Fork in the Call
Step into a real call where it turned; real outcomes judge your line.
The job
Teaches agents what to say at the moments calls go wrong, judged by counted outcomes from real calls rather than a trainer's opinion or a simulated customer. The winning lines go to both human agents and the AI agent.
The moment
A trainee agent steps into a real call at the moment it turned, and has to choose what to say next.
The trainee types "I'll call you back before 4 today with the adjuster's decision".
The answer joins the widest calm stream: 140 real calls where agents said roughly the same thing. Next to it runs a thin red stream of the 210 calls where agents repeated the policy.
There is no actor and no guesswork, only how real customers reacted.
What it does
- Rebuild forks
- Save the winning branch as a Save Card
- Assign a fork drill
- Propose the move to the AI agent
What you see
Choose-your-own-adventure transcript that ends in a branching river of real outcomes
What it moves
The share of calls where the customer stays calm in the moments after each turning point, for trained agents compared with their own earlier calls, with the number of calls shown.
Built for
- QA, training & knowledge
- Contact-centre operations
- Frontline staff
