Conversation intelligence
In-Person Sales Conversation Analysis: What to Analyse, How, and What It Changes
In-person sales conversation analysis is the work of turning what was said between an advisor and a customer into something a sales organisation can act on: which moments of the pitch were present, whether the offer was quoted correctly, which objection ended the conversation, which competitor was named, whether a number was taken, and how the conversation ended. Done by a manager with a notebook it covers a handful of conversations a week; done by AI on every consented conversation it covers the network every day. This guide sets out what to analyse, the three ways to do it, what each costs in an area manager's week, and the decisions the analysis changes.
The eight things to analyse
| Question | What it changes | |
|---|---|---|
| Moments present | Which of greet, need, demonstrate, offer, objection, close happened? | The coaching topic for that advisor and store |
| Offer quoted | Was this month's price, exchange or finance stated as published? | Mis-selling risk and launch execution |
| Objection | Which objection ended it, if unconverted? | Whether the store, the brand or the product owns the fix |
| Competitor | Was a rival named, and how was it answered? | Market intelligence and the competitor-defence script |
| Need discovered | How many need questions were asked before the pitch? | Whether advisors sell the product on offer or the right product |
| Questions asked by the customer | What did the customer want to know? | Unmet demand and the FAQ the training never covered |
| Number taken | If no sale, was a number taken with a reason to call? | Whether the lost sale can be recovered |
| Outcome | Bill, booking, number, or walk-out? | The denominator for everything above |
Three ways to do it
The third method is retail conversation intelligence, defined at /learn/what-is-retail-conversation-intelligence/. The metrics it produces and the reports that use them are at /learn/in-store-conversation-analytics/, and the plan to a baseline at /learn/how-to-measure-in-store-sales-interactions/.
| Method | Coverage | An area manager's week | What it misses |
|---|---|---|---|
| Manual review of observed conversations | A few a week, in stores the manager visited | Two store visits, a notebook, a theory | Everything the manager did not see, which is almost everything |
| Sampled review of recordings | Ten to twenty a week, listened to | Three hours of listening, honest notes on a sample | The pattern across a region; the busiest hour |
| AI scoring of every consented conversation | Most conversations, every day | Fifteen minutes per store on the exceptions and three transcript lines | Conversations the advisor did not record; coverage has to be watched |
The decisions it changes
- Which two stores the area manager visits this week, and with which three transcript lines.
- Which step the training team teaches next, because it is the one skipped across a region.
- Whether the offer is the problem or the quoting of it.
- Which competitor the brand team hears about this week rather than next quarter.
- Who gets a call tomorrow: every walk-out with a number and a reason.
On the floor
A furniture chain's regional head reviews the analysis of a month of consented conversations. The surprise is the need questions: the best store asks three before showing anything, the median store asks none and shows the sofa on offer. Conversion tracks the question count. The coaching note is one line and a recording of the best store's advisor asking 'what is the room like?' in Hinglish.
Frequently asked questions
How do you analyse an in-person sales conversation?
Against a written standard: which moments of the pitch were present, whether the offer was quoted correctly, the objection, the competitor, the need questions asked, the customer's questions, whether a number was taken and the outcome. Manually that covers a few conversations a week; AI scoring of every consented conversation covers the network daily with a transcript line behind each score.
Do you need to record customers to analyse conversations?
To analyse more than a sample, yes, with consent under the DPDP Act. Manager observation and sampled listening give a few conversations a week; the pattern across a region needs most of them.
What should be analysed first?
The offer quoted and the number taken: the first is where mis-selling and lost sales hide, the second is the cheapest conversion lever in assisted retail.
Related reading
- In-store conversation analytics
- How AI can analyze in-store sales conversations
- How to measure sales conversation quality
- What is face-to-face conversation intelligence?
Where Borentis applies this
- Objection Intelligence: The reason they did not buy, in their own words.
- Competitor Defence: Hear the rival the moment your customer names them.
- Coaching from Best Conversations: Your best advisor, teaching everyone.
Borentis is the Agentic Operating System for Customer Interactions, built for Indian retail floors: consented one-tap capture on the advisor's phone, every conversation scored against your playbook with the evidence behind every number, leads created when a number is heard, and coaching from your own best conversations.