CTO of One
[ ← CTO of One ]

Buyer preparation guide

Put the pager in writing.

For the CTO who still carries the pager: which systems a rotation would cover, how severity and response are defined, who is reachable when, and what has been tested before anyone relies on it.

Choose your questions ↓

A preparation aid, not an assurance pack. This guide lists what to ask, who answers, and the record that would settle each question. It contains none of those records. Provider documents and written terms still need review by their owners on your side, and finishing the list is not an approval, a certification or a go-ahead for live data.

Your request

17 of 17 questions included

Every question starts included, because an unknown is not an answer. Leave one out only when you do not need to ask it this time.

Choose questions ↓

All 17 questions are listed below, followed by a plain-text version you can copy or print. Choosing questions, copying and downloading need JavaScript.

Who reads what

CEO or budget holder
Stage 1 coverage, and the minimum commitment.
Engineers on call
Severity, escalation, runbooks and handback, across all three stages.
Access approver
Production access, agent scope and the merge rule.

Minimum commitment

Consult: $2,000 for a 5-hour block. Build: $45,000 minimum over three months ($15,000 a month). Build Plus: $75,000 minimum. Build Program: $150,000 minimum. Levels and fit

What this site lists

Package terms as published. Confirm each one in writing: none of them is evidence that a control or document exists.

  • Build lists a shared 24/7 rotation that covers only the systems built with you, with response terms in writing. Everything else stays with your own on-call. FAQ
  • Build Program ($50,000 a month, $150,000 minimum over three months) lists a named person on call around the clock as your backup. Build levels
  • Agents work in branches, open draft pull requests and stop at the gates you set; an engineer who understands the change approves every merge. FAQ
  • A Consult block does not need production access. Consult
  • Services touching payment, health or student records are listed in Build Plus ($25,000 a month, $75,000 minimum over three months). Fit & price
  • Scope and price are agreed before paid work begins, and an enquiry does not book or buy a service. Fit & price

Not on this page

  • Response figures for any severity or stage
  • Severity definitions, coverage hours or a rota
  • Named contacts or an escalation path
  • Exclusions or dependency ownership
  • Runbooks, restore tests or drill records

Ask for each by name. If a record does not exist yet, “not recorded” is an answer too: write it down.

The questions, in three stages

Each one names who answers it, who on your side checks the answer, and the record that would settle it.

A ticked box puts the question in your request. Leaving one out does not mark it answered: this page records no answers.

17 of 17 questions included

Define the bounded work

List exactly what a rotation would cover. Anything not on the list stays with your own on-call.

  1. 1.1

    Which services, repositories and environments would the rotation cover, listed by name?

    Answered by
    Both sides
    Checked by
    Engineering lead
    Settled by
    A coverage list marking each service and environment in or out.
  2. 1.2

    What stays with your own on-call, and how does alert routing keep that boundary?

    Answered by
    Both sides
    Checked by
    Engineering lead
    Settled by
    An out-of-scope list and the alert routing that enforces it.
  3. 1.3

    How is each severity level defined, with an example incident for each?

    Answered by
    Both sides
    Checked by
    Engineers on call
    Settled by
    Written severity definitions agreed by both sides.
  4. 1.4

    What may each agent read, run or change, and what must it never touch?

    Answered by
    Both sides
    Checked by
    Engineering lead
    Settled by
    An agent scope sheet you approve, one per agent.
  5. 1.5

    Which production access would the work need, and who approves and removes each grant?

    Answered by
    Your team
    Checked by
    Access approver
    Settled by
    An access list with an approver and an expiry for each grant.

Review the evidence and agree terms

Get each response commitment separately and in writing. “Best effort” and “not promised” are valid answers when they are written down.

  1. 2.1

    For each severity, what is committed separately for detection, acknowledgment, response, workaround and restoration?

    Acknowledging a page is not restoring the service. The site says response terms are set in writing; take the figures from those terms, not from this page.

    Answered by
    Provider
    Checked by
    Engineering lead
    Settled by
    Written response terms for each severity and each stage.
  2. 2.2

    What are the coverage hours and time zone, and how is the rota staffed overnight and on holidays?

    Answered by
    Provider
    Checked by
    Engineering lead
    Settled by
    A role-based rota or coverage arrangement for the covered systems.
  3. 2.3

    Who are the named contacts on each side, in what escalation order, through which channels?

