Tianze Wang王天泽
中文/EN
← Offering

AI CRM Decision-first CRM

A CRM for pharma field reps, built around decisions

MVPIndependent · 2026Field-rep app

The problem

A rep’s day is really a string of decisions: who to see, what to discuss, what to record, what to do next. But a CRM is a system of record, filled in mostly because the company checks. And AI suggestions pushed to the front line usually go ignored, because they were never wired into quotas and bonuses.

The approach

Change the CRM’s core object from record to decision. Before the visit, a briefing is already prepared, and two minutes of reading is enough before going to see the doctor. Afterwards the rep talks into the phone, and the system turns it into a profile update, objections, compliance flags and a follow-up draft. The system lines up the next step; the rep accepts or vetoes it, picking a reason for a veto. The record becomes a by-product.

Try it

Try it: you’re the rep, with five decisions today

Tuesday morning · you cover 126 doctors · AI has prepared 5 decisions
  1. Priority visit

    See Dr. Chen in cardiology at City No. 1 Hospital, 10 am

    WhyCity No. 1’s purchases last month ran 18% below its three-month average, and you haven’t answered Dr. Chen’s question about safety data in elderly patients.

    ReadyThree talking points and a compliance-approved safety summary.

  2. Reschedule

    Move this afternoon’s visit with Dr. Li to Thursday morning

    WhyDr. Li added a clinic session this afternoon; Thursday morning is open.

    ReadyA draft message proposing the new time.

  3. Send material

    Send Dr. Wang a patient education leaflet

    WhyLast time she said her patients often mix up when to take their medicine. The leaflet is in the approved library.

    ReadyThe leaflet and a one-line note.

  4. Pause

    Skip District Central Hospital this week

    WhyThe product is still going through the hospital’s formulary process; the committee meets next month, and the last two visits moved nothing.

    ReadyThe freed half-day went to three doctors on the west side.

  5. Report

    Flag to your manager: a competitor started patient assistance at two east-side hospitals

    WhyIt came up in three of your call notes last week.

    ReadyA drafted field report.

10:40. Walking out of Dr. Chen’s office, you talk into your phone:

Dr. Chen read the elderly-patient data. Now he wants the subgroup data for patients with renal impairment. He said he might move to another campus next month.

The system turns it into:

Profile
May move to another campus next month; confirm who takes over
Request
Subgroup data for patients with renal impairment
Compliance
Only the medical-affairs-approved subgroup data may be shared
Follow-up draft
Send the approved subgroup summary within two weeks

You didn’t fill in a single field. The record already exists.

Concept demo. People, hospitals and data are all made up.

Want the full demo, or to try it on your own business? →

How it’s built

  • It decides inside the boundary a CIU sets: which doctors a rep covers and roughly how much each gets this quarter are headquarters’ calls.
  • Results and veto reasons flow back. When many reps veto the same kind of suggestion, some allocation-level assumption has usually gone stale.
  • As in the CIU, numbers come from code; the model understands the question, organizes the steps and explains the result.

Further reading