Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for ctos

Architecture Decision Review Agent

Every significant architecture decision reviewed and recorded

Architecture Decision Review Agent: what goes in, what the agent does and what you get

What it does

Engineering teams make architecture decisions that affect cost and security for years, yet proposals often arrive missing key parts. When a proposal is submitted, this agent reads it with the linked designs and compares it with company standards and earlier decision records. It checks that the proposal is complete and in line with standards. If security, cost or another section is missing, it asks the author to fill the gap and rechecks. It then estimates cost, risk and effort for each option and summarizes the trade-offs on one page. Before the review, it checks that every trade-off cites a source or estimate, and goes back to fill any that do not. The CTO or architecture board makes the decision, and any exception to standards needs explicit approval. Edge case: a proposal that reverses an earlier decision is linked to it, with affected systems listed.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueYes, continueApprovedNoNo 1 STARTS WHEN Architecture proposal submitted 2 USES A TOOL Read the proposal and linked designs 3 USES A TOOL Compare with standards and previous decision records 4 CHECKS THE RESULT Is the proposal complete and in line with standards? If not: list missing sections and conflicts, ask theauthor to revise, then reread. Back to step 2. 5 DOES Estimate cost, risk and effort for each option 6 DOES Summarize options and trade-offs on one page 7 CHECKS THE RESULT Does every trade-off cite a source or estimate? If not: fill gaps from the author or mark them asassumptions. Back to step 5. 8 YOU APPROVE CTO or architecture board decides 9 RESULT Decision record filed and linked
Read the steps as a list
  1. Architecture proposal submitted
  2. Read the proposal and linked designs
  3. Compare with standards and previous decision records
  4. Is the proposal complete and in line with standards?If not: list missing sections and conflicts, ask the author to revise, then reread. Back to step 2.
  5. Estimate cost, risk and effort for each option
  6. Summarize options and trade-offs on one page
  7. Does every trade-off cite a source or estimate?If not: fill gaps from the author or mark them as assumptions. Back to step 5.
  8. CTO or architecture board decidesThe agent waits here for your OK.
  9. Decision record filed and linked

How it decides

It checks proposals against standards and history and requires missing sections before review.

  • Security and cost sections required
  • Reversals linked to earlier decisions
  • Exceptions to standards need explicit approval

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Standards
  • Required sections
  • Review cadence
  • Log format

What keeps you in control

It always asks you first

  • Architecture decisions
  • Exceptions

Hard limits

  • Never merges code
  • No decisions without board

It stops when

  • Done: recorded
  • Stop: proposal withdrawn

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensOn May 6 a team submits a proposal to add a message queue for order events. The agent compares it with 42 past decisions and finds no data retention section and a conflict with the 2025 rule on managed services. It asks the author to revise. The second version passes. The agent estimates $1,200 a month for option A and $300 for option B, writes the summary, and the CTO picks option B on May 9.

More agents for ctos