The company was cold-calling its own paying customers. We stopped that in week one.

BijliRide (BIJLIRIDE PRIVATE LIMITED) rents electric two-wheelers to gig-economy drivers, the people delivering food and packages across India’s metros. It is Hyderabad-led, with riders in Bengaluru, Delhi, Mumbai and Pune. Growth at that pace produces a specific kind of mess: more customers than any spreadsheet was built for, spread across more tools than anyone can reconcile by hand.

The customer record lived in three places at once, a set of spreadsheets, a calling platform used by the telecaller team, and the app’s own database. None of the three agreed with the others. That is not a reporting inconvenience; it is a daily operational tax.

• [01]Around 600,000 “leads” and over 138,000 registrations, with heavy duplication and no reliable single view of a person.
• [02]Nearly half of the “leads” were already in their own customer base. More than 1,500 active, paying riders were still being cold-called by the outbound team.
• [03]The work between systems, reconciling records, updating them after every call, keeping the lists clean, was done by hand, or not at all.

Mindlyft did not replace their tools or ask them to migrate. We built the execution layer between the tools they already ran, inside their own accounts.

• [01]One source of truth. We reconciled the three systems into a single, deduplicated view of each person, matched on phone number, and kept it current automatically.
• [02]Active customers, off the cold list. The system now removes anyone who is an active paying rider from the outbound cold-call lists, so the team stops burning goodwill on people who already pay.
• [03]Every call becomes tracked work. Call outcomes turn into the follow-ups, updates, and next steps that were promised, with a human approving anything a customer would see before it goes out.
• [04]Their data stays theirs. Mindlyft reads from their systems and never overwrites their master records. Every change is logged and reversible, and none of it touches sensitive KYC data.

The manual record-update work that used to eat the team’s day dropped by 70 to 80 percent, and overall administrative load fell about 30 percent. The outbound team stopped calling their own paying customers. The record everyone works from is now one record, not three.

BijliRide is Mindlyft’s paying design partner: a real company, in production, paying for the work, not a pilot slide.

BijliRide’s tools were spreadsheets and a dialer. Yours might be Salesforce or HubSpot, Gong, Slack and Jira. The shape of the problem is identical: work gets promised on a call, and then it either happens by hand or it slips. That is the exact gap Mindlyft engineers away, inside whatever stack you already run.

Your numbers will not be BijliRide’s numbers. They depend on your stack and where you start. What carries over is the approach: engineered, human-approved, reversible, and built where your data already lives.