AI dispatch vs traditional taxi dispatch software: what is different

The difference between AI dispatch and traditional dispatch software is smaller than the sales pitch suggests, and bigger than the sceptics think. Both hold your bookings, prices, drivers and records. The real change is at the front door: who answers the phone and the messages, and what that does to the controller's day. This guide compares the two function by function, with the costs and the honest limits.

What is traditional taxi dispatch software?

Traditional taxi dispatch software is a booking system that a person runs. A call-taker or controller answers the phone and types the job in, and the system holds it, prices it and gets it to a driver.

A traditional taxi dispatch system covers much the same ground from one supplier to the next:

• A booking screen for the office, with addresses, times and notes.
• Zones and fixed prices, or a tariff worked out from distance.
• A driver app or an in-car terminal that shows the job.
• Allocation by the controller, or by the system's own rules, such as the next car in a zone queue.
• Booking records, account customers and reports.
• On some systems, an online booking form or a passenger app.

All of it waits for a person to pick up the phone. When the office is busy, or closed, the call rings out and the customer rings someone else.

What does AI dispatch change?

AI dispatch changes who takes the booking. Software answers the phone or the message, asks the questions your controller would ask, quotes from your price list and saves the job. The rest of the system carries on as before.

The word "AI" gets used for two different things, so ask which one you're being sold. One is the booking side: a phone agent and a chat agent that take bookings. The other is software that suggests, or picks, the driver for each job. A system can have either, both or neither. Our plain guide to what an AI taxi dispatch system is explains both.

How do AI and traditional dispatch compare, function by function?

Most of the system stays the same. Here's where the difference sits:

The last row is the one sales pitches skip. Outside London, section 56 of the Local Government (Miscellaneous Provisions) Act 1976 makes the operator record the particulars of every booking before the journey starts, in the form the council sets. That duty doesn't move to the software, whoever or whatever took the call.

Do you still need a controller with AI dispatch?

Yes. In the AI vs human controller question, the answer is both: the controller's job moves from typing routine bookings to running the exceptions.

On a Friday night the agent takes the airport runs and the pre-books for Saturday. The controller deals with the caller who's lost, the account customer disputing a fare, the driver running late and the job that needs moving to another car. The software doesn't know the ring road is shut tonight unless someone tells it.

Your controller does.

The rules on people don't change either. The Department for Transport's statutory standards ask councils to require operators to keep a register of everyone who takes bookings or dispatches vehicles, each with a basic DBS check, so your controller stays on that register. The same standards say you can outsource booking and dispatch but can't pass on your duty to protect children and vulnerable adults. A safeguarding worry raised on a call has to reach a person in your office.

Where do the costs move?

From wages to software. With a traditional system, covering more hours means paying more people. With AI dispatch, the cost follows the software's price and how many calls and messages it handles.

Here's an example of the night controller cost, worked out from the official minimum rate. It's a floor, not a typical wage.

And one person can't cover all of that alone. GOV.UK sets the weekly limit at an average of 48 working hours, normally over 17 weeks, unless the worker opts out, and night workers must not work more than an average of 8 hours in 24. Paid holiday has to be covered as well: at least 28 days a year for someone on a five-day week. So in practice the nights get shared between two people, or the owner takes the phone to bed. Our guide to taxi night cover without a night controller goes through the options.

On the software side, ask how the price is built: a monthly fee by fleet size, a charge per booking, call minutes, or a mix. Then put it next to the work you're losing now, using one sum: missed calls a day that would have been a job × your average fare × 30. Our guide to missed calls in a taxi office shows how to count them from your phone provider's reports.

What can't AI dispatch do?

Plenty, and a fair supplier will say so. These are the limits of AI dispatch worth knowing before you switch anything on:

• It only knows what you've told it. Your zones, prices, regular customers and road closures have to be in the system.
• It can mishear. A street name, a postcode or a surname. A good agent reads the booking back and asks again when it isn't sure, but listen back to calls in the first weeks anyway.
• It shouldn't handle the hard calls. Complaints, lost property disputes, a safeguarding worry or an emergency belong with a person.
• It can't ring your customers to sell to them without the right consent. The ICO says an automated marketing call needs the person's specific consent to that type of call, and general marketing consent isn't enough.
• It doesn't change the Equality Act. Under section 170 of the Equality Act 2010, refusing a booking because the passenger has an assistance dog, or charging extra for the dog, is an offence. Set the agent up to pass those bookings to you, never to turn them away.
• It brings data work with it. The ICO says that in the vast majority of cases, using AI triggers the legal requirement to carry out a data protection impact assessment, so do one before you go live.

And some callers just want a person. Give them a quick way to reach one.

Which suits your firm?

Which one suits you depends on where your calls go missing.

If you answer the phones yourself between jobs, the gain from AI is on the calls you can't take while you're driving, and overnight. If you have a busy office with good controllers, the gain is on the overflow in the rush and the messages nobody has time to read. If your phones are covered and calls rarely ring out, a traditional system may be all you need, and that's a fine answer.

There's a middle way too: a person in the office in the day, and the agent on the overflow and the nights. Dispatch AI is booking and dispatch software for taxi and private hire firms. We put it on the calls your office can't answer, when the line rings out and overnight, or on your whole booking line if you'd rather. You choose the hours. You can ring the demo line to hear it, and our guide to choosing the best AI taxi dispatch system has a test plan that works on any supplier.