# Before you write the brief. Source: https://customlabs.io/faq/ Updated: 2026-09-12 Questions # Before you write the brief. The questions we get asked most, answered plainly: no hedging, no marketing copy. 01 How do engagements usually start? + With a discovery session, not a sales call. We map the problem, the data, and the constraints, then send a written brief and estimate. If the scope is still fuzzy after that, we'll sometimes run a small paid discovery phase first rather than guess. See the Process page for the full shape of an engagement. 02 How much does this cost, and how long does it take? + It depends on scope, so we quote ranges, not a fixed price, once we understand the problem. Most first engagements ship a production slice within a few weeks. The budget ranges on the contact form are a reasonable starting point for sizing your brief. 03 Who owns the code and the models? + You do. Everything we build is documented and observable. We hand it over at the end of the engagement. We're consultants, not vendors with a kill-switch. 04 Do you use our data to train models? + No. Your data stays in your environment. We build on top of it, not off it. 05 We're not sure AI is even the right answer. Will you tell us? + Yes. Our Strategy & Readiness engagements exist specifically to give you an honest answer, including "don't build this" when that's the right call. 06 Which models or stack do you build on? + We're model-agnostic — Anthropic, OpenAI, open-weight, or self-hosted — and we build on open standards rather than locking you into one vendor's stack. Every other answer ## Every other question, answered on the page that owns it. The six questions above come up before a brief. Every other question the site answers already has a home on a real page. Search below, or browse by group. Showing all 206 questions ### Before you start What it costs, and who ends up owning the result. #### AI Integrations - [Can you build on our existing stack?](https://customlabs.io/services/ai-integrations/#faq-can-you-build-on-our-existing-stack) - [Which model providers do you support?](https://customlabs.io/services/ai-integrations/#faq-which-model-providers-do-you-support) - [How do you keep latency and cost predictable?](https://customlabs.io/services/ai-integrations/#faq-how-do-you-keep-latency-and-cost-predictable) - [Do we own the integration afterward?](https://customlabs.io/services/ai-integrations/#faq-do-we-own-the-integration-afterward) #### AI Readiness / Ship Audit - [What do we walk away with?](https://customlabs.io/diagnostic/ship-audit/#faq-what-do-we-walk-away-with) - [How long does it take?](https://customlabs.io/diagnostic/ship-audit/#faq-how-long-does-it-take) - [Does our data leave our environment?](https://customlabs.io/diagnostic/ship-audit/#faq-does-our-data-leave-our-environment) - [What does the audit cover?](https://customlabs.io/diagnostic/ship-audit/#faq-what-does-the-audit-cover) - [How is this different from the free AI Readiness Scorecard?](https://customlabs.io/diagnostic/ship-audit/#faq-how-is-this-different-from-the-free-ai-readiness-scorecard) - [What happens after the audit?](https://customlabs.io/diagnostic/ship-audit/#faq-what-happens-after-the-audit) #### AI Technical Diligence - [Can you do investor-side technical due diligence?](https://customlabs.io/diagnostic/technical-diligence/#faq-can-you-do-investor-side-technical-due-diligence) - [Who is this for?](https://customlabs.io/diagnostic/technical-diligence/#faq-who-is-this-for) - [How fast can you turn this around?](https://customlabs.io/diagnostic/technical-diligence/#faq-how-fast-can-you-turn-this-around) - [Does the target’s data leave their environment?](https://customlabs.io/diagnostic/technical-diligence/#faq-does-the-target-s-data-leave-their-environment) - [Do you talk to the target’s team directly?](https://customlabs.io/diagnostic/technical-diligence/#faq-do-you-talk-to-the-target-s-team-directly) - [What if the finding changes the deal?](https://customlabs.io/diagnostic/technical-diligence/#faq-what-if-the-finding-changes-the-deal) #### Choosing a partner - [Is this page trying to talk me out of hiring you?](https://customlabs.io/choosing-a-partner/#faq-is-this-page-trying-to-talk-me-out-of-hiring-you) - [Do