentic AI removes the human. It removes all of those protections simultaneously.