Solution Templates¶
Ready-to-build agent designs. Each template gives you everything needed to go from zero to a deployed agent — and they come in two flavors depending on how far up the ramp the job pushed you.
- Studio templates are low-code specs: a copy-paste system prompt, knowledge source guidance, topics to configure, starter prompts, a test case table, and a deployment checklist. A maker can build any of them in a half-day on Copilot Studio's low-code canvas.
- Foundry templates are pro-code blueprints: a representative code scaffold, an evaluation gate, and a deployment checklist for developers. Reach for these only when a Studio agent genuinely can't carry the job.
Most teams should start in Studio. Climb to a Foundry template only when you hit a real ceiling.
Template or walkthrough? Use both.
A solution template is the blueprint — the spec for what to build and why: the system prompt, knowledge-source guidance, topics, a test-case table, and a deployment checklist. Its matching Studio walkthrough (under Stage 6) is the how — the click-by-click build in Copilot Studio. Most function templates link straight to their walkthrough, and each walkthrough links back to its blueprint. Read the template to understand the design; follow the walkthrough to build it.
Studio templates (low-code)¶
Grouped by functional area. Start with the cross-functional Policy FAQ Agent — it adapts to any team with a document library — then add the function-specific agents your org needs.
Cross-functional¶
| Template | Best for | Time to build |
|---|---|---|
| Policy FAQ Agent | Any team with a policy or process doc library — HR, IT, Finance, Legal | 2–3 hours |
| Meeting Notes & Action Tracker Agent | Any team wanting meeting notes turned into tracked decisions, owned actions, and follow-up drafts | 3–4 hours |
Human Resources¶
| Template | Best for | Time to build |
|---|---|---|
| Onboarding Buddy Agent | HR and people teams wanting a personalised first-week guide for new starters | 3–4 hours |
| HR Benefits & Leave Agent | HR teams deflecting benefits, PTO, and leave questions while escalating personal cases safely | 3–4 hours |
| Performance Review Prep Agent | Employees and managers preparing for review cycles with self-assessment drafts from their own notes | 3–4 hours |
| Recruiting & Interview Scheduler Agent | Hiring managers and recruiters drafting job descriptions, interview questions, and process guidance | 3–4 hours |
IT¶
| Template | Best for | Time to build |
|---|---|---|
| IT Helpdesk Triage Agent | IT teams looking to deflect tier-1 tickets and automate ticket creation | 4–5 hours |
| IT Access Request Agent | IT teams wanting guided software, hardware, and permissions requests with automatic ITSM ticket creation | 4–5 hours |
| IT Knowledge & How-To Agent | IT teams giving employees step-by-step how-to answers from KB articles with clean helpdesk escalation | 3–4 hours |
| Software License & Asset Query Agent | IT teams answering license entitlement, device request, and asset-return questions from policy | 4–5 hours |
Finance¶
| Template | Best for | Time to build |
|---|---|---|
| Finance Expense & Procurement Agent | Finance and ops teams wanting self-service answers on expense policy, procurement rules, and approval thresholds | 3–4 hours |
| Budget & Spend Q&A Agent | Teams asking budget questions — spend categories, thresholds, cost-centre owners, month-end deadlines | 3–4 hours |
| Invoice & AP Status Agent | Employees and vendors checking invoice submission, payment terms, and AP status guidance | 3–4 hours |
Sales¶
| Template | Best for | Time to build |
|---|---|---|
| Sales Enablement Agent | Sales teams needing instant competitive intel, case studies, and talk tracks | 3–4 hours |
| Sales Proposal & RFP Agent | Sales teams needing approved proposal content, RFP responses, and case studies assembled on demand | 3–4 hours |
| Account Research & Briefing Agent | Sellers wanting a pre-meeting account brief — background, activity, opportunities, and talking points | 3–4 hours |
| Deal Desk & Pricing Approval Agent | Sellers navigating pricing and discount policy, approval thresholds, and deal-desk sign-off | 3–4 hours |
Legal & Compliance¶
| Template | Best for | Time to build |
|---|---|---|
| Legal & Compliance Guidance Agent | Legal and compliance teams wanting to deflect routine process questions with strict escalation controls | 3–4 hours |
| Contract Review & Clause Agent | Business teams finding approved clause language and flagging terms that need legal review | 3–4 hours |
| NDA Intake & Triage Agent | Employees requesting NDAs — right template, required info, approval path — with non-standard cases routed to legal | 3–4 hours |
Marketing¶
| Template | Best for | Time to build |
|---|---|---|
| Marketing Campaign Agent | Marketing teams turning a campaign brief into an asset checklist and on-brand first-draft copy | 3–4 hours |
| Content Repurposing Agent | Marketing teams turning one approved asset into on-brand social, email, and abstract derivatives | 3–4 hours |
| Brand & Messaging Guardian Agent | Marketing teams checking draft copy against brand voice, terminology, and messaging guidelines | 3–4 hours |
Customer Support¶
| Template | Best for | Time to build |
|---|---|---|
| Customer Support Deflection Agent (internal) | Support teams answering from approved help content, drafting replies, and deflecting common questions | 3–4 hours |
| Knowledge Article Drafting Agent | Support agents turning a resolved ticket into a clean, reusable KB article in the team's format | 3–4 hours |
| Escalation Routing Agent | Front-line support finding the right tier, team, and severity, and drafting the escalation summary | 3–4 hours |
Procurement¶
| Template | Best for | Time to build |
