A public-sector agency is scaling a Copilot Studio agent from pilot to production and must justify ongoing funding with clear ROI criteria. The…

The correct sequence for a pilot-to-production project is to first establish the goals, then gather data, analyze that data to build a financial model, and finally make a go/no-go decision.

1. Define baseline outcomes and success criteria (2): You must first know what success looks like before starting the pilot.

2. Run a time-boxed pilot and capture telemetry (4): The pilot generates real-world data on usage, performance, and costs.

3. Build a full TCO model with run costs (1): Use the telemetry from the pilot (step 4) to build an accurate, data-driven Total Cost of Ownership (TCO) model.

4. Produce the scale decision with sensitivity ranges (3): With the success criteria and the TCO model, you can make an informed decision to scale.

A. 1 -> 2 -> 4 -> 3: Building a full TCO model before defining success criteria or running a pilot to gather data is premature and based on assumptions, not evidence.

B. 4 -> 2 -> 1 -> 3: Running a pilot without first defining the success criteria (what you are measuring against) is inefficient and lacks clear direction.

D. 3 -> 2 -> 4 -> 1: Making the scaling decision at the beginning is illogical and defeats the purpose of conducting a pilot to inform that very decision.

1. Microsoft Cloud Adoption Framework for Azure. (2024). Strategy methodology. Microsoft Learn. The framework emphasizes defining motivations and business outcomes (Step 2) before proceeding with pilots and adoption plans (Steps 4, 1, 3). (Section: "Understand motivations").

2. Microsoft Power Platform. (2023). Planning a Power Apps project. Microsoft Learn. This guide details the project phases, starting with "Phase 1: Plan," which includes defining the business problem and success metrics (Step 2), before moving to design and making (Step 4). (Section: "Project phases > Phase 1: Plan").