Concourse & Jira DC Portal

✅ Put Simply

Enable users to ask natural-language questions about Playbooks and receive instant, intelligent answers — powered by an integration with the Ask Concourse.

✅ Why It Matters

Playbooks are dense, detailed, and packed with value — but not all users know where to look or how to navigate them efficiently. The Ask Concourse Assistant can act as a conversational layer over the Playbook Library, allowing users to quickly ask things like:

• “What activities are in the Transformation Strategy playbook?”

• “What should I deliver in the Mobilize stage for Digital Transformation?”

• “What’s the difference between the Operate and Optimize stages?”

This drastically lowers the barrier to engaging with Playbook content, especially for new joiners, pursuit teams, and delivery leads working under time pressure. It supports better discovery, faster reuse, and deeper understanding of PwC’s delivery frameworks — without requiring users to master navigation or terminology.

✅ High-Level Acceptance Criteria

1. Given the user is viewing a Playbook, Framework, or Building Block page,
   When they activate Ask Concourse Assistant,
   Then they should be able to type or speak a natural-language question.

2. Given a question is asked about Playbook content,
   When ACA processes the request,
   Then it should return an accurate, concise, and sourced answer — including links to deeper content if needed.

3. Given the user asks a follow-up question,
   When the context is relevant to the previous thread,
   Then ACA should maintain that context and refine the response accordingly.

4. Given the Playbook content is structured (Playbooks → Frameworks → Activities/Deliverables),
   When ACA returns results,
   Then the assistant should present information with that structure in mind (e.g., grouped by PDM stage).

5. Given some questions may reference scoping or Offering selections,
   When ACA has access to the scoped engagement context,
   Then it should prioritize content relevant to that scope in its responses.

6. Given users may ask ambiguous or incomplete questions,
   When ACA identifies gaps,
   Then it should ask clarifying questions or provide examples of valid queries.

7. Given this is an AI-based interaction,
   When a response is generated,
   Then users should have a way to rate accuracy and flag issues to improve model quality over time.