Demo mode. Write-backs to production data are blocked.

About the intelligence

ZM
OnlineLast synced --:--

About the intelligence

How the scores are produced, what the AI does, and what every term means.

The numbers come from your own inputs. The four engines compute the scores deterministically; the language model only writes the explanation in plain words. The predictions sharpen as real field history accumulates. Nothing here is a black box: tap Why? on any score to see the inputs and values behind it.

What it does

  • Computes every score deterministically from your own field inputs.
  • Shows the working: the top factors with their real values, on every score.
  • Phrases the brief and the reasons in plain language (Haiku), with a deterministic fallback.

What it does not do

  • Invent labels. Every term is defined below and grounded in the SRS.
  • Produce a number or a prediction with the language model.
  • Move float. The Float Rebalancer System is the system of record.
Source: Float Rebalancer System

Glossary

Every score the platform shows, with its definition, inputs and the SRS reference it is grounded in.

Priority score

FR14 (value-based route advice)

How soon an outlet should be visited, built only from the four FR14 drivers. Higher means visit sooner.

Last visit date35%

How overdue a visit is against the target cadence.

Float levels30%

The linked agent's float against a healthy balance.

Sales performance20%

Sales this period against the outlet target.

Recruitment targets15%

Territory GA recruitment against target.

0 to 100. Higher visits first.

Time-to-stockout

FR45 (float demand forecast). Source: Float Rebalancer System.

Projected hours until an agent's float hits zero at the current transaction velocity.

Float level

Current float balance read from the Rebalancer.

Transaction velocity

Daily burn rate derived from recent float readings.

Hours of cover. Under 12h is red, under 36h is amber.

Authenticity score

FR23 (proof-of-visit authenticity)

How trustworthy a visit looks, from the FR23 checks. Low means hold for review.

GPS vs outlet

Device location within 150m of the outlet master. -25 if not.

Photo location

Photo GPS within 150m of the outlet. -20 if not.

Dwell time

Time on site plausible for the outlet. -45 if too short.

Proof photo

Photo not reused from an earlier visit. -60 if reused.

QR / OTP handshake

QR scanned and OTP confirmed. -30 if missing.

0 to 100, starting at 100 and deducting per failed check. Below 60 is flagged.

Conversion score

FR51 to FR59 (acquisition funnel)

Likelihood a captured lead activates, from interest level and history.

Interest level80%

Interest expressed at capture, 1 to 10.

Customer type15%

Strategic types (master agent) score higher.

Follow-up

An overdue follow-up discounts the score.

0 to 100. Higher is more likely to convert.

Territory forecast

FR60 to FR62 (territory gap forecasting)

Projected period total for a territory against its GA target, from the current run rate.

Run rate

Achieved to date over the share of the period elapsed.

Target

The GA or Mobile Money GA target for the period.

A projected count. A gap over 20% of target is a severe miss.

Churn risk

FR58 (activation and churn-risk)

How likely an outlet is going dormant, from recent activity.

Last activity

Days since the outlet last transacted.

Outlet state

A dormant flag carries a high base risk.

0 to 100. 70 and above is treated as going dormant.