Written by the AgileFever editorial team. Program details were checked against the live program pages in October 2026. We run the course discussed here. Course content comes from its program page, and the roadmap steps are our own suggestions.
Most architects who ask about this move are not short of architecture skills. They have signed off designs for years. What changes is that AI components bring decisions with no settled playbook in most companies yet. The roadmap below is organised around those decisions, and it uses our GenAI & Agentic AI for Solution Architects course as one of the steps. We sell that course, so we say where the course ends and your own practice has to start.
What you already have
| You already do | What changes with AI | Where the course covers it |
|---|---|---|
| Choose platforms and vendors | Compare model vendors on cost, latency, data residency and lock-in (Azure AI Foundry, AWS Bedrock, Claude, and agent frameworks) | Module 1 |
| Write architecture documents and answer RFPs | Use AI to draft them, then review what it produced | Module 2 |
| Design data access and search | Retrieval pipelines: chunking, embedding model choice, vector databases, hybrid search and re-ranking | Module 3 |
| Design service orchestration | Agent orchestration, tool use and human approval points | Module 4 |
| Plan deployment and operations | Model serving, versioning, rollback, drift monitoring, retraining triggers | Module 5 |
| Budget infrastructure | Token economics, caching, model tiering, inference cost forecasts | Module 6 |
| Pass security reviews | Prompt injection, tool misuse, data leakage through embeddings and PII | Module 7 |
| Integrate with existing systems | Add RAG or agents to a monolith without a rewrite | Module 8 |
| Get designs approved and adopted | Guardrails, evaluation harnesses, rollout and adoption plans | Module 9 |
10 modules in total
A roadmap in five steps
- Close the Python gap first. The course lists Python for AI as a prerequisite, and it is 6 hours. You do not need to become a developer. You need to read agent and retrieval code well enough to review it.
- Pick one real design problem at your company before the course starts. Module 1 starts a running thread from your own use case, and the capstone is a complete architecture package for it. A real problem makes the course far more useful than a sample one.
- Take the course. It is 16 hours of live sessions, with the capstone presented at the end.
- Use the output on a real review. A cost model, a security review and a vendor comparison have value only when someone with authority reads them.
- Read the field. The course gives you the frameworks. Production practice changes quickly, so follow vendor documentation for the platform you actually use.
What a course does not give you
Production experience. You will not have run an AI system under load, handled an incident or lived through a model change. Building a track record means putting a small AI component into a real system, even a low-stakes one, and writing up what happened. If you are looking at career moves, that record matters more than the certificate.
Course facts
| Item | Detail |
|---|---|
| Fee (US) | $450 |
| Format | Live virtual, 16 hours, 10 modules |
| Next cohort listed | November 30 to December 2, 2026, Monday to Wednesday, 9:30 AM to 1:30 PM Eastern |
| Prerequisites | Python for AI (6 hours); working knowledge of software architecture; comfort with a cloud platform (AWS, Azure or GCP); no prior GenAI or LLM experience required |
| Capstone | A complete architecture package for a real use case from your own company, presented at the end |
| Certificate | AgileFever certificate on completion |
Additional discounts are available. Contact us to know more.
For the interview side of the move, read interview questions and prep. For the difference between the two roles, see Solution Architect vs AI Solution Architect, and for the technical patterns, architecture patterns for agentic AI systems. The program page is GenAI & Agentic AI for Solution Architects.
Questions people ask
Do I need hands-on AI experience before the course?
No. The page says no prior GenAI or LLM experience is required. It does expect Python for AI, architecture fundamentals and comfort with one cloud.
Will this make me an AI solution architect?
It gives you the decision frameworks and a documented architecture for your own use case. The role itself comes from applying them in real projects.
Are sessions recorded?
Yes. Live sessions are recorded.
