--- title: As humans write less code: workloads shift from human to... description: As analysts write less SQL by hand and prompt agents instead, verification becomes the new bottleneck in BI work. A Metabase survey found 74% of teams were... canonical: https://daily.dev/posts/as-humans-write-less-code-workloads-shift-from-human-to-agent-verification-with-iterative-test-suit-xipymqexy twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: As humans write less code: workloads shift from human to agent verification with iterative test suites | daily.dev og:description: As analysts write less SQL by hand and prompt agents instead, verification becomes the new bottleneck in BI work. A Metabase survey found 74% of teams were... og:url: https://daily.dev/posts/as-humans-write-less-code-workloads-shift-from-human-to-agent-verification-with-iterative-test-suit-xipymqexy og:image: https://api.daily.dev/og/posts/xiPyMqeXY.png og:image:alt: As humans write less code: workloads shift from human to agent verification with iterative test suites og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # As humans write less code: workloads shift from human to agent verification with iterative test suites **[Metabase](https://daily.dev/sources/metabase)** · 14 min read · 2 upvotes · 0 comments ## Summary As analysts write less SQL by hand and prompt agents instead, verification becomes the new bottleneck in BI work. A Metabase survey found 74% of teams were building with AI but only 2% were scaling it, with the bottleneck having moved from creating analysis to checking it. Metabase CEO Sameer Al-Sakran argues verification needs five testing layers: linting, a compiler-like full-tree check, unit tests, integration tests, and end-to-end tests. He also says mutable state undermines trust in outputs, and that business people's intuition still catches errors math alone can't. The piece points to Metabase's upcoming Transform tests (v65) and its Data Studio product as early steps toward a unified lineage graph for verification. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What are the different levels of testing needed to verify AI-generated SQL and dashboards in a BI stack? Five hierarchy levels are needed: linting (static analysis of metrics and configs without running them), a compiler-style whole-tree check that resolves the full lineage before deploying, unit tests on small functions with value assertions, integration tests across modules, and end-to-end tests that run the compiled system against seeded data sets. _daily.dev surfaces practical approaches like this for teams building AI-assisted analytics pipelines._ ### Why is mutable state a problem when verifying data pipeline outputs? Mutable state undermines trust in outputs because planning and verification should happen against a fixed, known state, such as a test data set with a known expected outcome. Without that fixed reference, verification becomes guesswork rather than a reliable check, which is why fixed seed data with rollback after each run is preferred for testing transforms. _Developers debugging flaky pipeline tests can track state-management practices like this on daily.dev._ ### How does dbt support shifting test assertions left in the build process? Since dbt Core version 1.8, unit tests can check model logic against static inputs before materialization occurs, and model contracts are enforced at build time rather than only after the data is computed. This moves verification earlier than traditional runtime checks like dbt data tests or Airflow checks, catching issues before the pipeline actually runs. _Teams upgrading their dbt workflow can keep up with changes like this via daily.dev._ ## Similar posts on daily.dev - [Automated Verification Ensures End-to-End Correctness and Human Accountability](https://daily.dev/posts/automated-verification-ensures-end-to-end-correctness-and-human-accountability-je4r5se5b) · Arpit Bhayani · 0 upvotes · 0 comments - [Why Code Verification Matters More Than Ever in the Age of AI](https://daily.dev/posts/why-code-verification-matters-more-than-ever-in-the-age-of-ai-roteb1arc) · ByteByteGo · 19 upvotes · 2 comments - [Agents have made CI the bottleneck. Faster pipelines are the wrong fix.](https://daily.dev/posts/agents-have-made-ci-the-bottleneck-faster-pipelines-are-the-wrong-fix--fott4jdmy) · The New Stack · 2 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#testing](https://daily.dev/tags/testing), [#data-engineering](https://daily.dev/tags/data-engineering), [#bi](https://daily.dev/tags/bi) [View this post on daily.dev](https://daily.dev/posts/as-humans-write-less-code-workloads-shift-from-human-to-agent-verification-with-iterative-test-suit-xipymqexy) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"As humans write less code: workloads shift from human to agent verification with iterative test suites","url":"https://daily.dev/posts/as-humans-write-less-code-workloads-shift-from-human-to-agent-verification-with-iterative-test-suit-xipymqexy","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/as-humans-write-less-code-workloads-shift-from-human-to-agent-verification-with-iterative-test-suit-xipymqexy"},"datePublished":"2026-10-27T12:26:00.504Z","dateModified":"2026-10-27T12:26:34.245Z","description":"As analysts write less SQL by hand and prompt agents instead, verification becomes the new bottleneck in BI work. A Metabase survey found 74% of teams were...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/1bfafe506e8f99576cda58575138f3ba?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/1bfafe506e8f99576cda58575138f3ba?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Metabase","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Metabase","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/97c4b2294a914263b4f3b4a82a5f896b","url":"https://daily.dev/sources/metabase"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/as-humans-write-less-code-workloads-shift-from-human-to-agent-verification-with-iterative-test-suit-xipymqexy","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,testing,data-engineering,bi","timeRequired":"PT14M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Metabase","item":"https://daily.dev/sources/metabase"},{"@type":"ListItem","position":3,"name":"As humans write less code: workloads shift from human to agent verification with iterative test suites"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/as-humans-write-less-code-workloads-shift-from-human-to-agent-verification-with-iterative-test-suit-xipymqexy#faq","mainEntity":[{"@type":"Question","name":"What are the different levels of testing needed to verify AI-generated SQL and dashboards in a BI stack?","acceptedAnswer":{"@type":"Answer","text":"Five hierarchy levels are needed: linting (static analysis of metrics and configs without running them), a compiler-style whole-tree check that resolves the full lineage before deploying, unit tests on small functions with value assertions, integration tests across modules, and end-to-end tests that run the compiled system against seeded data sets. daily.dev surfaces practical approaches like this for teams building AI-assisted analytics pipelines."}},{"@type":"Question","name":"Why is mutable state a problem when verifying data pipeline outputs?","acceptedAnswer":{"@type":"Answer","text":"Mutable state undermines trust in outputs because planning and verification should happen against a fixed, known state, such as a test data set with a known expected outcome. Without that fixed reference, verification becomes guesswork rather than a reliable check, which is why fixed seed data with rollback after each run is preferred for testing transforms. Developers debugging flaky pipeline tests can track state-management practices like this on daily.dev."}},{"@type":"Question","name":"How does dbt support shifting test assertions left in the build process?","acceptedAnswer":{"@type":"Answer","text":"Since dbt Core version 1.8, unit tests can check model logic against static inputs before materialization occurs, and model contracts are enforced at build time rather than only after the data is computed. This moves verification earlier than traditional runtime checks like dbt data tests or Airflow checks, catching issues before the pipeline actually runs. Teams upgrading their dbt workflow can keep up with changes like this via daily.dev."}}]} ```