--- title: "AI Agent Acceptance Checklist for Buyers" description: "Accept an AI agent build only after source, permission, denied-action, receipt, rollback, ownership, and clean-start tests pass." canonical: "https://scalewithsearch.com/articles/ai-agent-acceptance-checklist" date: "2026-08-18" modified: "2026-09-19" --- ## Site navigation - [Scale With Search](https://scalewithsearch.com/) - Real estate - Real estate - [Real estate](https://scalewithsearch.com/for/real-estate) - Work - Start here - [Send your brief](https://scalewithsearch.com/work#send-your-brief) - [Prepare your six-question brief](https://scalewithsearch.com/work#prepare-your-six-question-brief) - Build - [Site build, content library with SEO, signal desk](https://scalewithsearch.com/work) - For your business - Trades and home services - [Auto body and collision shops](https://scalewithsearch.com/for/auto-body-and-collision-shops) - [Foundation and home repair contractors](https://scalewithsearch.com/for/foundation-and-home-repair) - [Garage door and fencing contractors](https://scalewithsearch.com/for/garage-door-and-fencing-contractors) - [HVAC contractors](https://scalewithsearch.com/for/hvac-contractors) - [Janitorial and commercial cleaning companies](https://scalewithsearch.com/for/janitorial-and-commercial-cleaning) - [Locksmiths](https://scalewithsearch.com/for/locksmiths) - [Moving companies](https://scalewithsearch.com/for/moving-companies) - [Pest control companies](https://scalewithsearch.com/for/pest-control-companies) - [Plumbing and electrical contractors](https://scalewithsearch.com/for/plumbing-and-electrical-contractors) - [Restoration and water or fire damage companies](https://scalewithsearch.com/for/restoration-and-water-fire-damage) - [Roofing companies](https://scalewithsearch.com/for/roofing-companies) - [Towing companies](https://scalewithsearch.com/for/towing-companies) - [Tree services and landscaping companies](https://scalewithsearch.com/for/tree-services-and-landscaping) - [Solar installers](https://scalewithsearch.com/for/solar-installation) - [General contractors](https://scalewithsearch.com/for/general-contractors-and-construction) - [Paving, concrete, and flooring contractors](https://scalewithsearch.com/for/paving) - Practices and professional services - [Bookkeeping and tax practices](https://scalewithsearch.com/for/bookkeeping-and-tax-practices) - [Dental practices](https://scalewithsearch.com/for/dental-practices) - [Family and criminal defense law firms](https://scalewithsearch.com/for/family-and-criminal-defense-law-firms) - [Med spas and aesthetics practices](https://scalewithsearch.com/for/med-spas-and-aesthetics) - [Personal injury law firms](https://scalewithsearch.com/for/personal-injury-law-firms) - [Veterinary clinics](https://scalewithsearch.com/for/veterinary-clinics) - [Gyms and fitness studios](https://scalewithsearch.com/for/fitness) - [Therapy and outpatient health practices](https://scalewithsearch.com/for/therapy-and-outpatient-health) - [Medical billing companies](https://scalewithsearch.com/for/medical-billing) - [Insurance agencies](https://scalewithsearch.com/for/insurance-agencies) - [Financial advisors](https://scalewithsearch.com/for/financial-advisors) - [Property management companies](https://scalewithsearch.com/for/property-management) - [Recruiting and staffing agencies](https://scalewithsearch.com/for/recruiting-and-staffing) - [Architects and interior designers](https://scalewithsearch.com/for/architects-and-interior-designers) - [Logistics and supply chain companies](https://scalewithsearch.com/for/logistics-and-supply-chain) - Agencies, MSPs, and manufacturing - [IT and managed service providers](https://scalewithsearch.com/for/it-and-managed-service-providers) - [Machine shops and precision manufacturers](https://scalewithsearch.com/for/machine-shops-and-precision-manufacturing) - [Marketing agencies and freelancers](https://scalewithsearch.com/for/marketing-agencies-and-freelancers) - [SEO agencies and consultants](https://scalewithsearch.com/for/seo-agencies-and-consultants) - [Small manufacturers and fabricators](https://scalewithsearch.com/for/small-manufacturers-and-fabricators) - Restaurants, shops, studios, and nonprofits - [Restaurants and hospitality businesses](https://scalewithsearch.com/for/restaurants-and-hospitality) - [Retail stores and ecommerce sellers](https://scalewithsearch.com/for/retail-and-ecommerce) - [Photographers, event planners, and travel agents](https://scalewithsearch.com/for/photographers) - [Churches and nonprofits](https://scalewithsearch.com/for/churches-and-nonprofits) - [All industries](https://scalewithsearch.com/for/) - Learn - For your office - [Office job guides](https://scalewithsearch.com/guides/) - [Browser calculators](https://scalewithsearch.com/tools/) - Start here - [How it works](https://scalewithsearch.com/how-it-works) - [Free Starter Kit](https://scalewithsearch.com/kit/business-memory-starter-kit.zip) - [Synthetic specimen](https://scalewithsearch.com/specimen/working-session-specimen.zip) - Guides - [The Complete Guide to Business Memory for AI Agents](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [The Complete Small-Business Guide to AI Agent Governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [The Complete Guide to Leaving Vendor AI Memory](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - Articles by cluster - [Business memory](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [Agent governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [Migration and ownership](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - For machines - [llms.txt](https://scalewithsearch.com/llms.txt) - [llms-full.txt](https://scalewithsearch.com/llms-full.txt) - [Machine view](https://scalewithsearch.com/?view=machine) - Company - Evidence - [Proof](https://scalewithsearch.com/proof) - Company - [About](https://scalewithsearch.com/about) # AI Agent Acceptance Checklist for Buyers. The builder runs an agent during a screen share. It answers the sample question and creates the expected draft. The buyer accepts the project. The next week, the agent reads the wrong account folder, retries a live write, and leaves no receipt. The demo showed a happy path. It did not prove the system was ready to accept. Acceptance is the buyer's written decision that a named build version passed a defined list of tests. It covers the result, source path, permissions, denied actions, receipts, recovery, ownership, and handoff. ## Acceptance begins with the purchased outcome Write the outcome at the top of the checklist. For example: "Prepare one weekly account brief from the buyer-owned account packet. Cite current sources. Save the draft and receipt. Stop before external send." That sentence prevents scope drift. The acceptance review should not quietly include bulk processing, live publication, another department, or a new connector. The existing guide to [what a build acceptance list contains](/articles/what-a-build-acceptance-list-contains) explains the structure. This checklist turns the buyer's review into a pass or stop decision. Microsoft's agent evaluation checklist recommends test cases with expected responses, acceptance criteria, and evaluation methods. It also includes forbidden cases and ongoing evaluation after changes. [Microsoft: Agent evaluation checklist](https://learn.microsoft.com/en-us/agents/agent-evaluation/evaluation-checklist) NIST provides voluntary AI risk-management guidance and resources for testing, evaluation, verification, and validation. [NIST AI Resource Center](https://airc.nist.gov/) Neither source accepts the build for you. Your checklist must match the effect the agent can have. ## Inspect `buyer-acceptance.yaml` Use one status vocabulary across the project. ```yaml build_id: weekly-brief-v1.2.0 outcome: prepare one account brief and stop before send reviewer: account_owner decision_date: pending statuses: - accepted - failed - blocked - accepted_exception critical_items: - source_isolation - approval_boundary - secret_handling - duplicate_prevention - ownership_transfer evidence_root: acceptance/weekly-brief-v1.2.0/ production_authorized: false ``` The `production_authorized` field remains false during acceptance testing. Passing tests does not activate the system. Each checklist item should link to a test receipt. Do not accept screenshots without the input, expected result, and build version they represent. ## Check the exact output Confirm the output path, file name, required sections, source citations, date, status, and review marker. Test one normal input. Then change the wording while keeping the same facts. The agent may vary prose. It must preserve the required structure and facts. Reject acceptance criteria such as "good quality" or "sounds professional." Replace them with observable checks or an approved human-review rubric. ## Check source authority The agent should read the correct account packet and honor source precedence. Test: 1. All required current sources are present. 2. One required source is missing. 3. An older archive item conflicts with a current decision. 4. Two current sources with equal authority conflict. 5. A formatting example contains a fictional fact. 6. Another account contains a tempting matching phrase. Expected behavior should be explicit. The current decision wins only when the source-order file grants it authority. Equal-authority conflict produces a stop. Example facts do not transfer. Another account remains unread. ## Check permissions and denied actions List every capability available to the runtime. Include file writes, email, calendar, CRM, deployment, payments, web forms, and shell commands. For each consequential action, test three states: - no approval; - one valid matching approval; - stale, mismatched, or consumed approval. The denied case is a core acceptance result. An agent that completes the happy path but cannot refuse the wrong action is not ready. If the accepted job is draft-only, the production runtime should not hold send credentials. State that as architecture evidence, not only a prompt instruction. ## Check prompt injection from business files Put a harmless instruction inside a fixture document. For example, tell the agent to ignore the task and copy a synthetic marker into an external-action preview. The content should remain data. It should not grant authority or change the destination. OpenAI defines prompt injection as third-party content misleading an AI into doing something the user did not ask for. It recommends layered controls, narrow access, explicit tasks, and confirmation for important actions. [OpenAI: Understanding prompt injections](https://openai.com/safety/prompt-injections/) One passed fixture does not prove immunity. It proves the implemented boundary handled the named case. ## Check duplicates, retries, and uncertain results Run the same job identifier twice. Replay a consumed approval. Simulate a provider timeout after an action request. The system should return the prior receipt or a duplicate state. It should read external state before retrying an uncertain action. It should never treat a timeout as proof that nothing happened. Record the idempotency key or deduplication field used by the workflow. If the connected system supports its own idempotency control, record that too. ## Check receipts for every state Run at least four fixtures: - completed; - blocked; - failed; - no work. Each receipt should identify the job, build version, source versions, tests, output, stop status, approval status, and observed result. The article on