SUPERWISE services
Managed AI operations
For teams that want a dependable partner to monitor, maintain, report on, and improve live AI.
$3,500 to $15,000 per month12-month term
Outputs you keep
Operating baseline
Runbook and escalation path
Incident and change record
What you get
Working outputs, not a presentation about them
Your live AI has a named operating team, an agreed service level and a regular record of performance, incidents, changes and next improvements.
Service agreement and responsibility map
The written coverage, business-hours service level, response and escalation expectations, client contacts, approval boundaries and exclusions for the agreed agent, Sentinel fleet or Chat workspace. Use it to know who acts when something changes or fails and which requests are maintenance versus a separately scoped build.
Operating baseline
A month-one record of the system, configurations, dependencies, expected operating signals, known issues and the measures the monthly report will follow. Use it to distinguish a real change from the condition accepted at handover and to onboard a future operator.
Runbook and escalation path
The checks, routine maintenance steps, issue categories, contacts, approval points and handoff instructions used to operate the covered system. Use it during an incident, a staff change or a provider handoff so the response does not depend on one person's memory.
Incident and change record
A running record of reported issues, investigation, resolution, configuration and tuning changes, approvals and any open follow-up. Use it to answer what happened, what changed and why, and to find repeated problems that need a larger fix.
Monthly operating report
A monthly summary of the agreed performance and availability signals, incidents, service-level handling, maintenance, changes, risks and decisions needed from your team. Use it for the operating review, internal reporting and evidence that the live system is being watched rather than merely left on.
Quarterly service review and improvement plan
A quarterly review of performance, risk, recurring issues, product or dependency changes and a ranked set of maintenance and improvement actions. Use it to approve the next quarter's changes, budget separate build work and decide whether the service still fits the business.
Configuration history and handover package
The current runbook, operating records, reports and history of configuration and approved changes maintained through the term. Use it to renew with clear context, bring operation in-house or hand the service to another provider.
Outcomes
What will we have at the end?
Someone owns the system after launch
Monitoring, maintenance, issue handling and routine tuning sit in one operating cadence, so the live agent, Sentinel fleet or Chat workspace does not become an unowned side project.
Problems follow an agreed response path
The service agreement defines coverage and response expectations, and the runbook defines who investigates, who approves a change and when an issue comes back to your team.
Leadership can see what changed
Monthly operating reports and quarterly reviews give the business a durable record of performance, incidents, configuration changes, risks and the next improvement decisions.
Examples
What could this look like in our business?
These are examples of suitable work, not claims about past client engagements.
Example: a live agent needs a steady owner
An agent depends on changing data, integrations and business rules. Managed operations watches the agreed signals, handles routine maintenance, records incidents and brings material changes to the client owner.
Example: a Sentinel fleet needs policy upkeep
A team has Sentinel on its covered PCs and needs one place to review fleet status, policy changes, issues and the documented limits of coverage. The service maintains the agreed fleet and includes it in the operating report.
Example: Chat workspaces need administration
Departments use approved Chat workspaces and need routine configuration, user and issue handling under a named owner. Managed operations keeps the agreed workspaces in the service cadence.
Example: an update changes behaviour
A model, dependency, data source or business rule changes and the live workflow behaves differently. The team investigates against the baseline, records the decision and applies an approved fix or tuning change.
Example: a quarterly review exposes the next risk
The operating record shows repeated exceptions, new data conditions or a control that no longer fits. The review turns that pattern into an owned improvement rather than leaving it in support tickets.
Success measures
How do we know it worked?
We agree the measures at kickoff and use them in the final decision.
Measures agreed at kickoff
- Service-level performance: covered issues are acknowledged and handled against the response terms agreed in the service agreement.
- Operating visibility: every monthly report contains the agreed performance signals, incidents, maintenance, changes, risks and open decisions.
- Change control: material configuration and tuning changes have a reason, approval and outcome in the operating record.
- Issue learning: recurring incidents and exceptions become owned corrective or improvement actions at the quarterly review.
- Handover readiness: the current runbook, configuration history and open issues are sufficient for the client or another provider to take over.
Your team
What does our team need to give?
About 4 to 8 hours across your team in month 1, then about 2 to 4 hours a month plus the quarterly review and any material incident. These are planning figures; the covered systems and response model determine the actual need.
Service owner
Owns the business priority, receives escalations, approves material changes and attends the monthly and quarterly reviews.about 1 hour, with more time during a material incident
System or workspace administrator
Coordinates access, users, dependencies and client-side changes and works with us during onboarding and issue investigation.about 2 to 3 hours in month 1, then about 1 hour
Domain owner
Judges whether outputs and workflow behaviour still fit the work and helps evaluate exceptions or requested tuning.about 1 hour when review or tuning is needed
Security or governance owner
Reviews material incidents, control changes and the quarterly risk record where the covered system requires it.about 1 to 2 hours each quarter, plus incident time when needed
Duration
A build path your team can see
12-month term
Month 1
Complete handover, set service levels, and establish the operating baseline.
