Reference document for how the APA scoring engine works. All data driven from apa.yaml.
| ID | Label | Description |
|---|---|---|
agent_builder |
Agent Builder | No-code agents within Microsoft 365 |
m365_copilot |
Microsoft 365 Copilot | Built-in Copilot experiences (fast-track path only) |
copilot_studio |
Copilot Studio | Low-code agents connecting to external systems |
foundry |
Microsoft Foundry | Full-code agents with custom AI models |
M365 Copilot is excluded from the scored assessment. It is only recommended via the prescreen fast-track path ("I want to enhance Microsoft 365 Copilot"). In the full wizard, m365_copilot is always zeroed.
Five questions, each scored 0–3 per platform. Max raw score: 15 (5 × 3).
Sets persona weight. Determines whether the builder has the skills for each platform.
| Option | ID | Agent Builder | CS | Foundry |
|---|---|---|---|---|
| Business user / SME — no coding | q1a | 3 | 1 | 0 |
| Low-code maker / IT pro | q1b | 1 | 3 | 0 |
| Professional developer | q1c | 0 | 2 | 3 |
| Data scientist / ML engineer | q1d | 0 | 0 | 3 |
CS gets 2 for q1c because it supports pro developers via YAML authoring and VS Code extension.
External-facing is a hard constraint that eliminates Agent Builder and M365 Copilot.
| Option | ID | Agent Builder | CS | Foundry | Hard Rule |
|---|---|---|---|---|---|
| Internal employees only | q8a | 3 | 3 | 1 | — |
| External users | q8b | 0 | 3 | 3 | Zeros AB, M365 |
| Not decided yet | q8d | 2 | 2 | 1 | — |
Deployment surface. Agent Builder can only run inside Microsoft 365.
| Option | ID | Agent Builder | CS | Foundry | Hard Rule |
|---|---|---|---|---|---|
| Inside Microsoft 365 apps | q2a | 3 | 3 | 1 | — |
| Custom app (website/mobile) | q2b | 0 | 3 | 2 | Zeros AB |
| Background (event-driven) | q2c | 0 | 2 | 3 | Zeros AB |
| Multiple / not decided | q2d | 1 | 3 | 2 | — |
Foundry gets 1 for q2a because Foundry agents can be surfaced in Teams via custom bot frameworks.
Single strongest discriminator between Copilot Studio and Foundry.
| Option | ID | Agent Builder | CS | Foundry | Hard Rule |
|---|---|---|---|---|---|
| Simple Q&A / lookups | q4a | 3 | 3 | 1 | — |
| Conversational (multi-turn) | q4b | 2 | 3 | 2 | — |
| Multi-step tasks | q4c | 0 | 3 | 3 | Zeros AB |
| Complex orchestration | q4d | 0 | 1 | 3 | Zeros AB, M365 |
Foundry gets 1 for q4a because Foundry can do Q&A via prompt flow.
Data source requirements. Agent Builder is the most limited in terms of data access.
| Option | ID | Agent Builder | CS | Foundry | Hard Rule |
|---|---|---|---|---|---|
| Microsoft 365 data | q3a | 3 | 2 | 0 | — |
| Business systems (CRM, etc.) | q3b | 1 | 3 | 2 | — |
| Advanced / private data | q3c | 0 | 1 | 3 | Zeros AB |
| Mixed sources | q3d | 1 | 3 | 2 | — |
Hard rules zero out platforms before scores are summed. They represent real platform limitations.
| Trigger | Platforms zeroed | Reason |
|---|---|---|
| q8b (external users) | AB, M365 | Cannot publish externally |
| q4d (complex orchestration) | AB, M365 | Requires Foundry |
| q4c (multi-step tasks) | AB | Cannot follow processes or take actions |
| q2b (custom app) | AB | Can only run inside Microsoft 365 |
| q2c (background) | AB | No event-driven capabilities |
| q3c (advanced data) | AB | Can only access Microsoft 365 data |
Additionally, M365 Copilot is always zeroed in the full assessment (hard-coded in JS).
For each platform not zeroed: sum the scores from all answered questions. Range: 0–15.
Persona preferences force one platform above another in ranking regardless of scores. Unlike hard rules, all scores are preserved — the override only affects sort order. A rationale message is displayed as a key factor on the recommendation card.
| Trigger | Prefer | Over | Rationale |
|---|---|---|---|
| q1d (data scientist / AI-ML) | Copilot Studio | Agent Builder | CS supports custom AI model integration, code-first development, and flexible orchestration that AB lacks |
This means that when a data scientist selects answers where Agent Builder would otherwise outscore Copilot Studio, CS is still recommended — but AB's actual scores remain visible in the score breakdown.
| Score | Label |
|---|---|
| 12–15 | Strong fit |
| 8–11 | Good fit |
| 4–7 | Possible fit — review tradeoffs |
| 0–3 | Not recommended |
Platforms are sorted by score descending. The highest-scoring platform is the primary recommendation. The second-highest is shown as "Also consider."
