Maintenance & Optimization Retainer

AI systems age in months. Keep yours sharp.

The chatbot that impressed everyone in March is confidently wrong by September. Models deprecate, prompts drift, prices move. The question is whether you find out from a dashboard or from a customer.

Put my systems on watch

Monthly, sized to what we maintain. Month to month after the initial term.

The week a model update ships Same event, two endings

A month on the retainer

Where the effort goes, week by week

Hover or tap a bar. Share of a typical month's effort, illustrative and sized to your inventory.

The watchlist

Five things that break AI systems while nobody looks

Deprecations land on someone else's schedule

Providers retire models, change defaults, and reshape APIs with short notice. We track every announcement that touches your stack and migrate on our calendar, not in a panic on theirs.

The same prompt does not stay the same

A prompt tuned for one model behaves differently on its successor, and quality sags so gradually nobody flags it. Monthly scoring against a fixed baseline catches sag while it is still invisible to customers.

Automations die at the joints

A renamed CRM field, an expired token, a webhook that silently stops firing. The workflow does not error, it just stops helping. We ping every joint on a schedule so dead links cannot hide.

Usage grows, and so does the bill

Costs creep as adoption spreads, and pricing changes land without ceremony. The monthly cost pass keeps spend proportionate and moves work to cheaper models when quality allows.

Last year's setup ignores this year's capability

Models improve fast enough that a workflow designed twelve months ago usually undersells what is now possible. We flag upgrades with expected payback, and you decide what is worth doing.

Where this sits

The retainer attaches to everything that ships

Whatever the audit justifies and your team adopts, this is what keeps it true a year later.

Start here: AI Readiness Audit The audit decides what deserves to exist before anything needs maintaining. Pair with: Team Training & Workshops Maintained systems plus a fluent team is the combination that compounds.

Questions

The fine print, answered plainly

What counts as maintenance versus new work?

Maintenance keeps what exists working and improving: monitoring, patches, prompt tuning, model migrations, cost passes. New workflows or integrations are scoped separately, and the monthly report flags candidates so nothing is sold by surprise.

Do you only maintain systems you built?

No. We adopt systems built by others after a short onboarding review that documents how they work and where they are fragile. If something is unmaintainable, we say so before taking your money.

What happens when something breaks?

You get a named contact and an agreed response window for outages in customer-facing systems. Fixes for covered systems are included in the retainer, not billed as extras.

How is the retainer priced, and can we leave?

Pricing is sized to the inventory we maintain: how many systems, how critical, how customer-facing. After the initial term it runs month to month, and everything stays documented so you can leave cleanly.

Get covered

Cheaper than the month nobody was watching.

Tell us what is already running and we will quote coverage. Every retainer includes:

Monthly health review, fixes, and prompt tuning for every covered system.
A cost pass that routinely finds more than it costs to run.
A one-page monthly report your leadership can read in two minutes.