SUPERWISE services · Step 4 of 7

Agent build

Turn one proven workflow into a working agent in your environment, tested with your team.
$35,000 to $75,000 pilot · $90,000 to $200,000 production6 to 10 weeks pilot · 3 to 6 months production

What you walk away with

  • Use case brief and success measures
  • The working agent, in your environment
  • Guardrail and scope definition
  • Test results and user acceptance record

Where this fits

Step 4 on the services path

Each step answers the question the last one raised. Start wherever the assessment points you.

Why it matters

What changes for your team

In 6 to 10 weeks, one real workflow runs on an AI agent inside your environment that your team has tested, understands and can extend.

Your people stop doing one task by hand

The agent takes a defined, repeatable task, such as sorting requests or pulling facts from documents, and your team reviews its output. Time saved is measured against a kickoff baseline.

The next AI decision rests on evidence

Test results, user acceptance feedback and a production decision show how the agent did on your own work. Nobody has to take a demo on faith.

You can show leaders and clients the real thing

When a board member, client or auditor asks what the AI does and what stops it doing more, you show them the live agent, its guardrails and its test record.

How it works

Three stages, with your team in the room

6 to 10 weeks pilot · 3 to 6 months production

Weeks 1–2

Pick the workflow and agree what success means, which data it uses and where its limits are.

Weeks 3–6

See a working agent early, built and tested with your team on real examples.

Weeks 7–10

Your users accept it, your team gets training and the guide, and you make the production decision on evidence.

The details

Everything you need to decide

Price, fit, your team’s time, how data is handled, and what you keep.

Price and scope

Draft pricing for review: we confirm a fixed scope and price after a scoping call.

TierPrice rangeDurationIncluded
Pilot$35,000 to $75,0006 to 10 weeksOne pilot workflow
Production$90,000 to $200,0003 to 6 monthsSandbox or limited data connections

Not included

  • A second workflow, which raises the price and is scoped separately.
  • Third-party software subscriptions, including the model, data and system licenses your agent depends on.
  • Fixing your source systems: data found missing, inconsistent or locked in a legacy system is reported, and clean-up is separate work you decide on.
  • A production rollout: the pilot runs on sandbox or limited data, and production is the next engagement.
  • Ongoing monitoring and maintenance after handover: that is Managed AI operations.
  • Legal advice or regulatory sign-off for your use of AI.
Is this the right starting point?

Good fit if

  • For teams ready to turn one defined workflow into a tested agent in their environment.
  • Teams of 50 to 500 people with a selected workflow
  • You want a defined handover your team keeps.
What you keep

Use case brief and success measures

One to two pages, set with your sponsor at kickoff: the workflow, its users and data, what the agent must never do and the measures that decide whether it worked. It is the yardstick for the pilot and the template for choosing the next workflow.

The working agent, in your environment

A running agent for the one pilot workflow, built with SUPERWISE Agent Studio components and connected to the sandbox or limited-scope data agreed in scope. Your team runs real work through it during acceptance testing and after handover.

Guardrail and scope definition

The written limits on what the agent may read and produce and which actions need a person’s approval, with tests that show each guardrail working. Use it to answer security, legal or compliance reviewers, and to decide later whether to widen what the agent may do.

Test results and user acceptance record

The functional, edge-case and guardrail tests, plus notes from your people who tried the agent on real examples: what passed, what failed and what was fixed. It is the evidence behind the production decision, and proof that the agent was tested before it was trusted.

Training session for your team

A hands-on walkthrough on using the agent, reading what it did, overriding it and reporting a problem. Your team runs the agent without us in the room, and new colleagues learn from the session notes.

Operating and extension guide

How the agent works: its inputs and outputs, data connections, guardrails, configuration and how to change or extend it. Use it to hand the agent to a successor, add a step or move to another provider.

Next use cases roadmap

A short, ranked list of the next workflows worth building, based on what the pilot showed about your data, systems and people. Use it to plan the next quarter’s AI work and budget, so the first agent starts a sequence.
What your team gives

About 8 to 12 hours a week across your team, most of it from the domain expert, and more in the acceptance weeks. These are planning figures; we confirm them against your workflow when we scope.

