CTO of One
[ 00 — CTO of One ]

Every page routes to you. Build the backstop with us.

Triage agents, runbooks and review rules built with your team. Our shared 24/7 rotation takes pages only for the systems we build with you.

See a worked example Find your starting point

[ Chapter 01 ]

Billing fails at 3 a.m. The page finds you.

One rotation, one name on it, because only you know how that service really works.

[ Chapter 02 ]

Your memory is the runbook.

Every fix waits on you, every holiday is a risk, and the team cannot grow past what you can carry.

[ Chapter 03 ]

The queue spirals inward.

Agents take the first pass: gather the logs, match the runbook, run the tests and draft a fix for an engineer to review.

[ Chapter 04 ]

Every merge has a name.

An engineer who understands the change approves it. Agents never merge their own work.

[ Chapter 05 ]

Lead the system. Stop being the system.

Picture a 3 a.m. page someone else can answer, because the runbook, the dashboards and the first triage are already there.

[ Chapter 06 ]

Every service mapped. Every agent scoped.

Each agent gets a repo list, a tool list, what it may never touch, and a rule for when it stops and calls a person.

[ 07 — How we work ]

Pair on the code. Keep the know-how.

Teach

Your practice as the technical lead: agent policy, review rules, incident drills and eval habits, learned on your own repositories.

Build

The systems that take the first page: triage agents, runbooks, observability and evals, built beside your engineers so your team can operate them.

[ 08 — Worked example ]

See an agent scope sheet and a triage drill.

A fictional billing service: what the triage agent may touch, where it stops, and who it pages.

Illustrative example · fictional data · not a client result

An agent scope sheet and triage drill for a billing service

Before
Alerts page the CTO directly. The billing runbook lives in one person’s head, and agents use broad tokens from a shared account.
After
A scoped triage agent gathers context and drafts a first diagnosis. The on-call engineer decides, and only the escalation path reaches the CTO.
  1. Scope the agent
  2. Write the runbook
  3. Run a drill
  4. Set the review rule
Triage agent scope sheet: billing service (fictional): 2 of 5 rows
Area
Code
Allowed
Read the billing and invoicing repositories
Never
Push to main or merge its own pull requests
Stop and page when
A fix touches payment-provider code
Area
Logs and metrics
Allowed
Query the last 24 hours of billing logs and dashboards
Never
Export customer data outside the logging tool
Stop and page when
The error rate stays high after the runbook steps

See the full example → Is this a fit? →

[ Chapter 09 ]

Your agents touch other people’s systems.

Egress allowlists, narrow tools and audit trails are engineering work now, reviewed like any other code.

[ Chapter 10 ]

Swap the model. Keep the tests.

A standing eval suite outlasts any single model choice, so an upgrade becomes a test run, not a gamble.

[ 11 — Ways to begin ]

Two ways to clear the queue.

Consult

$400 per hour 5-hour block · $2,000 minimum

Sold in 5-hour blocks

A $2,000 block (five hours) can work toward a written agent policy and review rule, a first runbook from your incident history, or a scope sheet for one agent. We agree the focus first; a block does not take over your on-call. We agree the scope before starting.

Ask about Consult → Details

On Build, we get projects up and running within a week, depending on complexity and how quickly you give us access to your systems.

[ 12 — Your next step ]

Hand off the pager.

What is the stack, who is on the team, and what keeps paging you?

Send a short enquiry, then choose a time to talk or wait for our reply. We agree the scope and price before paid work begins. Sending this form does not book or buy a service.

Which way in?

Thanks. Pick a time below, or watch your inbox for our reply.

Pick a time →

[ Questions ]

Does your rotation take every page?
No. Our shared 24/7 rotation covers only the systems we build with you in Build, with response terms in writing. Everything else stays with your own on-call.
Will agents push to production?
Not in what we build with you. Agents work in branches, open draft pull requests and stop at the gates you set, and an engineer who understands the change approves every merge.
What does a triage drill prove?
That the agent gathers the right context and stops where it should, on a replayed incident in staging. It checks the setup; it is not a promise about the next outage.
What is CTO of One?
CTO of One (ctoofone.ai), published by Utlyze, serves technical leaders at small companies who are the backstop for every system. It covers coding-agent policy, code review, on-call design and architecture for agents.
Who is CTO of One for?
CTOs, heads of engineering and technical co-founders who are accountable for all of the systems and are still the person everyone calls when production breaks.
Should my team use AI coding agents?
Most teams can, with written rules: which repositories agents may touch, which tests must pass, how secrets are kept out of reach, and a human approval on every merge. Start with one repository and widen from there.
Who reviews AI-written code?
An engineer who understands the change approves every merge. Agents can do a first pass, such as running tests, flagging risky diffs and summarizing what changed, but the approval stays with a named person.
Does AI-written code create technical debt?
It can, like any code that ships faster than the team understands it. The usual guards are the same ones good teams already use: tests, small diffs, clear ownership and review by someone who could have written the change.
How do I stop being the only one who can fix production?
Write down what only you know as runbooks, make the systems observable enough that someone else can see what broke, and let agents take first-line triage under clear escalation rules. Then practice handoffs before the next incident.
What is a fractional CTO, and is this one?
A fractional CTO is a part-time technical executive who leads technology for one or more companies. Consult gives you hours with our coaches on specific technical decisions; it supports your technical leadership rather than replacing it.
How much are Consult and Build?
Consult: $400 per hour, sold five hours at a time, so the minimum is $2,000. Build: $15,000 a month. Each month, pick up to 5 business processes for our build-with-you coaches to work on. They build your AI systems with you and teach your team to run them. If you ever leave, the systems stay with you. Agents work only on your systems, and a shared 24/7 rotation answers pages within response times set in writing. Build has a three-month minimum: $45,000 at the base level. Higher tiers and separately quoted work cost more.
How is this different from Business of One?
Business of One serves owners building an AI-first company. CTO of One is for the person accountable for the technology: agent policy, reliability and architecture. Engineer of One, a sister site, is for the person writing the code.
Is this security, legal or employment advice?
No. What we publish on security, law, employment, tax or investing is for learning only. For your own decisions, use qualified advisers who have seen your systems and your company.
What happens after I enquire, and what is the minimum commitment?
Send a short description of your work, then choose a time to talk or wait for our reply. The enquiry does not buy a service. We agree scope and price before paid work. Consult: 5-hour block · $2,000 minimum. Build: Three-month minimum · $45,000 minimum commitment. Build Plus is $25,000 a month ($75,000 minimum); Build Program is $50,000 a month ($150,000 minimum). New software products and core-system replacement are quoted separately.

Built on the Of One method: ask the right questions, map what you find, make the best move. of1.ai ↗