AI Inside
Decisions your fleet didn't know it needed —made automatically
Six AI engines run quietly underneath TaxiDex, making the small operational decisions that used to require a senior dispatcher's gut. The result: lower wait times, higher driver earnings, fewer empty miles.
200ms
Avg dispatch decision
12
Signals per dispatch
94%
ETAs within ±2 minutes
30 min
Demand forecast horizon
Six AI engines
Specialized models, not one giant black box
Each model is purpose-built, audited, and explainable. You can see why a ride was assigned, why a surge fired, why a driver got flagged.
Smart Dispatch
Assigns the optimal driver in under 200ms using 12 weighted signals: proximity, rating, vehicle class, fatigue, earnings target, route compatibility. It is the same trade-off a senior dispatcher makes by gut — scored, ranked, and applied consistently on every job.
- Auto-dispatch rules engine decides how offers cascade by zone, vehicle type, and rating
- Driver pre-assignment keeps VIP and pre-booked work with named drivers
- Controllers override any decision in one click from the drag-and-drop dispatch board
Demand Forecasting
Predicts where ride demand will spike 30 minutes ahead by learning from your own booking history and time-of-day patterns. Drivers get push notifications steering them toward high-demand zones before the rush forms — fewer empty miles, shorter passenger waits.
- Forecasts roll up by the same zones you draw on the live operations map
- Instant broadcasts push or SMS repositioning nudges to all drivers or a chosen few
- Feeds scheduled reports so planners see demand patterns week over week
Predictive ETA
Combines live traffic, historical patterns, and driver behavior into arrival estimates passengers can actually plan around. 94% within ±2 minutes across our 32 active markets — quoted before booking, updated continuously while the car is en route.
- Quoted in the passenger app and the embeddable web booker before the ride is confirmed
- Flight tracking re-times airport pickups automatically when arrivals shift
- Public track-my-ride links share the same live ETA with anyone the passenger sends them to
Fraud Detection
Catches GPS spoofing, fake rides, and chargeback rings before they cost you money. It quietly blocks bad actors without harassing legitimate customers — no extra checkout friction, no false-positive lockouts on your regulars.
- Watches card payments across Stripe and Square, including Apple Pay and Google Pay
- Flags suspicious refund and promo-code patterns before payouts settle
- Every flag lands in the audit log with the exact signals that triggered it
Dynamic Pricing
Surge tuned for sustainability — sharp enough to reward drivers, gentle enough not to scare passengers. Caps, floors, and explainability are built in, so a quote is never a mystery to the customer or to the regulator who asks about it.
- Surge and surcharge rules live inside the enterprise fare engine, scoped per zone
- Multipliers stack with hard caps — and fixed fares always stay fixed
- Every fare is backend-computed and auditable; nothing is improvised on the client
Driver Coaching
Spots habits that hurt driver ratings or earnings — harsh braking, excessive idle time, breaks taken at peak hours — and sends private nudges, not blame. Drivers improve their own numbers; managers only step in when a trend refuses to bend.
- Reads driver sessions, shifts, and daily job sheets for full context
- Nudges arrive as private push notifications inside the driver app
- Managers see anonymized trends in driver and job-earning reports, not call-outs
On a real shift
Where the AI shows up in your console
None of this lives in a separate 'AI tab'. The engines surface inside the screens your team already works in — dispatch rules, surge pricing, and the reports suite.
Smart Dispatch
A rules engine you tune like a senior controller
Smart Dispatch is not a take-it-or-leave-it algorithm. The dispatch rules console lets you decide how its 12 signals are applied on your patch: which zones offer to which vehicle classes first, how long a driver has to accept, when an offer escalates to the next-best candidate, and when a human gets pinged instead.
Pre-bookings and VIP work can be pre-assigned to named drivers days in advance, and the drag-and-drop dispatch board keeps the final say with your controllers — the AI handles the routine 95% so people can concentrate on the exceptions.
Dynamic Pricing
Surge with caps, floors, and a paper trail
Surge rules are scoped to the zones you draw on the live map and run inside the same enterprise fare engine that prices every other ride. Multipliers stack with hard caps so a busy Friday never produces an embarrassing quote, and fixed fares — airport runs, corporate contracts — always stay fixed, no matter what demand is doing.
Because every fare is computed on the backend, each surge fire is logged with the demand signals that justified it. When a passenger, a corporate client, or a regulator asks “why was this ride priced this way?”, you answer with a record, not a shrug.
Forecasting & Fraud
Predictions delivered in reports your team already reads
Demand forecasts and fraud flags land inside the reports suite alongside your booking, customer, job-earning, and call-history reports — not in a science-project dashboard nobody opens. Scheduled reports deliver the forecast to your planners’ inboxes automatically, every morning or every Monday.
When the forecast says a zone is about to spike, instant broadcasts push the nudge to drivers in one click. When the fraud engine flags a booking, the evidence is attached — GPS trace, payment fingerprint, account history — ready for a human decision.
Explainable AI
Every decision, traceable
No black boxes. Every dispatch decision, surge fire, or driver flag is logged with the exact signals that drove it. Auditable, exportable, defensible to regulators.
- Decision logs per ride, per driver
- Bias-detection reports per market
- Model version pinning & rollback
- GDPR-compliant data handling
The same discipline runs through the whole platform: a full audit log records who changed what and when, and the compliance export packages decision history for licensing authorities on demand. If an engine’s new model version underperforms, you pin the previous one and roll back — no redeploy, no downtime.
FAQ
Questions fleets ask before switching on the AI
How long does it take to go live with TaxiDex?
Typical fleets go live in 14 days. We migrate driver data, historical bookings, and customer accounts from Autocab, iCabbi, Cordic, or any legacy system with zero downtime.
Does TaxiDex work for limo and chauffeur fleets?
Yes. TaxiDex Enterprise includes flight tracking, hotel concierge web booker, B2B corporate invoicing, and VIP workflows specifically designed for premium chauffeur and corporate transport operations.
What payment gateways are supported?
Stripe (default global), Worldpay (UK), Trust Payments, Apple Pay, Google Pay, plus regional gateways: STC Pay & mada (Saudi), Paystack & Flutterwave (Nigeria), M-Pesa (Kenya), and any custom merchant on request.
Is there a per-driver fee?
No trip overage. TaxiDex is priced per active driver — about £11 ($14) per driver a month on Starter, £9 ($11) on Professional, and from £7 ($9) on Enterprise. That is at least 20% cheaper than Autocab, iCabbi, Cordic and TaxiCaller, and white-label apps, setup and migration are all included.
Can passengers book without installing an app?
Yes. The TaxiDex Web Booker is an embeddable booking widget for desktop and mobile browsers — perfect for hotel concierge desks, airports, and corporate intranets. One HTML snippet to embed.
See the AI run on a real shift
We'll point our engines at a sample of your historical data and show you the lift — before you commit.