An insurance sales network finds out where to focus its coaching with AI Role Play
The company wanted to see how its professionals responded under pressure, not just what they knew. With an AI avatar, every conversation became data by skill and by person.
of the team scored above 90 points in at least one simulation
practice sessions with the AI avatar
points between each professional’s average and best score
team average score
of the sessions started were completed
professionals in the sales network
In this case
An insurance sales network had 24 professionals talk with an AI avatar. Over 121 sessions, 58% of the team scored above 90 points in at least one simulation, and the company got a map of who needed support and in what.
The challenge
In insurance, knowing the product isn’t enough: you have to know how to respond when the customer asks questions and raises objections, and do it with confidence. That skill is rarely visible until the real conversation, and by then the customer has become the training ground.
The company wanted to see how its professionals applied what they know in the middle of a conversation and where it was worth reinforcing, instead of repeating the same training for everyone.
- Check how they respond under pressure, not just what they know.
- Measure objection handling and command of the information.
- Spot who is ready and who needs support before meeting a customer.
- Point managers’ time at the skill that really needs it.
The solution
The 24 professionals talked by voice, in real time, with an AI avatar in sales conversations with questions and objections included. At the end of each conversation they got feedback and could try again before taking that behaviour to a real customer.
The avatar’s questions about the product, with the right information.
Countering the customer’s objections with arguments, without losing the conversation.
Each simulation scored both skills out of 100. In total, 121 sessions were started and 71% were completed: enough to compare each professional with the group, skill by skill.
The results
The team average was 79.4 out of 100. But the useful data was beneath the average: the breakdown by skill showed where there was room to improve, and practice brought potential to light. Each professional’s best score was, on average, 7.5 points above their own average, and 58% of the team scored above 90 points at least once.
This is potential shown through practice, not an improvement attributed to the tool: it shows how far each person can go when they repeat the conversation.
One average, three coaching plans
Behind the average there were very different profiles, and each one called for a different plan:
Very high overall score in both skills. Needs to maintain, not reinforce.
Counters objections very well but is weaker on command of the information. Support goes to the product.
Starts lower and reaches the top by repeating. More practice works for them.
One team average, three different coaching plans.
Find out where each person needs coaching.
We configure the conversation with your product and your objections, and in the demo we show you what data you would get from your team.
How to replicate it with your team
- Choose the conversation. For example, a price objection or a renewal at risk.
- Decide what is assessed. The feedback modules: objection handling, product knowledge, forbidden messages…
- Look at the data by person. The team average says little; the breakdown by skill tells you whom to support and in what.
More information in the case study on gamestrategies.io.
Frequently asked questions
What did the insurance sales network train with AI Role Play?
Two skills in sales conversations with an AI avatar: answering the customer’s questions with confidence and handling their objections. Each simulation scored them out of 100.
What results did it get?
24 professionals completed 121 sessions with the avatar. 58% of the team scored above 90 points in at least one simulation, the average was 79.4 out of 100 and each person’s best score was, on average, 7.5 points above their own average.
Why is the case anonymous?
The company prefers not to appear by name. That is why the case is published as “an insurance sales network”, with anonymised data.
Can the same be done in another industry?
Yes. The scenario is configured with each company’s product, ideal customer and objections, and the feedback assesses the modules you choose.