    Answered by
    Both sides
    Checked by
    Engineers on call
    Settled by
    A current escalation path: roles, order and channels.
  4. 2.4

    What is excluded, such as third-party outages, changes your team ships or untested deploys?

    Answered by
    Provider
    Checked by
    Engineering lead
    Settled by
    The exclusions list in the support terms.
  5. 2.5

    Who owns each dependency, such as the cloud host, payment processor, model API and alerting tool?

    Answered by
    Both sides
    Checked by
    Engineering lead
    Settled by
    A dependency map with an owner and an escalation route for each.
  6. 2.6

    If the person on rotation is unavailable, who takes over, and is a named backup part of your level?

    A named person on call around the clock as your backup is listed in Build Program ($50,000 a month, $150,000 minimum over three months).

    Answered by
    Provider
    Checked by
    CEO or budget holder
    Settled by
    The backup arrangement in writing.

Accept and operate

Before anyone relies on the rotation, see it work on a replayed incident and a tested restore.

  1. 3.1

    Is there a runbook for each covered service, reviewed by an engineer who would follow it at night?

    Answered by
    Both sides
    Checked by
    Engineers on call
    Settled by
    Runbooks in your repository, each with a review date.
  2. 3.2

    Has a restore or rollback been tested in staging, with timestamps and the result recorded?

    If the provider supports other systems, ask for a redacted record from one of them too.

    Answered by
    Both sides
    Checked by
    Engineering lead
    Settled by
    A dated restore or rollback test record, including what failed.
  3. 3.3

    Has a replayed incident shown the triage agent stopping and paging at the right point?

    Answered by
    Both sides
    Checked by
    Engineers on call
    Settled by
    A drill record from staging.
  4. 3.4

    Who approves agent-written changes before merge, and does the repository enforce it?

    Answered by
    Your team
    Checked by
    Engineering lead
    Settled by
    Branch protection settings and the written review rule.
  5. 3.5

    How is each incident handed back to your team, and who owns the follow-up actions?

    Answered by
    Both sides
    Checked by
    Engineers on call
    Settled by
    A handover note template and one example from a drill.
  6. 3.6

    If the engagement ends, how and when do pages route back to your own on-call?

    Answered by
    Both sides
    Checked by
    Engineering lead
    Settled by
    A dated cut-over plan for alert routing.

17 of 17 questions included

The full request as plain text
CTO of One: questions to settle before live work
Guide: https://ctoofone.ai/readiness/
Prepared on: _________________________________
Prepared by: _________________________________

PREPARATION AID, NOT ASSURANCE
- This is a preparation aid: the questions still to answer, who answers each
  one, and the record that would settle it.
- It is not an assurance pack, a certification, a security approval or legal
  advice. A question left out of this list is not answered or approved.
- Provider records and written terms still need review by the owner named on
  your side. Finishing this list does not make it safe to release live data.

HOW TO FILL THE BLANKS
Under each question, note the document that answers it, its date or version,
and the person on your side who checked it. Leave a blank empty until a real
record exists. "Not recorded" and "not offered" are answers too.

QUESTIONS IN THIS REQUEST: 17 OF 17

STAGE 1. DEFINE THE BOUNDED WORK

1.1 Which services, repositories and environments would the rotation cover,
    listed by name?
    Answered by: Both sides. Checked by: Engineering lead.
    Settled by: A coverage list marking each service and environment in or
    out.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

1.2 What stays with your own on-call, and how does alert routing keep that
    boundary?
    Answered by: Both sides. Checked by: Engineering lead.
    Settled by: An out-of-scope list and the alert routing that enforces it.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

1.3 How is each severity level defined, with an example incident for each?
    Answered by: Both sides. Checked by: Engineers on call.
    Settled by: Written severity definitions agreed by both sides.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

1.4 What may each agent read, run or change, and what must it never touch?
    Answered by: Both sides. Checked by: Engineering lead.
    Settled by: An agent scope sheet you approve, one per agent.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

1.5 Which production access would the work need, and who approves and
    removes each grant?
    Answered by: Your team. Checked by: Access approver.
    Settled by: An access list with an approver and an expiry for each
    grant.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

STAGE 2. REVIEW THE EVIDENCE AND AGREE TERMS

2.1 For each severity, what is committed separately for detection,
    acknowledgment, response, workaround and restoration?
    Answered by: Provider. Checked by: Engineering lead.
    Settled by: Written response terms for each severity and each stage.
    Note: Acknowledging a page is not restoring the service. The site says
    response terms are set in writing; take the figures from those terms,
    not from this page.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