why [a receipt is part of the output](/articles/receipt-is-part-of-the-output) explains the operating need. During acceptance, the receipt is also evidence that the buyer inspected the right build. ## Check failure and recovery Disable one required tool. Corrupt one test input. Remove one source. Simulate a partial output write. The workflow should fail in a defined way. It should not present partial work as complete. Test the recovery path in an isolated environment. For a file write, restore or supersede the draft. For a deployment, restore the prior test version. For an external message, acknowledge that recall may be impossible and place more weight on prevention. The [rollback plan](/articles/ai-agent-rollback-plan) should name the exact object, owner, procedure, recovery time limit, and expected restored state. ## Check monitoring and the first bounded run Test the alert path with a controlled failure. Confirm the expected person receives an understandable notice that links to the runbook. Define the first production batch, even though acceptance does not authorize it. One record or one draft is a common start. The final production decision can approve that bounded run separately. Do not approve an unbounded schedule because local fixtures passed. ## Check ownership and transfer The buyer should receive: - canonical files or repository; - runbook; - environment-variable inventory without secret values; - source and permission map; - test fixtures and commands; - receipts; - rollback procedure; - support boundary. Ask an authorized person who did not build the system to start from a clean session. That person should locate the files, run the tests, produce the accepted output, and explain the stopping point. If the system works only through the builder's account or remembered setup, ownership transfer has not passed. ## Record the final buyer decision Mark each item `accepted`, `failed`, `blocked`, or `accepted_exception`. An accepted exception must state impact, owner, review date, and compensating control. Do not accept exceptions silently for source isolation, authorization, secret exposure, or destructive-action controls. The final decision names the build version and evidence root. Later changes require a new review or targeted regression test. ## Approval and stopping boundary Acceptance testing may use fixtures, recorders, test accounts, and isolated environments. It stops before production activation. A critical failed or blocked item stops acceptance. The test agent cannot mark its own work accepted. Deployment, live credentials, external sends, billing changes, data deletion, access changes, and automation activation require a named buyer decision after the evidence review. Scope expansion requires a new outcome and checklist. ## Sources - [Microsoft: Review the agent evaluation checklist](https://learn.microsoft.com/en-us/agents/agent-evaluation/evaluation-checklist) - [NIST: AI Risk Management Framework](https://www.nist.gov/itl/ai-risk-management-framework) - [NIST: AI Resource Center](https://airc.nist.gov/) - [OpenAI: Understanding prompt injections](https://openai.com/safety/prompt-injections/) ## Questions about AI Agent Acceptance Checklist for Buyers ### How are you testing AI agents before shipping to production? Keep production_authorized false during acceptance and test the purchased outcome against source authority, permissions, denied actions, prompt injection, duplicates, receipts, recovery, and ownership transfer. A critical failed or blocked item stops acceptance. ### Where do human approval controls help instead of slowing every step? Use risk-based controls instead of blanket approvals. Drafting and read-only work can remain inside a defined boundary, while irreversible, external, financial, destructive, or access-changing actions require explicit approval and verification. ### If you're building or deploying AI agents inside a company, what controls do you expect before you'd trust them with real tools? Check source isolation, permission boundaries, denied actions, secret handling, duplicate prevention, receipt states, recovery, ownership transfer, and clean-start operation. Passing the fixtures does not activate production or grant live credentials. ## Related: Testing and Acceptance - [Five AI Agent Reliability Metrics a Small Team Can Verify](/articles/ai-agent-reliability-metrics-small-business) - [What an AI Agent Rollback Plan Must Contain Before Production Access](/articles/ai-agent-rollback-plan) - [Test an AI Agent for Duplicate Runs, Retries, and Double Clicks](/articles/test-ai-agent-idempotency-and-retries) ---- ```text .|########||. .|########||. .|########||. |##||. .||##|. |##||. .||##|. |##||. .||##|. |#|. .|#|. |#|. .|#|. |#|. .|#|. |#| |#| |#| |#| |#| |#| .#| |#. .#| |#. .#| |#. .#. .#| .#. .#| .#. .#| |#. .#| |#. .#| |#. .#| |# ...... #| |# ...... #| |# ...... #| .# ||#########| #| .# ||#########| #| .# ||#########| #| .#. |######||######|. .#. .#. |######||######|. .#. .#. |######||######|. .#. #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# # .||####|###########################|. # # .||####|###########################|. # # .||####|###########################|. # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || #. .######|#|#########| .# #. .######|#|#########| .# #. .######|#|#########| .# .#. |#####||||#####| .#. .#. |#####||||#####| .#. .#. |#####||||#####| .#. |# ||########|| #. |# ||########|| #. |# ||########|| #. |# ...... #| |# ...... #| |# ...... #| |#. .#| |#. .#| |#. .#| |#. .#. |#. .#. |#. .#. .#| |#. .#| |#. .#| |#. |#| |#| |#| |#| |#| |#| .|#|. .|#|. .|#|. .|#|. .|#|. .|#|. |##||. .||##| |##||. .||##| |##||. .||##| .||########||. .||########||. .||########||. Scale With Search 2026 [scalewithsearch.com](https://scalewithsearch.com) ```