Monday CS standup
Accounts with sentiment shift in the last 30 days. Customers whose tone moved meaningfully, before the renewal call gets weird.
Why this is memory-only. Sentiment is not stored anywhere as a field. It lives in the cumulative tone of every call, email, and ticket - and whether it changed. A health score can be gamed and is often stale; the actual shift happens in between the moments anyone writes it down. Only a memory layer that reads every interaction over time can detect that an account's tone moved before anyone noticed.
Side by side: token usage, with and without Paperbase
Memory onlyWithout Paperbase
Not recoverable from ordinary fields
- Scan the CRM for accounts with low health scores.
- Ask CSMs which accounts feel off.
- Skim recent tickets for complaints.
- Rely on whoever happens to have sensed it.
- Miss the account that went quiet without filing a ticket.
With Paperbase
Persistent memory layer
- Ask Paperbase to measure Sentiment per account over the last 30 days.
- Flag accounts where the trend moved meaningfully.
- Pull the interaction that started the shift, quoted.
- Rank accounts by renewal proximity plus sentiment change.
Agent prompt
You are preparing the Monday CS standup. Using Paperbase: 1. Measure Sentiment for each account over the last 30 days. 2. Flag accounts where sentiment shifted meaningfully - positive or negative - versus the prior 30 days. 3. For each flagged account, pull the interaction where the shift started, quoted and dated. 4. Overlay renewal dates and open escalations. Output: - Accounts with sentiment shift, ranked by how sharp the change is. - For each: the shift, the quote that signals it, and what to do this week. - Accounts at renewal in the next 60 days with any negative movement.
Placeholders in {double_braces} are inputs the agent will ask for at runtime. Give the prompt to any agent connected to Paperbase, and the rest grounds in your own memory graph.