2.2 What are the coverage hours and time zone, and how is the rota staffed
    overnight and on holidays?
    Answered by: Provider. Checked by: Engineering lead.
    Settled by: A role-based rota or coverage arrangement for the covered
    systems.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

2.3 Who are the named contacts on each side, in what escalation order,
    through which channels?
    Answered by: Both sides. Checked by: Engineers on call.
    Settled by: A current escalation path: roles, order and channels.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

2.4 What is excluded, such as third-party outages, changes your team ships
    or untested deploys?
    Answered by: Provider. Checked by: Engineering lead.
    Settled by: The exclusions list in the support terms.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

2.5 Who owns each dependency, such as the cloud host, payment processor,
    model API and alerting tool?
    Answered by: Both sides. Checked by: Engineering lead.
    Settled by: A dependency map with an owner and an escalation route for
    each.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

2.6 If the person on rotation is unavailable, who takes over, and is a named
    backup part of your level?
    Answered by: Provider. Checked by: CEO or budget holder.
    Settled by: The backup arrangement in writing.
    Note: A named person on call around the clock as your backup is listed
    in Build Program ($50,000 a month, $150,000 minimum over three months).
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

STAGE 3. ACCEPT AND OPERATE

3.1 Is there a runbook for each covered service, reviewed by an engineer who
    would follow it at night?
    Answered by: Both sides. Checked by: Engineers on call.
    Settled by: Runbooks in your repository, each with a review date.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

3.2 Has a restore or rollback been tested in staging, with timestamps and
    the result recorded?
    Answered by: Both sides. Checked by: Engineering lead.
    Settled by: A dated restore or rollback test record, including what
    failed.
    Note: If the provider supports other systems, ask for a redacted record
    from one of them too.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

3.3 Has a replayed incident shown the triage agent stopping and paging at
    the right point?
    Answered by: Both sides. Checked by: Engineers on call.
    Settled by: A drill record from staging.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

3.4 Who approves agent-written changes before merge, and does the repository
    enforce it?
    Answered by: Your team. Checked by: Engineering lead.
    Settled by: Branch protection settings and the written review rule.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

3.5 How is each incident handed back to your team, and who owns the
    follow-up actions?
    Answered by: Both sides. Checked by: Engineers on call.
    Settled by: A handover note template and one example from a drill.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

3.6 If the engagement ends, how and when do pages route back to your own
    on-call?
    Answered by: Both sides. Checked by: Engineering lead.
    Settled by: A dated cut-over plan for alert routing.
    Document: ____________________________________
    Date or version: _____________________________
    Owner: _______________________________________

LISTED ON THIS SITE
Package terms as published, to confirm in writing. None of them is evidence
that a control or document exists.
- Build lists a shared 24/7 rotation that covers only the systems built with
  you, with response terms in writing. Everything else stays with your own
  on-call. (https://ctoofone.ai/faq/)
- Build Program ($50,000 a month, $150,000 minimum over three months) lists
  a named person on call around the clock as your backup.
  (https://ctoofone.ai/build/#levels)
- Agents work in branches, open draft pull requests and stop at the gates
  you set; an engineer who understands the change approves every merge.
  (https://ctoofone.ai/faq/)
- A Consult block does not need production access.
  (https://ctoofone.ai/consult/)
- Services touching payment, health or student records are listed in Build
  Plus ($25,000 a month, $75,000 minimum over three months).
  (https://ctoofone.ai/fit/)
- Scope and price are agreed before paid work begins, and an enquiry does
  not book or buy a service. (https://ctoofone.ai/fit/)

MINIMUM COMMITMENT
Consult: $2,000 for a 5-hour block. Build: $45,000 minimum over three months
($15,000 a month). Build Plus: $75,000 minimum. Build Program: $150,000
minimum.

LIMITS
- These are scoping questions, not legal advice. Ask counsel what applies
  where you operate.
- Keep records, credentials and personal data out of the enquiry form. Agree
  in writing how documents will be exchanged.
- The worked examples on this site use fictional organizations and data.
  They are not client results.
- The shared rotation is not a guaranteed resolution time, and it does not
  take over pages for systems outside the written coverage.

NEXT STEP
To talk it through: https://ctoofone.ai/start/?path=consult
An enquiry does not book or buy a service.

When the questions are ready, talk them through with us.

Ask about a Consult block → Is it a fit?

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. Keep records, credentials and personal data out of the form.