all 24 questions apply to every engagement?](https://customlabs.io/choosing-a-partner/#faq-do-all-24-questions-apply-to-every-engagement) - [What if a partner can't answer one of these yet?](https://customlabs.io/choosing-a-partner/#faq-what-if-a-partner-can-t-answer-one-of-these-yet) - [Can I use this if I've already picked a partner?](https://customlabs.io/choosing-a-partner/#faq-can-i-use-this-if-i-ve-already-picked-a-partner) - [Why publish the routes where you're not the best fit?](https://customlabs.io/choosing-a-partner/#faq-why-publish-the-routes-where-you-re-not-the-best-fit) #### Custom Development - [Who owns the code?](https://customlabs.io/services/custom-development/#faq-who-owns-the-code) - [How do you prove the AI behaves?](https://customlabs.io/services/custom-development/#faq-how-do-you-prove-the-ai-behaves) - [Can our team maintain it after handover?](https://customlabs.io/services/custom-development/#faq-can-our-team-maintain-it-after-handover) - [How fast do you ship a first slice?](https://customlabs.io/services/custom-development/#faq-how-fast-do-you-ship-a-first-slice) #### Diagnostics - [Why do you charge for the diagnostic?](https://customlabs.io/diagnostic/#faq-why-do-you-charge-for-the-diagnostic) - [Can we skip it and just build?](https://customlabs.io/diagnostic/#faq-can-we-skip-it-and-just-build) #### Pricing - [Do we own the code?](https://customlabs.io/pricing/#faq-do-we-own-the-code) - [Why isn’t there a fixed build price?](https://customlabs.io/pricing/#faq-why-isn-t-there-a-fixed-build-price) - [How fast do engagements move?](https://customlabs.io/pricing/#faq-how-fast-do-engagements-move) #### Readiness & Diligence - [What does the assessment cover?](https://customlabs.io/services/readiness-diligence/#faq-what-does-the-assessment-cover) #### Services - [Which service should we start with?](https://customlabs.io/services/#faq-which-service-should-we-start-with) - [Can we combine more than one service in a single engagement?](https://customlabs.io/services/#faq-can-we-combine-more-than-one-service-in-a-single-engagement) - [How do you size and price a combined engagement?](https://customlabs.io/services/#faq-how-do-you-size-and-price-a-combined-engagement) - [What if we're not sure which situation fits our case?](https://customlabs.io/services/#faq-what-if-we-re-not-sure-which-situation-fits-our-case) #### Strategy & Architecture - [Will you tell us NOT to build something?](https://customlabs.io/services/strategy-architecture/#faq-will-you-tell-us-not-to-build-something) - [How is this different from a big-consultancy deck?](https://customlabs.io/services/strategy-architecture/#faq-how-is-this-different-from-a-big-consultancy-deck) - [Do you lock us into a stack?](https://customlabs.io/services/strategy-architecture/#faq-do-you-lock-us-into-a-stack) ### What we build Our products, and the architecture and tooling underneath them. #### Architectures - [What exactly counts as "a reference architecture" here?](https://customlabs.io/architectures/#faq-what-exactly-counts-as-a-reference-architecture-here) - [How do I pick which one applies to us?](https://customlabs.io/architectures/#faq-how-do-i-pick-which-one-applies-to-us) - [Do these replace the patterns library?](https://customlabs.io/architectures/#faq-do-these-replace-the-patterns-library) - [Why do two of these mention your own products?](https://customlabs.io/architectures/#faq-why-do-two-of-these-mention-your-own-products) - [What if our system is a mix of two of these?](https://customlabs.io/architectures/#faq-what-if-our-system-is-a-mix-of-two-of-these) #### Beemy - [Is Beemy available yet?](https://customlabs.io/products/beemy/#faq-is-beemy-available-yet) - [Will it mix my work and personal messages?](https://customlabs.io/products/beemy/#faq-will-it-mix-my-work-and-personal-messages) - [Who is Beemy for?](https://customlabs.io/products/beemy/#faq-who-is-beemy-for) #### BizBinder - [Who is it for?](https://customlabs.io/products/bizbinder/#faq-who-is-it-for) - [Is BizBinder available yet?](https://customlabs.io/products/bizbinder/#faq-is-bizbinder-available-yet) - [What happens to old jobs and