Tier 3 · scoped build
AI agent and workflow automation
Agents that triage, extract, route and draft inside the systems you already run, with a human decision point wherever the stakes require one.
Who this is for
A team whose people spend their day moving work between systems, in a workflow where some decisions can be automated and some must not be.
From
$22,000
- Typical
- $35,000 to $90,000
- Other currencies
- €19,900 · £16,900 · S$29,000
Thirty minutes with the engineer who would run the work, before any paperwork.
The work, stage by stage
In order, with what happens at each phase and what it produces. Phases rather than dates, because a build is scoped per engagement and the catalog deliberately publishes a price band instead of a duration.
Map the workflow as it runs today
Phase 1
Step by step, including the exceptions people handle by hand and the informal rules nobody wrote down. Automating the documented process instead of the real one is the most common way this fails.
Decide what the agent may do alone
Phase 1
Each step classified automatable, automatable with review, or human only, with the reasoning recorded. A governance decision taken before the build rather than discovered after it.
Build the tools before the agent
Phase 2
Each action the agent can take is a bounded, testable tool with its own permissions, so authority is granted per action rather than by handing the agent a credential.
Wire the human decision points
Phase 3
Approval, escalation and override paths, each tested with the approver refusing rather than only approving, because the refusal path is the one that matters and the one nobody tests.
Stop control and audit trail
Phase 3
A kill switch exercised before launch, and a trail recording what the agent did, what it was given, and which model version produced it.
Shadow run, then launch
Phase 4
The agent runs alongside the people doing the same work, and every disagreement is reviewed, before it is given the work on its own.
Everything that arrives, listed
What is handed over at the end, and what you own of it. If something here would not be useful to you, say so before we start and the scope changes.
- The running agent, deployed in your environment, with its tools permissioned per action
- Source code for what was written for you, which you own on payment in full under our terms
- A step-by-step authority map: what the agent does alone, what it does with review, and what stays human
- Human decision points built and tested from the approver's side, including refusal and override
- A stop control, exercised before launch, with the behavior of in-flight work documented
- An audit trail recording the action, the input, the model version and the accountable person
- Shadow-run results: where the agent and your team disagreed, and what changed as a result
What this covers, and who takes the rest
Stated up front, because scope is what derails an engagement in week two rather than at the proposal.
Inside this build
- The workflow as it actually runs, exceptions included
- Tools, permissions, human decision points, stop control and audit trail
- A shadow run against your team's own output before launch
Handled separately
Named so you can plan and budget for it, whether we do it or somebody else does.
- Changes to the upstream systems the agent reads from, which their owners make
- Ongoing operation and incident response, which the Operate retainer covers
Start with a call, not a proposal.
Thirty minutes, straight to the engineer who would run the work. We will tell you if a smaller engagement answers your question, or if nothing here does.
Typical reply within one business day · Engagements start at $2,500
Scoped per engagement