Today’s plays › All plays › Coaching
Stop picking calls to review at random: have AI send managers the best and worst calls each day, and why
Picked and checked by Olga Pearson, SalesUpLevel
Get tomorrow’s play in your inbox. Free ↓
Reported in Attention's case study: Engine managers save 2 hours a week each on call reviews. Across 500 calls a week, sellers save 30 minutes a call, or 250 hours a week (from CRM write-back and follow-up emails), and key data is 80% more accurate.
"Go review 10 calls today" sounds like coaching. With thousands of calls a week, it is a lottery.
Case study: Engine (customer) and Attention (vendor). Engine has more than 200 sellers and thousands of calls a week, so managers used to pick calls at random or on a hunch. Now daily alerts tell managers and sellers their best and worst calls and why. Engine tracks average call score by team week over week (if the average demo call is a 2.5, that is where enablement goes), and game film sessions use the five calls the tool recommends. Calls with strong negative sentiment go to a churn-risk channel with the summary.
By Mollie Bodensteiner (Engine), in an Attention case study, SVP, Operations, Engine · Case study (Attention customer story) ↗ · Shared by Attention on LinkedIn Feb 19, 2026 (page has no date)
Run it before your 10am
- Write a short scorecard for each call type, starting with demos.
- Have AI score every call and send each manager a daily list: top calls, bottom calls, and the reason for each.
- Review only those calls, and pick five for a weekly game film session.
- Track the average score by team and call type week over week.
- Send calls with strong negative sentiment to a churn-risk channel with the summary.
- Tell the team it is a coaching tool, not surveillance, before you launch.
Here are today's scored [demo] call summaries for my team: [call, seller, score, summary]. Our scorecard: [3 to 5 criteria]. Pick the 2 best and 2 worst calls. For each: one sentence on why, with a timestamp or quote, and one coaching question I can ask the seller. Flag any call where the customer sounded at risk of leaving.
If you lead a team
- Replace 'review 10 calls' with the daily best and worst list.
- Track average score by call type and put enablement where it is lowest.
Watch out for
- Vendor case study; Attention is selling against Gong.
- The 250 hours is seller time from CRM write-back and follow-up emails, not from coaching. The case gives no win rate change for the coaching.
- Recording calls needs consent rules in each place you sell.
- What I'd have my team try first: for one week, each manager reviews only the 2 best and 2 worst calls the AI picks, then compare notes on Friday.
Source: Attention customer story 'Why Engine stopped paying for features they didn't need' (shared by Attention on LinkedIn Feb 19, 2026 and again Jun 9, 2026; page has no date). Case study: customer Engine, vendor Attention. Results are as reported; we did not measure them.
Get tomorrow’s play before your first call
Free. One sourced play a day.