13.2 Pega GenAI Autopilot & Pega GenAI Coach
Key Takeaways
- Pega GenAI Autopilot covers the generative AI features available at design time in App Studio, suggesting case types, life cycles, data model fields, picklist choices, and personas.
- Autopilot can populate an empty view with suggested fields based on the case type, stage, and step names, generate sample records for a data object, and draft message content for a Send email step.
- Autopilot also provides a conversational assistant with stored, searchable conversations and links to Pega Knowledge resources.
- Pega GenAI Coach is a run-time mentor: it gives context-aware suggestions and guidance to end users working a case, including summarizing case data so a user can review progress and recent changes.
- An architect configures Coach by editing its definitions, instructions, and data sources so its guidance matches the organization's objectives.
13.2 Pega GenAI Autopilot & Pega GenAI Coach
The Pega GenAI domain names two capabilities that are easy to confuse because both offer AI-generated guidance. The distinction is when each one acts.
| Pega GenAI Autopilot | Pega GenAI Coach | |
|---|---|---|
| When it acts | Design time, while the application is being built | Run time, while a user works a case |
| Who it serves | Developers and low-code authors in App Studio | End users — caseworkers, agents, service representatives |
| What it produces | Suggested case types, life cycles, fields, personas, sample data, message content | Contextual advice, guidance, and case summaries |
| Who configures it | Enabled and used by the development team | Configured by an architect against organizational objectives |
1. Pega GenAI Autopilot
Pega GenAI Autopilot™ describes all generative AI features available at design time. Its purpose is to accelerate low-code application development, and in Pega Platform '25 Pega frames it as pairing App Studio accelerators for workflows and data with a conversational assistant trained on Pega technical documentation and developer training.
Design-time Suggestions
Working through a realistic example — a team building a home-loan application:
- Case type names. When you create a case type, Autopilot suggests names based on the application name.
- Workflow. Autopilot generates a suggested workflow to kickstart development. You then review and modify the suggested stages, processes, and steps.
- Data model. Autopilot proposes fields for the case type's data model, including suggested choices for a picklist field.
- Personas. Autopilot suggests the users, or personas, who interact with the case type.
- Views. When you author an empty view in the Step configuration pane, Autopilot can populate the form with suggested fields based on the names of the case type, stages, and steps.
- Smart shape guidance. Autopilot offers shape-specific help — for example, when you use a Send email step, Autopilot can generate the message content.
- Sample data. Autopilot generates sample records for a data object to make testing easier, and can fill out a form with sample data when you run the case type.
- Insights. You can use natural-language prompts to generate a recommended Insight for analysing application data.
The Conversational Assistant
Alongside the in-context suggestions, Autopilot provides an assistant in App Studio with:
- Detailed information and steps — comprehensive guidance on creating or editing application elements.
- Conversation storage — conversations are retained for future reference.
- Search — you can search through stored conversations.
- Resource links — links out to Pega Knowledge resources.
Two Facts the Exam Rewards
- Review obligation. Pega states plainly that it is not responsible for the content of Autopilot suggestions and that AI-powered recommendations should always be reviewed. Any answer that has a developer accept generated output unexamined is wrong.
- Enablement. In Pega Infinity '23 and Pega Infinity '24, Autopilot features are inactive by default and require additional configuration to enable. "It should have worked out of the box on our '24 environment" is a misconception, not a defect.
2. Pega GenAI Coach
Pega GenAI Coach™ acts as a run-time mentor, helping users achieve optimal outcomes from processes. It is an AI-driven capability in Pega Platform that enhances user interactions and decision-making by providing intelligent, context-aware suggestions and guidance, improving the efficiency and effectiveness of business operations.
What a User Experiences
The canonical example: while working on a case, a user who wants to review progress and recent changes can have Coach summarize the case data at run time. Instead of reading a long case history, the user gets a synthesized picture of where the case stands.
More broadly, Coach supplies task-specific guidance and expert advice relevant to the case the user is working — the run-time counterpart to the design-time help Autopilot gives a developer.
What an Architect Configures
Coach is not a switch; it is configured to suit the organization's objectives and the specific needs of its employees. Architects edit its definitions, instructions, and data sources so that the guidance users receive reflects the organization's own policies and knowledge rather than generic advice.
That configuration surface is the reason Coach appears in a system architect exam at all. The blueprint objective is phrased as enabling custom AI-powered capabilities to guide users through complex processes — the emphasis is on the word custom.
3. Choosing Between Them in an Exam Scenario
Read for the actor and the moment:
- "A developer building the case type needs help defining fields and a first-pass workflow" → Autopilot.
- "A claims adjuster opening a four-month-old case needs to understand where it stands" → Coach.
- "Stakeholders with no Pega access need to co-design the application before development starts" → Blueprint (Section 13.1).
- "The application must call a model to summarize text and write the result into a case field" → Connect Rules (Section 13.3).
Those four map cleanly onto four distinct blueprint objectives, and the exam tests the boundaries between them far more often than it tests the internals of any one.
A developer authoring a new Collect information step opens an empty view and wants the platform to propose the fields the form should contain. Which capability does this, and what does it base the suggestions on?
A service representative is handed an escalated complaint case that has accumulated months of correspondence, notes, and audit entries, and must understand the current position quickly. Which Pega GenAI capability is designed for this, and who configures what it says?
A team on Pega Infinity '24 reports that Autopilot suggestions never appear in App Studio and concludes the platform is defective. A second team plans to accept Autopilot's generated workflow without reviewing it, to save a sprint. Assess both positions.