Managed AI services. We build and run it with you.
Managed is the Precision plan taking discovery calls now. Your own named engineer builds your workflows, connects your systems and runs your agents. Approvals stay on by default, and implementation is included.
The handoff template has changed. Please use the new version for the next draft.
I’ve prepared the instruction change and a draft to check against the new template.
- Instruction
- Use the revised client handoff template.
- Check
- Compare the prepared draft with Jamie’s template.
- Owner
- Alex · Named engineer
What Managed includes
Your own named engineer, one accountable partner who builds your workflows and works alongside your team.
Implementation included: connecting your systems and setting up your Agent OS (AOS), agents and routines, then testing on real examples.
Monitored and kept current: we keep your sources connected and add new documents as your business works.
Continuous improvement as new data and corrections come in, with your reviewers' corrections handled as agreed changes to the agent.
Upkeep as models change, with every model still under the same zero-retention routing.
Approvals on by default, so your team reviews what agents prepare and an owner decides which routine work, if any, runs on its own.
Packages run from 100,000 to 500,000 credits in steps of 50,000, each at $2,550 a month per 100,000 credits on a 12-month agreement. From 600,000 to 1,000,000 credits in steps of 100,000, then from 1,250,000 to 2,500,000 credits in steps of 250,000, with no price shown. Above 2,500,000 credits is custom.
Is this the same as a fractional chief AI officer?
A fractional chief AI officer is an advisory role for strategy and oversight. Managed AI runs the system day to day: a named Precision engineer builds your workflows, connects your systems and runs your agents. Approvals stay on by default, and implementation is included.
Managed AI or your IT provider?
An IT provider often sets up access to an assistant. Managed runs the workflows, agents and approvals built around your work. The two can work side by side.
Ask what each scope covers. If an IT proposal covers an existing assistant, check whether it also includes the workflow, connected sources, agent upkeep and review process.
Agree who owns what: before launch, settle who manages each system, connection, permission and escalation.
Questions
What is managed AI?
Ongoing operation, monitoring and improvement of an AI system after launch. On Managed, our engineers run your Agent OS, agents and workflows and keep them current as your work changes.
Do we need technical staff?
No. Our engineers build and run it. Your team reviews what agents prepare and approves sends by default. IT joins when the question is data access or security.
How much does Managed cost?
Managed requires a minimum purchase of 100,000 credits, which is $2,550 a month on a 12-month agreement, with implementation included and no setup fee. We estimate a typical user uses about 10,000 credits a month. A discovery call scopes the workflows and systems involved.
What happens when something goes wrong?
We monitor your system, and our written incident process covers finding the problem, containing it, recovering and telling you what happened. Every model call is recorded: which job made it, which model answered and what it cost.
Can we start on Free or Team instead?
They're coming soon: Free to try it on your own team, Team to put it to work across your team. Join the waitlist on the pricing page, or start on Managed today.
Can we add agents later?
Yes. With your pipeline and Agent OS in place, each new agent is faster to build and starts from everything your business already knows.