invoices?](https://customlabs.io/products/bizbinder/#faq-what-happens-to-old-jobs-and-invoices) #### CodeHerder - [What is CodeHerder for?](https://customlabs.io/products/codeherder/#faq-what-is-codeherder-for) - [Does it work with any AI coding agent?](https://customlabs.io/products/codeherder/#faq-does-it-work-with-any-ai-coding-agent) - [Do I need to change how my agents work to use it?](https://customlabs.io/products/codeherder/#faq-do-i-need-to-change-how-my-agents-work-to-use-it) #### Context Engineering - [Isn't this just prompt engineering with a longer name?](https://customlabs.io/context-engineering/#faq-isn-t-this-just-prompt-engineering-with-a-longer-name) - [How is this different from /tool-design/?](https://customlabs.io/context-engineering/#faq-how-is-this-different-from-tool-design) - [We just bought a model with a much bigger context window. Does this still matter?](https://customlabs.io/context-engineering/#faq-we-just-bought-a-model-with-a-much-bigger-context-window-does-this-still-matter) - [What's the single most common mistake on this page?](https://customlabs.io/context-engineering/#faq-what-s-the-single-most-common-mistake-on-this-page) - [Do we need all 24 practices before our first agent ships?](https://customlabs.io/context-engineering/#faq-do-we-need-all-24-practices-before-our-first-agent-ships) - [Where does the memory surface stop and a vector database start?](https://customlabs.io/context-engineering/#faq-where-does-the-memory-surface-stop-and-a-vector-database-start) #### CostMon - [Is CostMon available yet?](https://customlabs.io/products/costmon/#faq-is-costmon-available-yet) - [What will it connect to?](https://customlabs.io/products/costmon/#faq-what-will-it-connect-to) #### CustomHosted - [What can I host on CustomHosted?](https://customlabs.io/products/customhosted/#faq-what-can-i-host-on-customhosted) - [Is there lock-in?](https://customlabs.io/products/customhosted/#faq-is-there-lock-in) - [Who handles support?](https://customlabs.io/products/customhosted/#faq-who-handles-support) #### FreeTier - [What counts as a "free tier" on FreeTier?](https://customlabs.io/products/freetier/#faq-what-counts-as-a-free-tier-on-freetier) - [Is FreeTier free itself?](https://customlabs.io/products/freetier/#faq-is-freetier-free-itself) - [How often is it updated?](https://customlabs.io/products/freetier/#faq-how-often-is-it-updated) #### GreatAPIs - [How is GreatAPIs different from FreeTier?](https://customlabs.io/products/greatapis/#faq-how-is-greatapis-different-from-freetier) - [What does the auth-type filter cover?](https://customlabs.io/products/greatapis/#faq-what-does-the-auth-type-filter-cover) - [Are the HTTPS/CORS flags kept current?](https://customlabs.io/products/greatapis/#faq-are-the-https-cors-flags-kept-current) #### MCP - [Isn't MCP itself the security boundary?](https://customlabs.io/mcp/#faq-isn-t-mcp-itself-the-security-boundary) - [We only connect first-party servers we wrote ourselves. Does the trust boundary still apply?](https://customlabs.io/mcp/#faq-we-only-connect-first-party-servers-we-wrote-ourselves-does-the-trust-boundary-still-apply) - [What's the single fastest way to get burned on this surface?](https://customlabs.io/mcp/#faq-what-s-the-single-fastest-way-to-get-burned-on-this-surface) - [Do we need all 24 controls before connecting a single server?](https://customlabs.io/mcp/#faq-do-we-need-all-24-controls-before-connecting-a-single-server) - [How is this different from the trust-boundary surface on /tool-design/?](https://customlabs.io/mcp/#faq-how-is-this-different-from-the-trust-boundary-surface-on-tool-design) - [Does Dynamic Client Registration still work?](https://customlabs.io/mcp/#faq-does-dynamic-client-registration-still-work) #### Reserver - [Is Reserver available yet?](https://customlabs.io/products/reserver/#faq-is-reserver-available-yet) - [Does it only cover EC2?](https://customlabs.io/products/reserver/#faq-does-it-only-cover-ec2) - [How is pricing structured?](https://customlabs.io/products/reserver/#faq-how-is-pricing-structured) #### Tool design - [How many tools is actually too many?](https://customlabs.io/tool-design/#faq-how-many-tools-is-actually-too-many) - [Do we need MCP, or is in-process