Shopper intelligence
What Is ShopperDNA? Visit Intelligence That Joins the Camera to the Conversation
ShopperDNA is a product on the Borentis roadmap that joins what the store's cameras see to what the advisor's consented conversation says, so a visit is understood as a whole: the wait before the greeting, the display that held attention, the group that came together, and then the conversation that explains what they wanted. It works in events, not identities. This guide explains what ShopperDNA would count, how the join works, and what it deliberately does not do.
The gap it is built for
A store that captures conversations knows why the customers who talked to an advisor bought or left. It does not know about the ones who never got an advisor. On a Saturday afternoon in a durables store, that can be a third of the footfall: the family that browsed for six minutes, found nobody free, and walked out. No transcript exists because no conversation happened.
Cameras see that family. They cannot hear the others. ShopperDNA, on the roadmap, is the layer that puts the two next to each other so the store's week reads as visits, not just conversations.
The events it would count
| Event | What it means | Why a store manager cares |
|---|---|---|
| Footfall | Entries by hour and zone | Staffing by hour instead of by guess |
| Group size | How many came in together | A family of four in the TV zone is a different sale from a single visitor |
| Wait time | Minutes from entering a zone to an advisor approaching | The most direct measure of an understaffed floor |
| Unattended walk-outs | Visits that left without an advisor contact | The lost sales no conversation records |
| Display hotspots | Where visits linger, by zone and hour | Whether the new launch display is being seen |
| Attended visits | Visits anchored to a consented conversation | The denominator that makes coverage honest |
How the visit joins the conversation
- The advisor taps to record after consent, as they do today in Borentis Floor. That tap has a time and a zone.
- The camera stream for that zone has already been counting events. The visit in progress at that time in that zone becomes the attended visit for that conversation.
- The two records are stored side by side: the conversation with its transcript and scores, the visit with its wait time, group size and dwell. Neither contains a face, a name or a description of a person.
- Visits with no tap are the unattended ones. They are counted, timed and reported by store and hour, and nothing more is known about them.
What it refuses to do
- No facial recognition, and no matching a visit to a previous visit. Every visit is new.
- No estimates of age, gender or emotion. The product does not look at people; it counts events in zones.
- No profiles. Identity exists only where a customer gave a phone number for follow-up inside the conversation, and that lives in the conversation record, not the visit record.
- No live monitoring of advisors. Wait time is a store measure, reported by hour, not a stopwatch on a person.
What a store would read each week
Who visits, by group size and hour. What they came for, from the conversations that were captured. How long they waited before an advisor arrived. How many left unattended, and at which hours. Which displays held attention and which were walked past. Set next to conversation scores, this tells a store manager whether a low conversion week was a selling problem or a staffing problem, which is a different fix.
What exists today
ShopperDNA is on the roadmap. What runs today is the conversation layer: Borentis Floor captures consented conversations on the advisor's phone, scores them against the playbook, clusters objections and drafts follow-ups a person sends. ShopperPersonas clusters those conversations into the customer types who actually walk in. Retailers who want the visit layer can prepare by mapping camera zones to selling zones and by getting coverage on the conversation side high, because the join is only as good as the taps that anchor it. The guide on how to use CCTV for retail analytics covers what existing cameras can and cannot do on their own.
Frequently asked questions
Is ShopperDNA facial recognition?
No. It counts and times events in zones and never identifies a person. It does not estimate age, gender or emotion. Identity exists only where a customer gave a phone number for follow-up in the conversation.
When will ShopperDNA be available?
It is on the roadmap and not yet a shipping product. Borentis Floor and ShopperPersonas are available now.
Does it need new cameras?
The intent is to work from existing CCTV where the feed can be accessed, with camera zones mapped to selling zones. Specifics will be published when the product leaves the roadmap.
What is an unattended walk-out?
A visit that entered a selling zone and left without any advisor contact, so no conversation was captured. ShopperDNA would count these by store and hour so the size of the gap is visible.
Related reading
- ShopperDNA, the product page
- What is multimodal AI for retail?
- How to use CCTV for retail analytics
- What is shopper intelligence?
Where Borentis applies this
- Walk-in Recovery: The customer who left is still yours.
- Execution Scorecards: See the floor before the P&L does.
- Unmet Demand Signals: Demand for what you did not have.
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.