Executive sponsor

Chooses the workflow with us, confirms the success measures, clears blockers and makes the production decision.about 1 hour

Domain expert

Knows how the work is done today, supplies real examples and rules, and judges whether the agent’s output is right.about 4 to 6 hours

System and data owner

Grants access to the sandbox or limited data, answers system questions and approves any connection to a live system.about 2 hours in weeks 1 to 6

Approver and acceptance testers

Two or three future users test the agent on real cases and report what works and what does not.about 2 to 3 hours each in weeks 7 to 10
Data, security, and ownership
  • Where it runs: the agent is built and demonstrated inside your environment, using SUPERWISE Agent Studio, so your team sees it working where it will be used and not in a separate demo.
  • What data it touches: the pilot connects only to a sandbox or limited-scope data sources that you approve in scope. Anything wider, such as a live system or regulated data, is a scope change we agree before we connect it.
  • Protected health information: if the workflow involves PHI, tell us at the start so the pilot is scoped around it.
  • Guardrails and audit trail: guardrails and observability are set up from the first week, so each limit on the agent is written down and tested and what the agent did can be reviewed afterwards.
  • What we never do: we do not widen the agent’s access or connect it to live data without your approval, and the agent does not take an action outside the approvals defined in its scope.

SUPERWISE signs a Business Associate Agreement with healthcare customers, and the scope is confirmed during scoping. See the trust center.

You own the use case brief, guardrail definition, test results, training notes and operating guide, plus the agent’s source and configuration as listed in the scope. The agent runs in SUPERWISE Agent Studio, so keeping it running after we leave depends on the platform subscription, which is separate from the engagement fee and agreed in the scope before you sign. If you stop working with us, everything you keep is written so another team can carry the work forward.

The pilot ends with a production decision. A production build is a separate engagement, from $90,000 over 3 to 6 months, for an agent on live data with integrations, monitoring and handover; Managed AI operations, from $3,500 a month on a 12-month term, keeps a live agent monitored and maintained. If the pilot shows you need to map more systems first, Data and systems discovery comes before another build, and support after handover is agreed in the scope.

Examples and how we measure success

These are examples of suitable work, not claims about past client engagements.

Intake triage

The agent reads each request from email or a form, tags its type and urgency and drafts the routing. A person approves anything that leaves the building.

Document extraction into a system

The agent pulls the fields your team keys in today from invoices, claim forms or order documents into a staging area in your system. A person checks exceptions before anything posts.

Internal knowledge questions on approved data

Staff ask about policies, procedures or product documents and get answers drawn only from the documents you approve, with the source shown. Anything else gets a plain “not in the approved material”.

Scheduling and follow-up drafts with approval

After a call or a job, the agent drafts the follow-up and proposes the next appointment. A team member edits and sends it; the agent sends nothing on its own.

Warranty or returns review

The agent checks a request against your written rules and recommends approve, deny or escalate. A reviewer makes the call, and the reasons stay on record.

Measures we agree at kickoff

  • Task completion without help: the share of real cases from a test set of your own work that the agent finishes correctly with no one stepping in.
  • Escalation rate and reasons: how many cases go to a person and why, so you can see whether it escalates the right things.
  • Time per task against the baseline: task time with the agent compared with today, measured at kickoff and again at acceptance.
  • Cost per completed task: what each finished task costs to run, compared with doing it by hand.
  • User acceptance: the people who will use the agent test it on real examples and sign off, or say what must change first.
Who delivers the work

No case study of an agent build engagement has been published, so none is shown. The nearest real stories on this site are Renova Health, where AI assistants support chronic care managers with every AI decision logged and traceable to its source, and Bridgestone, where a governed workflow connects warranty claim inspections, approvals and resolutions with end-to-end traceability. Neither is presented as the result of this service.

Agent build questions

What to expect before you start

Talk to us

Talk to our team

Tell us about your team and we will scope agent build with you.

Tools in scope (required)