CounterTalk AI
Built at Meta Global AI Developer Hackathon · Oct 26, 2026 · Remote

CounterTalk: your reps hear it at the counter. Your company never does. CounterTalk is enterprise software for companies that sell through field sales teams, such as consumer goods brands and their distributors. It's sold to the company, not to individual reps. Reps, the teams that fix problems, and managers all work in one shared system. Scale. India has an estimated 12–13 million kirana stores. More than 8.8 crore (88 million) small businesses were registered on the government's Udyam portal as of June 2026. Brands and distributors reach most of these shops through field reps who visit in person, and field teams are large: These conversations happen across India's 22 scheduled languages, often mixed with English, and almost nothing from them gets written down. The problem Shopkeepers tell reps what's going wrong, which products they want and which competitors they prefer. Much of it is outside the rep's own work, such as a payment dispute, a short delivery or a competitor's credit terms, so it gets lost. And reps spend evenings typing reports in English. What it does The rep records the visit in the browser on their own phone, or uploads a recording. CounterTalk transcribes it (Hindi, English, Marathi or Hinglish) and writes the report: each complaint, competitor offer and product request with the shopkeeper's exact words, which play back from the recording, plus follow-up tasks. The rep confirms it. Nothing reaches a team before that. Each finding goes to the team that can act on it. Findings with the same cause become one issue across shops, with an owner and a due date. When a team fixes an issue, the rep gets a line to tell the shopkeeper at the next visit. It also writes weekly and monthly summaries that lead back to shopkeepers' words, answers questions typed in Hindi, Marathi or English, and gives the rep a brief before each visit. Who it helps. Reps don't type reports at night, and they walk into each shop prepared. Managers and teams get the full picture from every conversation, not just what reps remembered to report. They can see which problems are widespread, so decisions on credit terms, pricing and products rest on real numbers and the shopkeepers' own words. Built and tried for FMCG. The same playbook fits other goods sold through neighbourhood shops (groceries, general stores and apparel The company learns what to stock, what terms competitors offer, and which problems keep coming up, from thousands of conversations instead of secondhand accounts. Shopkeepers' complaints reach someone who can fix them, and they hear back when it's done. Built for enterprises Each company brings its own products, competitors, teams and routing rules as data, with no code changes. Four roles come built in: reps, team members, managers and admins and more can be setup if required. Each company sets its own teams, and new roles can be added to fit its structure. Consent comes first, by link or paper form, and a refusal blocks recording.