Story Shapes
Reads the plot of every conversation and finds the line that writes tragedies.
The job
Sorts conversations by the shape of the customer's experience: rescued, crashed, getting worse or steady. Finds the exact agent or bot lines where things turn, ties each to its cost, and bans or promotes lines in saved replies and the AI agent's rules.
The moment
The app reads each conversation like a story and sorts it by shape: rescued, crashed, getting worse, or steady. It then finds the lines where the story turns.
In refund conversations, the line that turns 38% of them into tragedies is "as per our policy". Conversations containing it escalate 2.4 times as often.
With one approved click, the QA lead removes the line from the agents' saved replies and from the bot's rules.
A month later, refund conversations that end badly fall from 41% to 22%. The team's monthly short story now ends with the line the night team uses to rescue customers.
What it does
- Ban or promote a turning line
- Push into macros and AI-agent guardrails
- Assign coaching
- Record the month's mix
What you see
Story-shape glyphs, a shape-mix shelf per team and bot, a turning-lines panel, and a monthly team short story read at calibration
What it moves
The share of conversations that crash or keep getting worse, by request type and team, and how often conversations containing banned lines escalate.
Built for
- QA, training & knowledge
- Contact-centre operations
- Frontline staff
- CX leaders & VoC
