Retail Conversation Intelligence Platform: What It Must Do, What to Ask, and How Borentis Is Built

The six jobs a platform has to do

1. Capture on the floor, with consent, offline. The advisor's phone, one tap, a spoken consent line, local recording when the signal drops, sync later, audio deleted after transcription. A platform that needs hardware per store or a live connection will miss the conversations that matter most.
2. Transcribe the way the floor talks. Hinglish, Hindi, English and regional languages, switched mid-sentence, in a showroom with music and a generator. An English-first model produces noise and the scores built on it are fiction.
3. Score against your playbook, with evidence. Not a generic model of a good conversation but the brand's own six moments, each with a line of transcript behind the score, so a store manager can coach from a fact rather than an opinion.
4. Coach from the store's own best conversations. A library of how this store's best advisor handled the price objection in this store's language is the most-used coaching asset a network can own.
5. Recover the walk-out. A customer who left without buying but with a number and a reason becomes a lead with a brief, drafted for a person to send. Nothing goes to a customer automatically.
6. Read the selling score next to operations and footfall. A store with clean audits and low playbook adherence needs a coach; a store that sells well with a poor window needs the VM team. One card per store, read by the area manager, is where the platform earns its name.

Twelve questions that separate a platform from a demo

A vendor that answers all twelve on your own data is selling a platform. A vendor that answers them with slides is selling a roadmap. The buyer's guide at /learn/in-store-conversation-intelligence/ covers what a pilot costs and how long one takes; the shortlist is compared at /learn/best-retail-conversation-intelligence-tools-india/.

1. Record a conversation on my own floor at my busiest hour and show me the transcript. Not a demo clip.
2. Show me the consent line the advisor says, how a customer declines, and what happens to the audio after transcription.
3. Score that conversation against my playbook, not yours. Where is the line that shows the skipped step?
4. Show me coverage by advisor and by hour for a store. What do you do when it falls?
5. Run it on a store phone under ten thousand rupees with mobile data off.
6. Show me the coaching view for a store manager: the most-skipped step this week and three transcript lines.
7. Show me a walk-out becoming a lead, the brief, and the person who sends the follow-up.
8. Show me the one card per store with the selling score, the operations score and footfall together.
9. Which regional languages have you transcribed on a real floor, and can I hear one?
10. What is the all-in price per store per month for my store count, and the year-two price, in writing?
11. What do my advisors see about themselves, and before whom?
12. What leaves with me if I leave: transcripts, scores and leads as CSV and PDF?

How Borentis is built

Borentis is a retail conversation intelligence platform built for retail floors. Borentis Floor does the six jobs above: capture on the advisor's phone with consent per conversation, offline-first; transcription in Hinglish, Hindi, English and regional languages; the Borentis Score against the brand's playbook with the evidence trail; coaching from best conversations; follow-up briefs for a person to send; and the store card. The agents inside it have names so that a store team can talk about them: Saarthi captures and coaches on the floor, Prajna analyses and answers questions across stores, Cadence drafts the follow-up a person sends.

Around the conversation sit the products that read it: BorentisOps, the store operations platform, puts the checklist, audit and visual merchandising score on the same store card as the selling score, so that the area manager sends the right person to the right store. ShopperDNA, which joins camera-based visit signals to the conversation, is on the roadmap. Borentis Remote serves teams that sell over calls and meetings and is a separate product for a separate job. The conversation on the floor is the centre; everything else is read against it.

The platform is designed for the DPDP Act from the consent line up, and the public position on recording customers in stores is at /learn/is-recording-customers-legal-in-india-retail/.

How it compares with the alternatives a retailer already has

• Mystery shopping sees one scripted visit per store per quarter; the platform sees most conversations every day. The comparison is at /learn/mystery-shopping-alternative-india/.
• CCTV and footfall analytics count the walk-in and the walk-out and hear nothing; the platform explains the gap between them. The comparison is at /learn/conversation-intelligence-vs-cctv-footfall-analytics/.
• Retail execution and store operations software checks that the store was set up to standard; the platform checks whether the floor sold to the playbook. Both are needed and they are different jobs, set out at /learn/conversation-intelligence-vs-retail-execution-software/.
• Call-centre conversation intelligence cannot capture a conversation across a counter, at /learn/borentis-vs-gong/.

On the floor

A consumer durables brand with exclusive outlets runs a platform pilot in ten stores. By week three the store card shows the pattern the brand's monthly review had never seen: two stores with the best audits and the worst finance-offer adherence, one store with a poor window and the best conversion in the region. The area manager's next three visits are to coach finance in two stores and to send the VM team to the third. Conversion in the first two stores moves within the month; nothing about the stores changed except who said what to whom.