enough?](https://customlabs.io/tool-design/#faq-do-we-need-mcp-or-is-in-process-enough) - [What's the single highest-leverage rule to start with?](https://customlabs.io/tool-design/#faq-what-s-the-single-highest-leverage-rule-to-start-with) - [How does this relate to the Eval Stack?](https://customlabs.io/tool-design/#faq-how-does-this-relate-to-the-eval-stack) - [Isn't this just prompt engineering?](https://customlabs.io/tool-design/#faq-isn-t-this-just-prompt-engineering) - [What's context engineering, and how is it different from prompt engineering?](https://customlabs.io/tool-design/#faq-what-s-context-engineering-and-how-is-it-different-from-prompt-engineering) #### txtfetch - [Is txtfetch available yet?](https://customlabs.io/products/txtfetch/#faq-is-txtfetch-available-yet) - [What formats will it support?](https://customlabs.io/products/txtfetch/#faq-what-formats-will-it-support) - [Does it store the documents I send it?](https://customlabs.io/products/txtfetch/#faq-does-it-store-the-documents-i-send-it) ### Running it safely Governance and security review, plus the adoption and release controls that keep a system honest. #### Adoption - [Isn't this just change management with new labels?](https://customlabs.io/adoption/#faq-isn-t-this-just-change-management-with-new-labels) - [We already ran a successful pilot. Why isn't that enough?](https://customlabs.io/adoption/#faq-we-already-ran-a-successful-pilot-why-isn-t-that-enough) - [What's the fastest way to kill a rollout without meaning to?](https://customlabs.io/adoption/#faq-what-s-the-fastest-way-to-kill-a-rollout-without-meaning-to) - [How do we know adoption is real, rather than seat activation with nothing behind it?](https://customlabs.io/adoption/#faq-how-do-we-know-adoption-is-real-rather-than-seat-activation-with-nothing-behind-it) - [Who should own this, the platform team or engineering leadership?](https://customlabs.io/adoption/#faq-who-should-own-this-the-platform-team-or-engineering-leadership) - [Does this still apply once we're past the pilot and scaling org-wide?](https://customlabs.io/adoption/#faq-does-this-still-apply-once-we-re-past-the-pilot-and-scaling-org-wide) #### Governance - [We're not building a high-risk system. Do we need any of this?](https://customlabs.io/governance/#faq-we-re-not-building-a-high-risk-system-do-we-need-any-of-this) - [Isn't this the same thing as the security review?](https://customlabs.io/governance/#faq-isn-t-this-the-same-thing-as-the-security-review) - [Who actually owns AI governance in a 200-person company?](https://customlabs.io/governance/#faq-who-actually-owns-ai-governance-in-a-200-person-company) - [What actually changed with the EU AI Act in 2026?](https://customlabs.io/governance/#faq-what-actually-changed-with-the-eu-ai-act-in-2026) - [Do we need ISO 42001 certification?](https://customlabs.io/governance/#faq-do-we-need-iso-42001-certification) - [What do we actually hand an auditor on day one?](https://customlabs.io/governance/#faq-what-do-we-actually-hand-an-auditor-on-day-one) #### Release - [Isn't this just CI/CD with different names?](https://customlabs.io/release/#faq-isn-t-this-just-ci-cd-with-different-names) - [We just edit the prompt in a dashboard. Do we really need a release process for that?](https://customlabs.io/release/#faq-we-just-edit-the-prompt-in-a-dashboard-do-we-really-need-a-release-process-for-that) - [What's the difference between this and the Operate stage in the handbook?](https://customlabs.io/release/#faq-what-s-the-difference-between-this-and-the-operate-stage-in-the-handbook) - [How do you roll back a retrieval index that already got rebuilt?](https://customlabs.io/release/#faq-how-do-you-roll-back-a-retrieval-index-that-already-got-rebuilt) - [What counts as a severity-1 release incident if the system never actually crashes?](https://customlabs.io/release/#faq-what-counts-as-a-severity-1-release-incident-if-the-system-never-actually-crashes) - [Do we need all six surfaces before we ship anything?](https://customlabs.io/release/#faq-do-we-need-all-six-surfaces-before-we-ship-anything) #### Security review - [Is this legal advice?](https://customlabs.io/security-review/#faq-is-this-legal-advice) - [Do all four gates apply to every AI