When the top two platforms score within 2 points, they're presented as a complementary pair with a rationale banner:
| Pair | Rationale |
|---|---|
| Copilot Studio + Foundry | Build in CS, extend with custom code in Foundry |
| M365 Copilot + Copilot Studio | M365 Copilot for end users, CS for customization |
| Agent Builder + M365 Copilot | AB for M365-native agents, M365 for extensibility |
Persona-based tiebreakers — when two platforms score equally and a specific persona answer is selected, one platform is preferred:
| Trigger | Platforms | Prefer | Rationale |
|---|---|---|---|
| q1c (professional developer) | AB, CS | CS | CS supports code-first authoring via VS Code extension |
| q1d (data scientist / AI-ML) | CS, Foundry | CS | CS provides a faster path to production agents |
A "Why not?" explainer identifies the single question where the winner most outscored the runner-up.
Contextual warning banners when answer combinations are logically contradictory:
| Condition | Note |
|---|---|
| q2c + q4a | Background agent doing simple Q&A — contradictory |
| q8b + q2a | External users in M365 apps — external users can't access your tenant |
| q1a + q4d | Business user wants complex orchestration — requires dev skills |
| q1a + q3c | Business user needs advanced data sources — requires technical expertise |
When Foundry wins but the builder is a business user (q1a), a banner advises partnering with a development team.
Across all 768 possible answer combinations:
| Platform | Wins | % |
|---|---|---|
| Copilot Studio | 599 | 78.0% |
| Foundry | 159 | 20.7% |
| Agent Builder | 10 | 1.3% |
Note: The distribution above reflects raw scoring only. With persona preferences applied, the two MLEng combos where Agent Builder would win are overridden to recommend Copilot Studio instead, reducing AB's effective wins to 8.
Ties: 70 combos (9.1%) — 68 are CS/Foundry, 2 are AB/CS.
AB only wins in its narrow sweet spot: business user or undecided persona, internal audience, M365 deployment, simple Q&A or conversation, M365 data. All 10 winning combos share this profile.
| Combo | AB | CS | F | Override? |
|---|---|---|---|---|
| BizUser + Internal + M365 + SimpleQA + M365Data | 15 | 12 | 3 | — |
| BizUser + Internal + M365 + Converse + M365Data | 13 | 12 | 3 | — |
| BizUser + Internal + Multiple + SimpleQA + M365Data | 13 | 12 | 4 | — |
| BizUser + Undecided + M365 + SimpleQA + M365Data | 14 | 11 | 3 | — |
| BizUser + Undecided + M365 + Converse + M365Data | 12 | 11 | 3 | — |
| BizUser + Undecided + Multiple + SimpleQA + M365Data | 12 | 11 | 4 | — |
| ProDev + Internal + M365 + SimpleQA + M365Data | 12 | 12 | 6 | — |
| ProDev + Undecided + M365 + SimpleQA + M365Data | 11 | 11 | 6 | — |
| MLEng + Internal + M365 + SimpleQA + M365Data | 12 | 11 | 6 | CS via persona pref |
| MLEng + Undecided + M365 + SimpleQA + M365Data | 11 | 10 | 6 | CS via persona pref |
The ProDev rows are technically correct — even a developer building a simple Q&A bot on SharePoint data should consider Agent Builder. The MLEng rows are overridden by the persona preference: Copilot Studio is recommended instead, with the rationale shown on the card.
Foundry wins when answers include any combination of: pro dev or ML persona (q1c/q1d), background deployment (q2c), complex tasks (q4c/q4d), or advanced data (q3c). Its strongest signal comes from persona + task complexity.
CS wins 78% of combinations because it scores 2–3 on almost every option. It's the most versatile platform — usable by low-code makers, deployable anywhere, and capable of multi-step tasks. This distribution is intentional: CS is the default recommendation unless the user's needs clearly require Foundry's full-code capabilities or fit entirely within AB's narrow scope.
| Platform | Min | Max | Avg |
|---|---|---|---|
| Agent Builder | 11 | 15 | 12.5 |
| Copilot Studio | 8 | 15 | 11.9 |
| Foundry | 9 | 15 | 11.9 |
No combination produces a "best platform" below 8, so every user gets at least a "Good fit" recommendation.
| Note | Combos | % |
|---|---|---|
| Background + SimpleQA | 48 | 6.2% |
| External + M365Apps | 64 | 8.3% |
| BizUser + Orchestrate | 48 | 6.2% |
| BizUser + AdvData | 48 | 6.2% |
| Foundry + BizUser (persona mismatch) | 12 | 1.6% |
Notes are not mutually exclusive — a single combo can trigger multiple notes.