Last verified: October 11, 2026
What You Can Now Do
Context Memo now writes your chat starters for you: you can see suggested chat starters in the dashboard that are automatically generated from your brand profile and validated against the product index before they appear. Instead of staring at an empty prompt field and guessing which question will return something useful, you get a short set of starters grounded in what your brand profile says about your mission, ICP, differentiators, and proof, and filtered against what the index can actually answer. That changes one task: you no longer configure starters as a setup step before the chat is worth using. You open the chat and start working.
Where It Is in Context Memo
Open your dashboard and go to the chat interface. Suggested starters render above the input field as soon as the page loads, pulled live from your current brand profile. There is no separate settings screen to visit and no toggle to enable. If your brand profile changes, the suggestions on that screen change with it.
How to Use It
- Open the chat in your dashboard. The suggested starters appear above the input field, scoped to your brand profile rather than generic examples.
- Click a starter to run it. The chat answers using your indexed product content, so you can see immediately whether the underlying material supports the question your buyers are asking.
- If the suggestions look thin or off-target, open your brand profile and fill the gaps: positioning, ICP, differentiators, proof points. Return to the chat and the suggested starters reflect the updated profile.
- Treat a starter that returns a weak answer as a signal, not a bug. The starter passed index validation, which means something adjacent exists but the coverage is shallow. That is a memo topic.
- Use the starters as a cross-check against your scan data. When a suggested starter maps to a prompt your buyers are already running in ChatGPT, Claude, Gemini, or Perplexity, you have a direct line from your own positioning to the answers models are producing.
Why We Built It
Customers told us they were hand-crafting chat starters from scratch, which meant doing configuration work before they got any value from the chat at all. The brand profile already held the material those starters should have been built from, so the manual step was duplicate effort: writing down, in prompt form, what the profile already stated. Worse, hand-written starters were not checked against the index, so people wrote questions the index could not answer yet and concluded the chat was weaker than it was. Generating from the profile and validating against the index removes both problems in one pass.
What It Does Not Do Yet
Suggestions are derived from your brand profile, which means the quality of the output tracks the completeness of the input. A sparse profile produces fewer and more generic starters. There is no manual authoring, pinning, or reordering of starters in this release: if you want a different starter, the lever is the brand profile, not the chat screen.
Index validation is a filter, not a quality score. A starter appears because the index holds relevant content, not because that content is deep enough to produce a citation-grade answer. Use the answer itself to judge depth, then publish against the gaps you find.
This also assumes you have content in the index to validate against. Brands early in their build, with little published material on their own domain, will see a shorter list until foundational content exists. In that situation the sequence is the same as it has always been: establish the content and domain authority first, then work the AI visibility layer on top of it.