Skip to main content
The Knowledge Hub is not just a library. Its real value shows up when you connect it to AI roleplays. When a scenario references a page or a source, the AI uses that content to shape the prospect persona and to evaluate the rep’s performance.

​
Why connect them

AI roleplays come to life when the AI persona knows what your company sells, how your team talks about it, and what objections you actually hear. You can describe all of that inline in a scenario, but it gets long, hard to maintain, and impossible to reuse. Putting that content in the Knowledge Hub once and referencing it from every scenario is faster and keeps the whole system in sync.

​
What to put in the hub for roleplays

A few kinds of content are especially useful for scenarios:
  • Product pages - one-pagers describing your product, pricing tiers, and key features. The AI uses these to answer prospect questions accurately.
  • Objection libraries - a page covering the objections your team hears and the responses that work. The AI can draw on these when it plays a skeptical prospect.
  • Call recordings or transcripts - upload these as sources. The AI can mimic the language, pacing, and concerns real prospects raised.
  • Frameworks - your discovery framework, your qualification checklist, your demo flow. These inform the feedback the AI gives to the rep after the call.

​
Attaching content to a scenario

Inside the scenario builder, click Source, then Browse Knowledge Hub to pick the pages and sources you want. When you do, the builder reads that content and builds it into the scenario: the persona’s backstory and knowledge, the scenario context, and the evaluation criteria. The scenario keeps the version of each page it was built from, so every session and every piece of feedback runs on what the scenario holds. Later edits to a page reach the scenario when you update the scenario.
The Choose Content dialog in the scenario builder, listing Knowledge Hub pages, files, and links with checkboxes to select them
This connection is bidirectional in the data model: you can see which scenarios reference a given source from the source’s detail page, which makes it easy to find the impact of an edit before you publish.

​
When Knowledge Hub content changes

Feedback after a roleplay comes from the scenario’s own evaluation criteria and scenario details. When a page a scenario was built from changes, such as new messaging or an updated process, bring the scenario up to date:
1

Find the scenarios built from the page

Open the page in the Knowledge Hub. Its detail view lists the scenarios that reference it.
2

Ask the agent to update each scenario

Edit the scenario and point the agent at the updated page, for example: “We’ve updated our verification process page in the Knowledge Hub. Review it and update this scenario, including the evaluation criteria, so it matches.”
3

Review and publish

Check the changes the agent proposes, then publish the scenario.
For facts that change often, like rates or pricing, build the scenario around the conversation and its checkpoints rather than exact figures. The persona can react to whether an offer sounds strong or weak, and the scenario stays accurate between updates.

​
Letting the builder research the hub for you

Attaching content is not the only way in. While you are building a scenario, you can ask the agent to go and read the Knowledge Hub itself, and it will search the hub and come back with what it found:
  • “Review our Knowledge Hub and recommend a coaching scenario to build.”
  • “Use our scenario building guide in the Knowledge Hub, and build this around the Daniel Frost character page.”
  • “Check the hub for our discovery framework before you write the evaluation criteria.”
The agent searches the hub, tells you what it found, and builds its recommendations on top of it.
Scenario builder chat where the agent has searched the Knowledge Hub and lists the modules it found before recommending scenarios to build
This matters most when the material you want the agent to use is large. Pasting a long brief into a single scenario tends to produce weaker results than putting the same material in the hub and pointing at it, because the agent can pull the specific pages it needs instead of working through everything at once.
Break big documents into small, categorized pages rather than uploading one large file. A separate page per character, per competency, and per guideline gives the agent something specific to retrieve, and gives you something reusable to reference by name when you build.
The agent only reads what is in the hub. It will not research your company on the open web, so anything you want it to know about your business, your frameworks, or your terminology needs a page or a source first.

​
Pages and Programs

Programs can reference Knowledge Hub pages too. When a participant opens a program component and the component links to a page, they see the current published version of that page - not a frozen copy. Update the page once and every participant in every program sees the update.

​
Good practice

  • Keep one source of truth - keep product and pricing details in one Knowledge Hub page and build every scenario and program from it, so each update starts from the same place
  • Use pages for the curated story, sources for the raw material - pages are what you want the AI to reason from; sources are the underlying evidence
  • Publish before attaching - the builder uses the published version of a page, so publish it before you build a scenario from it

​
Getting Help

Questions about using the Knowledge Hub in roleplays? Contact us at [email protected].