Copilot Studio SME

Role Summary

Avanade India’s Data & AI practice is looking for an Agentic AI & Copilot Studio SME to lead client enablement and adoption for agent-based AI solutions built on Microsoft Copilot Studio and the wider Microsoft agent ecosystem. This is a training-led, client-facing role: the primary focus is designing and delivering structured skilling - instructor-led workshops, hands-on labs, train-the-trainer programmes and champion communities - for large organisations building and scaling agents.

What distinguishes this role from a platform-training role is agentic depth. You are expected to understand how agents actually work - orchestration and reasoning loops, tool and function calling, grounding and retrieval, memory and context handling, multi-agent patterns, evaluation and guardrails - and to teach Copilot Studio in that context rather than as a feature tour. You will build reference agents yourself, advise on live engagements, and be the person client teams escalate hard agent-design questions to.

Key Responsibilities

• Own the design and delivery of agentic AI and Copilot Studio skilling programmes for large enterprise audiences - curriculum design, instructor-led sessions, hands-on labs, self-paced content, assessments and certification pathways.
• Run train-the-trainer and champion-enablement programmes so client teams can sustain and scale agent development independently after the engagement.
• Teach agent design as a discipline, not a tool walkthrough - task decomposition, orchestration and reasoning patterns, tool/action design, grounding strategy, memory and context, human-in-the-loop checkpoints, failure modes and cost/latency trade-offs.
• Tailor content and depth to distinct audiences - business makers, pro-code engineering teams, and business and IT leadership.
• Build reference and demo agents in Copilot Studio - topics, generative answers, actions and connectors, agent flows, knowledge sources and multi-agent handoffs - to anchor training and seed client use cases.
• Design and maintain lab environments, sandbox tenants and exercise sets, keeping them current with the fast-moving agent-platform roadmap.
• Provide hands-on advisory on live engagements: agent design reviews, grounding and knowledge-source guidance, action/connector patterns, prompt and instruction design, testing and evaluation approaches, and go-live readiness.
• Advise on when to build in Copilot Studio versus a pro-code approach on Azure AI Foundry or an agent framework - and how the two coexist in a single client landscape.
• Establish evaluation and Responsible AI practice for agents: test sets and regression suites, groundedness and quality measurement, content safety, guardrails, human oversight and monitoring in production.
• Advise clients on agent governance and operating models - environment strategy, DLP policies, maker onboarding, publishing and approval workflows, agent lifecycle management, licensing considerations and usage analytics.
• Support use-case discovery workshops - identifying, qualifying and prioritising agent scenarios, and setting realistic expectations on what agents can and cannot do reliably.
• Act as the practice reference point for agentic AI enablement: build reusable training assets, playbooks, agent templates and accelerators, and mentor consultants joining the capability.
• Contribute to presales as the subject-matter expert - enablement approach, effort and duration estimates, demos and client-facing solution walkthroughs.

.

Required Skills & Experience

• 9-12 years of overall experience in software engineering, automation, data or AI roles, with a recent and substantial focus on generative and agentic AI.
• Hands-on experience designing and building agentic AI solutions - multi-step orchestration, tool and function calling, retrieval and grounding, memory and context management, and multi-agent collaboration patterns.
• Up to 2 years of hands-on experience with Microsoft Copilot Studio, covering both building agents and delivering training on the platform.
• Practical Copilot Studio depth: topics and trigger design, generative answers and knowledge sources, actions and connectors, agent flows, publishing across channels (Teams, web, M365 Copilot), and testing and troubleshooting agents.
• Demonstrated experience designing and delivering technical training to large organisations - instructor-led workshops, hands-on labs, train-the-trainer and champion programmes, for both business and technical audiences.
• Working knowledge of the wider Microsoft AI stack: Azure AI Foundry / Azure OpenAI Service, Azure AI Search for retrieval, and how these complement low-code agents.
• Familiarity with agent frameworks and protocols - Semantic Kernel, AutoGen, LangChain/LangGraph, MCP or similar - sufficient to position and compare approaches credibly.
• Understanding of prompt and instruction design, RAG patterns, and agent evaluation - groundedness, task success, regression testing and quality measurement.
• Understanding of Responsible AI in practice: content safety and filtering, guardrails, data boundaries, auditability and human oversight.
• Working knowledge of the Power Platform (Power Automate, Dataverse) and of Microsoft 365 data sources as grounding and action surfaces for agents.
• Comfort with code - Python or C#, plus REST APIs and integration patterns - enough to build custom actions, explain pro-code extensions and debug integrations.
• Strong facilitation and communication skills, with the confidence to hold a room of senior stakeholders as well as a lab of hands-on makers.
• Ability to produce enablement collateral independently - decks, lab guides, quick-reference material and recorded walkthroughs.

Preferred / Good to Have

• Microsoft certifications: Azure AI Engineer Associate (AI-102), Power Platform Fundamentals (PL-900) or Developer (PL-400), or Copilot/agent-related Microsoft credentials.
• Experience with declarative agents, the Microsoft 365 Agents Toolkit, Graph connectors or custom engine agents for deeper extensibility scenarios.
• Experience taking agents to production - observability, cost and latency optimisation, versioning and lifecycle management.
• Exposure to non-Microsoft agent platforms (AWS Bedrock Agents, Google Vertex AI Agents, open-source frameworks) for comparative positioning.
• Experience across multiple industry domains (manufacturing, BFSI, consumer goods, energy, healthcare, or others).
• Experience in a global delivery / SI environment (Avanade, Accenture, or similar) working with distributed teams across geographies.
• Prior contribution to reusable training assets, agent accelerators or practice-building initiatives.

.