feature?](https://customlabs.io/security-review/#faq-do-all-four-gates-apply-to-every-ai-feature) - [We're pre-launch, is it too early to prep for this?](https://customlabs.io/security-review/#faq-we-re-pre-launch-is-it-too-early-to-prep-for-this) - [What if we don't have a good answer to one of these yet?](https://customlabs.io/security-review/#faq-what-if-we-don-t-have-a-good-answer-to-one-of-these-yet) - [How is this different from a generic AI security checklist?](https://customlabs.io/security-review/#faq-how-is-this-different-from-a-generic-ai-security-checklist) - [Can a Ship Audit or Readiness & Diligence engagement help with this directly?](https://customlabs.io/security-review/#faq-can-a-ship-audit-or-readiness-diligence-engagement-help-with-this-directly) ### Cost and proof What AI actually costs to run, and the record of what we shipped and what broke. #### Agentic delivery - [Is this just vibe coding at scale?](https://customlabs.io/agentic-delivery/#faq-is-this-just-vibe-coding-at-scale) - [What's the actual bottleneck?](https://customlabs.io/agentic-delivery/#faq-what-s-the-actual-bottleneck) - [Do we need a framework or an orchestrator?](https://customlabs.io/agentic-delivery/#faq-do-we-need-a-framework-or-an-orchestrator) - [How do you stop two agents breaking each other's work?](https://customlabs.io/agentic-delivery/#faq-how-do-you-stop-two-agents-breaking-each-other-s-work) - [Does this hold up on regulated or audited code?](https://customlabs.io/agentic-delivery/#faq-does-this-hold-up-on-regulated-or-audited-code) - [Does it replace engineers?](https://customlabs.io/agentic-delivery/#faq-does-it-replace-engineers) - [How do we start on Monday?](https://customlabs.io/agentic-delivery/#faq-how-do-we-start-on-monday) #### Cost - [Why is our inference bill so much higher than what we estimated?](https://customlabs.io/cost/#faq-why-is-our-inference-bill-so-much-higher-than-what-we-estimated) - [What's the single fastest lever to pull first?](https://customlabs.io/cost/#faq-what-s-the-single-fastest-lever-to-pull-first) - [Do we need all six surfaces on day one?](https://customlabs.io/cost/#faq-do-we-need-all-six-surfaces-on-day-one) - [Isn't cutting tokens just going to hurt quality?](https://customlabs.io/cost/#faq-isn-t-cutting-tokens-just-going-to-hurt-quality) - [How is this different from the AI Cost Calculator?](https://customlabs.io/cost/#faq-how-is-this-different-from-the-ai-cost-calculator) - [Why no per-token price table?](https://customlabs.io/cost/#faq-why-no-per-token-price-table) #### Delivery record - [Did agents write all of it?](https://customlabs.io/delivery-record/#faq-did-agents-write-all-of-it) - [Doesn't this prove agents need heavy supervision?](https://customlabs.io/delivery-record/#faq-doesn-t-this-prove-agents-need-heavy-supervision) - [Why publish your own bugs?](https://customlabs.io/delivery-record/#faq-why-publish-your-own-bugs) - [Can we see the repo?](https://customlabs.io/delivery-record/#faq-can-we-see-the-repo) - [Does this transfer to our codebase?](https://customlabs.io/delivery-record/#faq-does-this-transfer-to-our-codebase) ### Learning the craft The Applied AI Handbook and the pattern library, plus the eval stack and the failure-mode field guide. #### Evals - [How big does a golden set actually need to be?](https://customlabs.io/evals/#faq-how-big-does-a-golden-set-actually-need-to-be) - [Do we need all six layers on day one?](https://customlabs.io/evals/#faq-do-we-need-all-six-layers-on-day-one) - [Can we just use an LLM to judge our own output?](https://customlabs.io/evals/#faq-can-we-just-use-an-llm-to-judge-our-own-output) - [What's the difference between an eval and observability?](https://customlabs.io/evals/#faq-what-s-the-difference-between-an-eval-and-observability) - [Isn't this a lot of infrastructure for a small feature?](https://customlabs.io/evals/#faq-isn-t-this-a-lot-of-infrastructure-for-a-small-feature) - [How is this different from the eval methodology on Capabilities?](https://customlabs.io/evals/#faq-how-is-this-different-from-the-eval-methodology-on-capabilities) #### Failure modes - [How do I know which failure mode I have?](https://customlabs.io/failure-modes/#faq-how-do-i-know-which-failure-mode-i-have) - [Can