|---|---|---|
| Procurement Sourcing Agent | Procurement teams giving requesters a compliant vendor shortlist with policy gates and approval routing | 3–4 hours |
| Supplier Onboarding & Compliance Agent | Teams onboarding a new supplier — required docs, compliance checks, approval steps, and status | 3–4 hours |
| Purchase Request Intake Agent | Employees raising a purchase request — category, threshold, preferred vendors, and justification | 3–4 hours |
Field Service¶
| Template | Best for | Time to build |
|---|---|---|
| Field Service Triage Agent | Field service teams triaging symptoms to likely cause and the right dispatch path, safely | 4–5 hours |
| Parts & Inventory Lookup Agent | Field technicians checking part availability, compatibility, and how to order or reserve a part | 4–5 hours |
| Work Order Status Agent | Technicians and dispatchers checking work-order status, next appointments, and status updates | 4–5 hours |
Engineering¶
| Template | Best for | Time to build |
|---|---|---|
| Engineering On-Call Runbook Agent | Engineering teams matching alerts to runbook steps and knowing the escalation path under pressure | 4–5 hours |
| Deployment & Release Notes Agent | Engineering teams turning a change log into clean release notes and a deployment summary | 3–4 hours |
| Architecture Decision Record Agent | Engineering teams drafting and finding ADRs from the team's template — context, decision, consequences | 3–4 hours |
Workplace & Operations¶
| Template | Best for | Time to build |
|---|---|---|
| Event Coordination Agent | Workplace and ops teams planning internal events — rooms, catering, comms — within policy | 3–4 hours |
| Facilities & Workspace Request Agent | Employees reporting a facilities issue or requesting a workspace change, guided through the form | 3–4 hours |
| Travel & Expense Policy Agent | Employees checking travel booking policy, per-diem rules, approvals, and how to book or expense a trip | 3–4 hours |
Foundry templates (pro-code)¶
Reach for these only when a Studio agent genuinely can't carry the job — custom code, your own evaluation gate, and identity you own.
| Template | Best for | Time to build |
|---|---|---|
| Pro-Code Grounded Q&A Agent | Developers whose grounded Q&A workload outgrew Studio — custom retrieval, an evaluation gate, code-owned identity | 1–2 days |
| Multi-Agent Workflow Orchestrator | Developers automating a multi-step process across specialist agents and tools, with deterministic control and a workflow-level eval gate | 2–4 days |
| High-Volume Document Processing Agent | Ops, finance, or legal teams extracting structured data from documents at scale, with schema validation and a precision/recall gate | 2–4 days |
| Customer-Facing Support Agent | Developers building an external, action-taking support agent with strict guardrails, content safety, and continuous evaluation | 3–5 days |
| NL-to-SQL Analytics Agent | Developers building "ask your data warehouse" self-service analytics with a SQL-safety layer and a correctness gate | 2–4 days |
| Browser-Using (Computer-Use) Agent | Developers automating web apps that have no API — sandboxed, allow-listed, and human-confirmed on irreversible actions | 3–5 days |
| Voice Agent / Contact-Center IVR | Developers building a real-time phone agent — speech-to-text, grounded conversation, text-to-speech, clean human handoff | 3–5 days |
| Code-Review / PR-Triage Agent | Developers and devex teams adding advisory, private-repo PR review and reviewer routing — comments only, never merges | 2–4 days |
| Multi-Modal Document Classification Agent | Developers sorting a high-volume document stream by vision + text, with confidence-based routing and a precision/recall gate | 2–4 days |
| Batch Summarization Pipeline Agent | Developers summarizing high volumes of documents or transcripts on a schedule, behind a groundedness eval gate | 1–2 days |
| Structured Data Extraction Agent | Developers extracting schema-constrained JSON from unstructured documents at scale, with validation and an accuracy gate | 1–2 days |
| Content Moderation & Safety Agent | Developers moderating user content against a policy taxonomy with Azure AI Content Safety and human-review routing | 2–4 days |
| Deep Research Agent | Developers building a multi-step research agent that plans, gathers, cites, and self-critiques, with a citation eval gate | 2–4 days |
| Translation & Localization Agent | Developers running batch translation with glossary enforcement, a quality gate, and human review of low-confidence segments | 1–2 days |
| Recommendation Agent | Developers producing grounded, explainable recommendations from user context and a catalog, with a relevance gate and guardrails | 2–4 days |
More templates coming
This library grows as new patterns are validated in the field. Each template starts as a walkthrough, gets tested with real customers, and graduates here when the pattern is repeatable. Have one to contribute? Open an issue on GitHub.
Borrow, don't build
These templates are starting points — the canonical reference for the products they're built in stays with Microsoft. Build against the official docs, not our paraphrase of them.
- Copilot Studio documentation — the build-to-govern reference for every low-code template
- Agent Library — Microsoft's own catalog of production-ready agent templates
- Microsoft Foundry documentation — the canonical reference for every pro-code blueprint
- Foundry Agent Service — build, deploy, and run pro-code agents with tools and orchestration
Thanks — glad it helped. Spotted something to improve anyway? Suggest it →
Sorry this didn't land. Tell us what broke → so we can fix that rung of the ramp.
Found an issue or want to suggest an improvement? Open a GitHub issue · Star on GitHub