Last verified: October 7, 2026
What You Can Now Do
Context Memo now gives you a dedicated Simulator page with a runs table that shows the history of all simulation executions, separate from the Prompts section. Every run you execute is recorded as a row you can come back to, so a simulation is no longer a one-time look at a model's answer that disappears when you close the window. The practical change: you can treat simulation as a repeatable check rather than a spot check. Run the same prompt set after you publish a memo, after a competitor changes their positioning, or at a fixed interval, then open the earlier run and read the two against each other instead of relying on memory or a screenshot in a Slack thread.
Where It Is in Context Memo
The Simulator is its own page under Research. It was previously a pop-up modal launched from the Prompts page; that modal is gone. The page now opens with the runs table listing your simulation executions, and you start new runs from the same screen. Prompts stays focused on prompt management, and simulation history lives where the rest of your research work sits.
How to Use It
- Open Research → Simulator. The runs table loads with your simulation executions listed as rows.
- Start a run from the Simulator page. It appears in the runs table as a new row once it executes, so you do not need to keep the screen open to know it was recorded.
- Open a row to review that run's output. You see what was simulated and how the models answered for that execution, held as a fixed record rather than a live view that changes under you.
- Re-run after you publish. Execute the same simulation once a memo is live on your domain, then open the earlier row and compare the answers to see whether the model's description of your brand moved.
- Use the table as your audit trail. When someone asks what AI models were saying about your category last quarter, point at the run from that date instead of reconstructing it.
Why We Built It
Customers told us they had to reach the simulator through a pop-up on the Prompts page and had no way to view or manage their simulation history in one place. That cost them the thing simulation is actually for: a before-and-after. A run would show something worth acting on, the modal would close, and the evidence went with it, which meant the next person asking "did the memo change the answer?" got an opinion instead of a record. Promoting the Simulator to its own page with a persistent runs table makes the history reviewable by anyone on the team with access, which also matters when AI visibility work has to be reported upward rather than just felt.
What It Does Not Do Yet
This release is about history and management. It does not change what the Simulator executes or how it scores an answer, so your existing interpretation of results carries over unchanged.
Two limits worth knowing before you build process around it:
- Comparison is manual. The runs table gives you the record of each execution. Reading two runs against each other is still something you do by opening both, not a diff view that highlights what changed between them.
- No launch point inside Prompts. Because the modal was removed, simulation now starts from the Simulator page. If your team had a habit of running a quick simulation while editing a prompt, that path has moved.
For teams that need alerting on changes in how models answer, scheduled scans and Slack alerts remain the mechanism for that; the runs table is a research record you go and read, not a monitor that pages you.