Pega is one of those platforms that quietly runs a lot of the world's bank, insurer and telco back offices, and Pega Academy is how most people who work on it learn. The courses, which Pega calls missions, are free. They are properly hands-on, with exercises built around a practice environment, and the System Architect path takes you from case types and data models through to integrations and security. The pace is sensible and the writing is clearer than most enterprise vendor training. The certification ladder is one of the better designed ones I have looked at.
System Architect, then Senior System Architect, then Lead System Architect for developers; Business Architect for analysts; Decisioning Consultant and Data Scientist for the customer decision hub side, with Lead Decisioning Architect sitting above both; plus Robotics System Architect and Customer Service Developer. Most exams are 50 to 60 questions and cost $190 per attempt. The Lead System Architect credential is where it gets serious: a written architecture exam followed by an application design and build exam run by Pega itself, at $750 an attempt. Very few vendor programmes make you build something, and that is why LSAs are valued by the big integrators. On AI, Pega has made a choice I agree with.
Rather than launching a standalone AI badge, the current System Architect exam has a weighted Pega GenAI section covering Blueprint, Knowledge Buddy and Autopilot. Blueprint in particular is changing how Pega projects start, because it generates a first-cut application design from a description of the process, and anyone doing the work now needs to know how to review what it produces. Folding that into the core exam treats AI as part of the job rather than a separate skill, which is how it plays out on real projects. The weaknesses are predictable. Nothing you learn here transfers outside Pega.
The academy never discusses when a process is a bad fit for case management, or what it costs to migrate off Pega once you have years of rules in it. The decisioning tracks teach Pega's Next-Best-Action approach thoroughly, but they are not a general data science education, and I would not treat the Data Scientist certification as evidence of broader ML skill. On price, $190 with no bundled retake is middle of the market, and Pega explicitly does not hand out free vouchers. Compare Mendix and Boomi, which give their certifications away, or Appian at $250. The LSA fee is steep, but it pays for human-marked practical work, and I think it is justified. My take: a well run academy for a platform with a real job market, with sensible AI coverage.
The limit is Pega itself.