Skip to content

Module 10 · Lesson 40

Capstone: launch and defend an agent service

Bring the course together in a complete, evidence-led offer, architecture, pilot, and operating plan.

Updated August 2026 · AI Agency Academy

What you will learn

Make a clear, safer operating decision.

You will be able to present a complete AI-agent service that a client, operator, and reviewer can understand, test, and run responsibly.

Why this matters

Good agent work is useful before it is impressive.

The capstone is deliberately more than a demo. A useful service combines a real workflow, narrow offer, accountable agent contract, data and safety boundaries, evaluation evidence, controlled launch, and a plan to keep improving.

Field note 40

Make the relationship visible.

Core concepts

The language that keeps the work clear.

Service thesisThe specific buyer, workflow problem, bounded intervention, and value hypothesis.
Saved on this device.
Operating architectureThe human owners, sources of truth, tools, approvals, handoffs, and measures that make the service work.
Launch evidenceEvaluation results, approvals, rollback readiness, and pilot plan required before real customer exposure.
90-day planA sequenced plan for pilot review, improvement, responsible expansion, and service operations.
The practical method

Work through the decision in order.

  1. 01

    Choose one narrow service

    Use the market and workflow evidence from the course; do not combine several unrelated use cases.

  2. 02

    Assemble the contract

    State trigger, permitted context, tool boundary, human owner, hard paths, and success condition.

  3. 03

    Prove the pilot is ready

    Include test cases, evaluation result, client approvals, source map, access matrix, stop conditions, and rollback.

  4. 04

    Defend the next 90 days

    Show measurement cadence, client review, likely improvement, capacity assumption, and the decision rules for expansion.

Worked example

A realistic, bounded implementation.

A final project serves independent home-service companies that receive incomplete after-hours web enquiries. The offer is a guided intake, CRM record, owner routing, and approved acknowledgement—not a promise to close every lead.

The architecture maps form data, service-area policy, CRM contact record, task owner, message template, exception path, and daily review. Tests include normal request, missing address, urgent safety concern, duplicate record, and unavailable integration.

The pilot starts with one service category. The 90-day plan reviews completion, correction, escalation, response time, and customer feedback before any expansion. The learner can explain every boundary and trade-off to a client.

Build it in practice

Use this copyable working template.

Adapt it to the client’s evidence, policy, people, and tools. Do not treat placeholders as approved instructions.

Buyer + workflow: [statement]. Offer: [bounded change]. Agent contract: [summary]. Sources + permissions: [map]. Evaluation + pilot: [evidence]. 90-day operating plan: [dates, owners, decisions].
Spacebrain implementation

Put the operating system around the agent.

Build the capstone as a real Spacebrain workspace: contacts and lifecycle, permitted intake, ownership, tasks, messages or calls where appropriate, automations, review reporting, and client-facing operating records.

Practice

Before you move on

  • Present your service to a skeptical operator or peer.
  • Ask them to identify an unclear owner, source, permission, or stop condition.
  • Revise the plan until the first pilot is both useful and defensible.
  • The service solves one observed workflow problem.
  • Safety, access, and human ownership are explicit.
  • Evidence supports launch readiness.
  • The next 90 days have measured decisions, not vague ambition.

Put the learning to work

Build an AI service people can trust.

Create a free Spacebrain account and use the operating layer around your AI service.

Start free trial