I diagnose this without an external audit?](https://customlabs.io/failure-modes/#faq-can-i-diagnose-this-without-an-external-audit) - [What does a Ship Audit add?](https://customlabs.io/failure-modes/#faq-what-does-a-ship-audit-add) #### Handbook: Build - [Why is notebook-to-production usually a bigger lift than the prototype?](https://customlabs.io/handbook/build/#faq-why-is-notebook-to-production-usually-a-bigger-lift-than-the-prototype) - [What's the highest-leverage thing to fix first?](https://customlabs.io/handbook/build/#faq-what-s-the-highest-leverage-thing-to-fix-first) - [How much guardrail work do we actually need?](https://customlabs.io/handbook/build/#faq-how-much-guardrail-work-do-we-actually-need) #### Handbook: Cost - [Why is our inference bill so much higher than we estimated?](https://customlabs.io/handbook/cost/#faq-why-is-our-inference-bill-so-much-higher-than-we-estimated) - [What's the fastest lever for cutting AI spend?](https://customlabs.io/handbook/cost/#faq-what-s-the-fastest-lever-for-cutting-ai-spend) - [Why aren't we getting prompt-caching discounts?](https://customlabs.io/handbook/cost/#faq-why-aren-t-we-getting-prompt-caching-discounts) #### Handbook: Decide - [How do we know if a problem is actually an AI problem?](https://customlabs.io/handbook/decide/#faq-how-do-we-know-if-a-problem-is-actually-an-ai-problem) - [What's the single biggest mistake teams make at this stage?](https://customlabs.io/handbook/decide/#faq-what-s-the-single-biggest-mistake-teams-make-at-this-stage) - [Should we use RAG or fine-tune a model?](https://customlabs.io/handbook/decide/#faq-should-we-use-rag-or-fine-tune-a-model) #### Handbook: Design - [Do we need an agent, or is a pipeline enough?](https://customlabs.io/handbook/design/#faq-do-we-need-an-agent-or-is-a-pipeline-enough) - [pgvector or a dedicated vector database?](https://customlabs.io/handbook/design/#faq-pgvector-or-a-dedicated-vector-database) - [Why does model-agnostic architecture matter this early?](https://customlabs.io/handbook/design/#faq-why-does-model-agnostic-architecture-matter-this-early) #### Handbook: Evaluate - [How big does an eval set actually need to be to be useful?](https://customlabs.io/handbook/evaluate/#faq-how-big-does-an-eval-set-actually-need-to-be-to-be-useful) - [Can we just use an LLM to judge our LLM's output?](https://customlabs.io/handbook/evaluate/#faq-can-we-just-use-an-llm-to-judge-our-llm-s-output) - [What's the difference between evals and observability?](https://customlabs.io/handbook/evaluate/#faq-what-s-the-difference-between-evals-and-observability) #### Handbook: Operate - [How do we stop an agent that runs forever?](https://customlabs.io/handbook/operate/#faq-how-do-we-stop-an-agent-that-runs-forever) - [Can a document in our own knowledge base actually hijack our agent?](https://customlabs.io/handbook/operate/#faq-can-a-document-in-our-own-knowledge-base-actually-hijack-our-agent) - [Why does our AI cite a policy we deleted months ago?](https://customlabs.io/handbook/operate/#faq-why-does-our-ai-cite-a-policy-we-deleted-months-ago) #### Patterns - [How is this different from the failure-mode field guide?](https://customlabs.io/patterns/#faq-how-is-this-different-from-the-failure-mode-field-guide) - [Do I need all of these?](https://customlabs.io/patterns/#faq-do-i-need-all-of-these) - [Are these specific to one model provider?](https://customlabs.io/patterns/#faq-are-these-specific-to-one-model-provider) ### From the field Questions a specific insight or comparison raised, answered inside that piece. #### Agents vs Pipelines: When Autonomy Is Worth the Reliability Cost - [Isn't a pipeline just a less impressive version of an agent?](https://customlabs.io/compare/agents-vs-pipelines/#faq-isn-t-a-pipeline-just-a-less-impressive-version-of-an-agent) - [How do I know if my workflow needs an agent?](https://customlabs.io/compare/agents-vs-pipelines/#faq-how-do-i-know-if-my-workflow-needs-an-agent) - [Can a system be part pipeline, part agent?](https://customlabs.io/compare/agents-vs-pipelines/#faq-can-a-system-be-part-pipeline-part-agent) - [What's the minimum bar before shipping agentic