Home / Services / AI System Continuity

Keep it running.AI System Continuity.

You shipped production AI. This keeps it running. 2 to 3 days a month on the system you already have: evals re-run against the baseline, prompts fixed when they break, vendor issues handled, and one small scope addition.

Book 30 minutes with Edwin

No published price. I put the number in writing within 48 hours of the call.

PriceIn writingWithin 48 hours of the call
Term2 to 3 daysA month, scope-locked
Minimum3 monthsMonthly renewal after
Credit$2,500 offDiagnostic fee, within 30 days

01 · What breaks

Production AIdecays quietly.

Nothing announces itself. It shows up as a user-visible failure, a leadership question you cannot answer, or a cost line nobody traced back to its source. You do not need another build. You need someone watching what you already shipped, on a schedule, against a baseline.

The model changes under you
Same prompt, different answers. Accuracy slides a few points a quarter. The eval suite would catch it if anyone were running it.
The vendor retires the model
A 90-day deprecation window and no migration plan. The replacement handles tool calls differently and the evals fail on day one.
Upstream data changes shape
A CRM field gets renamed. The agent keeps running, returns plausible answers, and quietly corrupts the next step in the workflow.
The evals stop running
No owner and no schedule. By month four the eval suite is a folder nobody opens, and regressions reach users unflagged.
Access outlives the org chart
Roles shift, permissions do not, and the AI inherits access nobody meant to grant.

02 · What's included

What shipped,kept running.

2 to 3 days a month, scope-locked. Each item has a cadence and ends up in the monthly report.

Eval monitoring. The eval suite runs against the baseline every month, same examples, same scoring. Drift gets flagged with a number.
Prompt triage. When a prompt breaks, I reproduce it in the evals, ship the fix, and re-baseline. No tickets for your engineers.
Vendor escalations and deprecation watch. Quota issues, billing surprises, rate limits, and model retirements come to me. The migration plan is drafted before the deprecation window opens.
One small scope addition a month. A new prompt or a small workflow change, sized to the day budget and held to the same eval bar as the original build.
A one-page monthly report. Accuracy, latency, cost, and errors against the baseline, plus what broke and what got fixed. Written so you can forward it without a redraft.
A quarterly refresh. The layer your team and your AI both read from gets updated, and who and what the AI can touch gets checked against the current org chart.

03 · Not included

Scope-lockedon purpose.

When the work grows, it moves to the engagement built for it. The retainer does not grow past 3 days a month.

New workflows

Scoped AI Build

One small addition a month fits. A new multi-step workflow does not. That is a build, priced per workflow from $7,500.

Owning the program

Fractional Head of AI Transformation

Sequencing new work, vendor strategy, hiring an AI lead, and leadership updates on the whole program. That is the Fractional retainer.

Budget and roadmap

AI Portfolio Allocation Review

Deciding where the next AI budget goes is its own 3 to 4 week module.

New training rollouts

AI Workforce Training

Continuity updates the team's material when a model or workflow changes. A full rollout is its own 4 to 8 week module.

04 · Price and terms

Priced in writing.Within 48 hours.

2 to 3Days a month

3 monthsMinimum

48 hoursTo a written price

1 daySev-1 acknowledged

There is no published number. The price depends on how complex the system is, how big the eval suite is, and how many vendors are in the stack, so I write it down after the call instead of guessing on a web page. If you start within 30 days of a diagnostic readout, the full $2,500 comes off.

The first month is reading the system. The second catches the first drift. By the third you have a real baseline, which is why the minimum is three months. After that it renews monthly, and either of us can cancel with 30 days notice.

The guarantee

Sev-1 in one business day, or the month is waived

Production AI broken in a way your users can see is acknowledged within 1 business day and triaged within 3. If I miss that, next month's retainer is waived.

The exit

14-day mutual exit

Either of us can end the engagement inside the first 14 days. Days worked are billed at the agreed rate and nothing else is owed.

05 · Who it is for

It shipped.Nobody is watching.

Teams whose AI work is in production, through a Scoped AI Build or on their own, and who have nobody keeping an eye on it. There is not enough work for a full-time AI lead, and the system still needs one.

It is the quieter of the two monthly doors. If there is still more to build and nobody to own it, the Fractional retainer fits better.

06 · FAQ

Questions.

What is AI System Continuity?

A monthly retainer that keeps the production AI you shipped running. 2 to 3 days a month: evals re-run against the baseline, prompt regressions triaged and fixed, model deprecations tracked, vendor escalations handled, and one small scope addition. Built for the gap between shipping and hiring a full-time AI lead.

How much does it cost?

There is no published price. The number depends on how complex the system is, how big the eval suite is, and how many vendors are in the stack. I put it in writing within 48 hours of the call. 3-month minimum, monthly renewal after. If you start within 30 days of a diagnostic readout, the full $2,500 comes off.

Who is it for?

Teams whose AI work has shipped, through a Scoped AI Build or on their own, and who need someone watching it. There is not enough work for a full-time AI lead, and the system still needs one.

How is it different from the Fractional retainer?

Continuity keeps what shipped running, 2 to 3 days a month. Fractional owns the program and ships the next builds, from $12,000 a month for one day a week, up to three days. If the work is done and needs watching, Continuity fits. If there is more to build, Fractional does.

What if I need more than 3 days a month?

Then you need a Scoped AI Build for the new work, or the Fractional retainer for ongoing ownership. I will tell you on the call which one fits. Continuity does not grow past 3 days a month.

What are the guarantee and exit terms?

Sev-1 incidents, meaning production AI broken in a way your users can see, are acknowledged within 1 business day and triaged within 3. If I miss that, next month's retainer is waived. Either of us can end the engagement inside the first 14 days. After the 3-month minimum, it renews monthly and either of us can cancel with 30 days notice.

GAT.

It shipped. Keep it running. Against the baseline.

30 minVirtual call

Walk me through the stack, the evals, and the vendors. You get the price in writing within 48 hours.