Problem and solution
How do we measure whether sales call training is working, without evaluating people?
Leadership wants evidence that training matters, and your reps want to know they are not being graded on practice. Both are possible if you measure behaviours and trends, not individuals.
Practical guidance for call training. Practice with DIRI AI is for coaching, not for rating or ranking people.
What it feels like
You ran the workshops, shared the recordings and set up the practice time, and now someone asks, "So did it work?" The honest answer sits behind a fog: revenue moved, but so did the lead mix, the season and two people's territories. You can say nothing causal, and a figure you invent would only come back to haunt you.
Then there is the other worry. The moment you start collecting numbers on practice, reps will assume they are being watched, and the practice will quietly become a performance. You want information that helps people improve, not a scoreboard that teaches them to avoid trying hard things.
Why it happens
Outcomes arrive late and have many causes
Meetings, deals and revenue depend on market, lists, pricing and timing as well as call skill. A change in results months after training says very little about whether the training caused it.
Measuring too many things hides the signal
A scorecard with fifteen items produces a vague average. It is hard to tell which behaviour changed, so nothing specific gets coached.
Individual scores change behaviour
When people believe practice results feed into reviews, they choose easy scenarios, hold back, or avoid practice altogether. The measurement then damages the thing it is measuring.
Practice scores are not validated measures
A score from a practice session reflects one scenario on one day. It is a useful guide for the person practising, but it is not a reliable measure of real performance, and treating it as one leads to bad decisions.
A method you can use this week
- Choose one behaviour per practice cycle Decide on a single observable thing for the next three or four weeks, for example "asks for a specific time before ending the call". A narrow target can be seen, discussed and coached, where a general target cannot.
- Define it so anyone could spot it Write one sentence: "Counts if the rep proposes two concrete times." A clear definition lets reps self-check and lets you discuss examples without argument about what the standard is.
- Compare attempts over time, not people against each other Look at each person's own first and latest attempt on the same situation. Progress against yourself is meaningful; a ranking of eight people is mostly noise and creates the wrong incentives.
- Track leading indicators before outcomes Early signs include how often the behaviour appears in a quoted moment, how reps describe the change in the debrief, and whether it turns up in live calls they bring to a 1:1. Outcomes can follow later, treated as context rather than proof.
- Ask the reps what changed At the end of the cycle, a ten-minute conversation: "What do you now do differently, and what still feels hard?" Their answer is the most informative data you will get, and it decides the next behaviour to work on.
Drill: The before-and-after check
A twelve-minute exercise for you and one rep, repeated per cycle. You need the same scenario at the start and end of the cycle and a sheet with the target behaviour defined in one line.
- Start of cycle, four minutes: the rep runs the situation and says what they did, in their own words. Note only whether the target behaviour appeared.
- Together, two minutes: write one sentence describing what the behaviour could sound like, for example "Would Tuesday at ten or Wednesday at two suit you?".
- Middle of cycle, three minutes: rep tries again with the sentence. Note whether it appeared and how natural it felt, from the rep's view.
- End of cycle, three minutes: same scenario, no coaching. Compare the three attempts side by side and ask the rep which sounded most like them.
How you know it worked: It worked if the rep can describe the change without prompting and the behaviour appears in their live calls too. A shaky result is not a failure; it tells you the next cycle should keep the same target.
How DIRI AI helps
DIRI AI gives you comparable attempts without any extra tracking work. A rep chooses a situation, sets the difficulty and client type, and speaks aloud to an AI customer, clearly labelled as AI, that reacts to what they actually say. They can repeat the same situation, and their history stays in their account, so the first and latest attempts can sit next to each other in the 1:1.
Each report has a rule-based score from 0 to 100, with AI feedback that explains the result but does not change the number, quoted key moments from the rep's own transcript, the effect of each moment, one concrete next change and better phrasings. For measuring a cycle, ask the rep to read the quoted moments for your target behaviour and describe what changed between attempts. Team plans exist, and a 90-day team pilot for up to ten advisors is available if you want to run one cycle with a small group.
Be careful about what the numbers can claim. The score is a guide for practice, not a validated measure, and an AI customer does not replace real customers or a human coach. Practice is for training and coaching, not for evaluating or ranking employees, and the report must never be the only basis for decisions about people. Use it to start a conversation, and judge the training by what reps say, what you hear in live calls and what happens over a longer period. Voice recordings are off by default.
Try it on a practice call
Free plan: 3 calls a month, up to 5 minutes each, no card. You talk to a customer played by an AI, then read a report that quotes your own words.
Questions people ask
What should I measure first?
One behaviour that you can hear on a call and that matters to the next step, such as asking for a specific time. Measure it for a few weeks, then choose another.
Is it fair to compare reps on practice scores?
It is better not to. A practice score reflects a scenario on a given day, and comparisons push people toward safe choices. Compare each person with their own earlier attempts.
How long before I can expect to see a change?
Behaviours can shift within a few weeks of focused practice, but connecting that to business results takes longer and involves many other factors. Treat results as context, not proof.
Can the report be used in performance reviews?
It should not be the only basis for any decision about a person, and it is designed for training and coaching. Keep it separate from evaluation, or reps will stop practising honestly.