autonomy?](https://customlabs.io/compare/agents-vs-pipelines/#faq-what-s-the-minimum-bar-before-shipping-agentic-autonomy) #### Build, Buy, or Skip: A Framework for AI Decisions - [When does it make sense to build an AI feature in-house instead of buying?](https://customlabs.io/insights/build-buy-or-skip/#faq-when-does-it-make-sense-to-build-an-ai-feature-in-house-instead-of-buying) - [Is 'skip' a real option, or just avoiding the decision?](https://customlabs.io/insights/build-buy-or-skip/#faq-is-skip-a-real-option-or-just-avoiding-the-decision) - [What's the most common mistake in a build-vs-buy decision?](https://customlabs.io/insights/build-buy-or-skip/#faq-what-s-the-most-common-mistake-in-a-build-vs-buy-decision) #### Evals Before You Ship: Why AI Features Need Tests Too - [Why do AI features need an eval suite instead of just testing manually?](https://customlabs.io/insights/evals-before-you-ship/#faq-why-do-ai-features-need-an-eval-suite-instead-of-just-testing-manually) - [What should an eval suite actually cover?](https://customlabs.io/insights/evals-before-you-ship/#faq-what-should-an-eval-suite-actually-cover) - [How do you judge AI output that isn't exact-match deterministic?](https://customlabs.io/insights/evals-before-you-ship/#faq-how-do-you-judge-ai-output-that-isn-t-exact-match-deterministic) #### From Notebook to Production: What Actually Breaks - [Why can't a working AI demo just ship to production?](https://customlabs.io/insights/notebook-to-production/#faq-why-can-t-a-working-ai-demo-just-ship-to-production) - [What's the biggest production risk in a RAG feature specifically?](https://customlabs.io/insights/notebook-to-production/#faq-what-s-the-biggest-production-risk-in-a-rag-feature-specifically) - [Does going from notebook to production require a full rewrite?](https://customlabs.io/insights/notebook-to-production/#faq-does-going-from-notebook-to-production-require-a-full-rewrite) #### Model-Agnostic by Design - [What does 'model-agnostic' actually mean in practice?](https://customlabs.io/insights/model-agnostic-by-design/#faq-what-does-model-agnostic-actually-mean-in-practice) - [Why not just hardcode the best model available today?](https://customlabs.io/insights/model-agnostic-by-design/#faq-why-not-just-hardcode-the-best-model-available-today) - [How much extra effort does a model-agnostic design cost?](https://customlabs.io/insights/model-agnostic-by-design/#faq-how-much-extra-effort-does-a-model-agnostic-design-cost) #### Open-Weight Models vs Frontier APIs: The Real Cost/Control Tradeoff - [Is self-hosting an open-weight model actually cheaper?](https://customlabs.io/compare/open-weight-vs-frontier-api/#faq-is-self-hosting-an-open-weight-model-actually-cheaper) - [What's the biggest hidden cost of self-hosting?](https://customlabs.io/compare/open-weight-vs-frontier-api/#faq-what-s-the-biggest-hidden-cost-of-self-hosting) - [Can I switch between the two later without a rewrite?](https://customlabs.io/compare/open-weight-vs-frontier-api/#faq-can-i-switch-between-the-two-later-without-a-rewrite) - [Do open-weight models match frontier API quality now?](https://customlabs.io/compare/open-weight-vs-frontier-api/#faq-do-open-weight-models-match-frontier-api-quality-now) #### Prompt Injection Is a Data Problem: A Threat Model You Can Ship Against - [What is prompt injection?](https://customlabs.io/insights/prompt-injection-threat-model/#faq-what-is-prompt-injection) - [Can you stop prompt injection by filtering the input?](https://customlabs.io/insights/prompt-injection-threat-model/#faq-can-you-stop-prompt-injection-by-filtering-the-input) - [How do you defend an LLM application against prompt injection?](https://customlabs.io/insights/prompt-injection-threat-model/#faq-how-do-you-defend-an-llm-application-against-prompt-injection) #### RAG vs Fine-Tuning: Which One Actually Solves Your Problem - [Can I use RAG and fine-tuning together?](https://customlabs.io/compare/rag-vs-fine-tuning/#faq-can-i-use-rag-and-fine-tuning-together) - [Does fine-tuning reduce hallucination?](https://customlabs.io/compare/rag-vs-fine-tuning/#faq-does-fine-tuning-reduce-hallucination) - [Which is cheaper to run at scale?](https://customlabs.io/compare/rag-vs-fine-tuning/#faq-which-is-cheaper-to-run-at-scale) - [How