Every month
Monitor, maintain, tune, resolve issues, and provide the operating report.
Every quarter
Review performance, risk, changes, and the next improvement plan.
Data and security
Is our data safe?
How the engagement handles data
- Access is limited to the covered system and operating tasks named in the service agreement. Your organization retains control of its accounts and credentials, and approval boundaries are recorded in the runbook.
- The operating record captures incidents, maintenance, approved configuration changes and the agreed performance signals; it does not replace the product's own audit record.
- For Sentinel, supported requests are checked on the PC and are not routed through SUPERWISE. Sentinel records metadata, not prompt content, and the service does not claim coverage beyond the documented Sentinel facts.
- For SUPERWISE Chat, the covered workspaces and their administration are agreed in scope. Chat uploads are stored in SUPERWISE's Google Cloud environment and are handled separately from Sentinel's on-device processing.
- The SUPERWISE platform is not covered by a BAA, so protected health information must stay out of it. We never widen production access or make a material change outside the approvals in the service agreement.
Ownership
What do we own when it is over?
You keep the operating records, monthly reports, configuration history, runbooks, decisions and handover material. Ownership of product subscriptions and any underlying source or configuration follows the product and build agreements already in place. If you stop working with us, the current package and open-issue list are handed back so your team or another provider can take over.
After the engagement
What happens after?
Managed AI operations runs on a 12-month term. Each quarterly review decides what continues as routine operation and what needs a separately scoped integration, new build or governance change. Before renewal, the current coverage, service level and improvement plan are reviewed. If the service ends, the handover package closes the operating record; support does not continue beyond the agreed term unless renewed.
Draft pricing for review
Know the range before the first call
We confirm a fixed scope and price after a scoping call.
| Tier | Price range | Duration | Included in a line |
|---|---|---|---|
| Service | $3,500 to $15,000 per month | 12-month term | One agent or Sentinel fleet |
Not included
- A new application, agent or workflow build, or a major new integration.
- Software, model, data, device-management or other third-party subscriptions.
- 24/7 coverage or a one-hour response unless it is priced and written into the service agreement.
- Additional agents, systems, sites or business units beyond the covered agent, Sentinel fleet or Chat workspace.
- Major data remediation, product development or compliance certification.
Fit
Know whether this is the right starting point
Teams with a live agent, Sentinel fleet, or Chat workspace
Good fit if
- For teams that want a dependable partner to monitor, maintain, report on, and improve live AI.
- Teams with a live agent, Sentinel fleet, or Chat workspace
- You want a defined handover your team keeps.
Start somewhere else if
You need to build or deploy the system before operating it.
Start with Agent buildWho delivers and proof
Delivered by the SUPERWISE services team
The SUPERWISE services team runs managed AI operations for Renova Health and Boca Raton Regional Hospital today. Renova Health has a published case study showing governed AI in chronic-care work. There is no published Boca Raton Regional Hospital case study on this site, so we do not invent a link or result for it.
Related published work
The nearest approved examples for this work.
Managed AI operations questions
What to expect before you start
The service runs $3,500 to $15,000 per month on a 12-month term. The low end covers one agent or Sentinel fleet, a business-hours service level, a monthly report and a quarterly review. More agents or systems, more sites, frequent tuning, 24/7 coverage or one-hour response move it up the range. This is draft pricing for review.
The base service uses a business-hours service level. Exact acknowledgement, response and escalation terms are agreed in the service agreement. We do not promise 24/7 or one-hour response unless that coverage is explicitly scoped and priced.
Yes, when the month-one handover can establish the architecture, access, dependencies, baseline and runbook needed to operate it. Missing documentation or a system that needs repair can require separate transition work before service levels begin.
Routine maintenance and agreed tuning are included for the covered system. A new workflow, major integration or application build is separate because it changes the system and its acceptance requirements.
Your service owner handles decisions and escalations, the administrator coordinates client-side access or changes, and a domain owner helps judge behaviour when needed. Plan on about 2 to 4 hours a month across the team in normal operation, plus quarterly review and incident time.
You keep the reports, records, configuration history, runbooks and handover material. We return the current open-issue list and operating context so your team or another provider can take over; support ends with the agreed term.
The service path
Where this fits
Learn about the decisions
What can you see while your AI is running?
Choose the operating signals that need regular review.
What does AI enforcement at runtime actually look like?
See where controls act while an AI request is running.
Can you prove afterwards who did what with AI?
Understand the activity record behind operating reports.
Products this service puts to work
SUPERWISE Sentinel
The service monitors and maintains the agreed Sentinel fleet.
SUPERWISE Chat
The service maintains the agreed Chat workspaces and reports on their use.
The SUPERWISE platform
The service operates the agreed agents running on the platform.
Talk to us
Tell us about Managed AI operations
A few details help the services team route your request.