long does it take to stand up each approach?](https://customlabs.io/compare/rag-vs-fine-tuning/#faq-how-long-does-it-take-to-stand-up-each-approach) #### Vector Database vs pgvector: Do You Actually Need a Dedicated Store - [At what scale does pgvector stop being good enough?](https://customlabs.io/compare/vector-database-vs-pgvector/#faq-at-what-scale-does-pgvector-stop-being-good-enough) - [Is a dedicated vector database always faster?](https://customlabs.io/compare/vector-database-vs-pgvector/#faq-is-a-dedicated-vector-database-always-faster) - [What's the real cost of running a separate vector database?](https://customlabs.io/compare/vector-database-vs-pgvector/#faq-what-s-the-real-cost-of-running-a-separate-vector-database) - [Can I migrate from pgvector to a dedicated store later without a rewrite?](https://customlabs.io/compare/vector-database-vs-pgvector/#faq-can-i-migrate-from-pgvector-to-a-dedicated-store-later-without-a-rewrite) #### What Actually Moved in the EU AI Act's 2026 Deadline Change - [Did the EU AI Act get delayed?](https://customlabs.io/insights/ai-act-high-risk-deadline-moved/#faq-did-the-eu-ai-act-get-delayed) - [Is my system still subject to the Article 5 prohibited-practices rules?](https://customlabs.io/insights/ai-act-high-risk-deadline-moved/#faq-is-my-system-still-subject-to-the-article-5-prohibited-practices-rules) - [What happens to systems already on the market before the transparency deadline?](https://customlabs.io/insights/ai-act-high-risk-deadline-moved/#faq-what-happens-to-systems-already-on-the-market-before-the-transparency-deadline) - [Where can I read the actual text instead of a summary?](https://customlabs.io/insights/ai-act-high-risk-deadline-moved/#faq-where-can-i-read-the-actual-text-instead-of-a-summary) #### What AI Actually Costs in Production, by Workload - [Are these real, measured costs or estimates?](https://customlabs.io/insights/ai-cost-benchmark-by-workload/#faq-are-these-real-measured-costs-or-estimates) - [Why is cost per successful outcome higher than the price on the pricing page?](https://customlabs.io/insights/ai-cost-benchmark-by-workload/#faq-why-is-cost-per-successful-outcome-higher-than-the-price-on-the-pricing-page) - [Which workload is most expensive, and why?](https://customlabs.io/insights/ai-cost-benchmark-by-workload/#faq-which-workload-is-most-expensive-and-why) #### What an AI Feature Actually Costs in Production - [What does an AI feature actually cost in production?](https://customlabs.io/insights/what-ai-actually-costs/#faq-what-does-an-ai-feature-actually-cost-in-production) - [Why do token costs balloon beyond the estimate?](https://customlabs.io/insights/what-ai-actually-costs/#faq-why-do-token-costs-balloon-beyond-the-estimate) - [How do you keep AI cost under control?](https://customlabs.io/insights/what-ai-actually-costs/#faq-how-do-you-keep-ai-cost-under-control) #### Why Your AI Agent Works in the Demo and Stalls in Production - [Why does an AI agent that works in the demo fail in production?](https://customlabs.io/insights/ai-agents-stall-in-production/#faq-why-does-an-ai-agent-that-works-in-the-demo-fail-in-production) - [How many steps can an AI agent chain before reliability drops?](https://customlabs.io/insights/ai-agents-stall-in-production/#faq-how-many-steps-can-an-ai-agent-chain-before-reliability-drops) - [How do you make an AI agent reliable enough to ship?](https://customlabs.io/insights/ai-agents-stall-in-production/#faq-how-do-you-make-an-ai-agent-reliable-enough-to-ship) #### Your RAG Demo Lied to You - [Why does a RAG demo look perfect but fail on a real corpus?](https://customlabs.io/insights/your-rag-demo-lied/#faq-why-does-a-rag-demo-look-perfect-but-fail-on-a-real-corpus) - [What's the most dangerous failure mode in RAG?](https://customlabs.io/insights/your-rag-demo-lied/#faq-what-s-the-most-dangerous-failure-mode-in-rag) - [How do you actually evaluate retrieval quality?](https://customlabs.io/insights/your-rag-demo-lied/#faq-how-do-you-actually-evaluate-retrieval-quality) No question matches that search. Try a different word. Still stuck? [Search the whole site](https://customlabs.io/search/), or [ask us the question directly](https://customlabs.io/contact/).