← All posts

Car finance

AI for car finance brokers in 2026: where it actually helps

Every broker's inbox has an AI pitch in it right now. This is the version from someone who takes finance calls all day: what actually moves deals, and what's just noise.

Emmanuel OnuohaFounder, MANOai8 August 20267 min read

I work a car finance floor and I build software for it. So when a broker asks me "should we be using AI?", I don't answer with a demo. I answer with a question: which part of your day is repetitive, describable, and happens every single week? That's where AI belongs. Everything else is a distraction dressed up as innovation.

Here's the honest map of where it earns its place in a brokerage, and where it doesn't.

1. Writing the messages nobody has time to write well

A chase text to a customer who's gone quiet. A polite nudge to a dealer. The "we've been approved, here's what happens next" note. These get fired off in ten seconds and it shows. AI drafts a better version in the same ten seconds, in your tone, with the customer's details already in it. The account manager just reads it and sends.

Worth it? Yes. This is the fastest win in the building. The work has no judgement in it and you do it fifty times a day.

2. Answering lender criteria questions instantly

"Will this lender take a zero-hours contract?" "What's the minimum income on that product?" Right now that answer usually lives in one experienced head, and everyone queues at their desk. An AI tool that knows your exact lender panel answers it in one line, for everyone, without the bottleneck.

Worth it? Yes, if it's trained on your panel and criteria. A generic chatbot that guesses at lender rules is worse than useless in a regulated business. It has to know your actual rules or it stays switched off.

3. Getting the quote out before the customer cools

Speed of answer wins finance. If a quote needs a calculator, a spreadsheet and a colleague who's at lunch, some of those customers have signed elsewhere before you reply. Put the calculation and the paperwork behind one button and the answer takes minutes.

Worth it? Yes, when the delay is mechanical rather than a real underwriting judgement. Measure how long a quote takes today and where the minutes actually go.

The test I use before building anything for a broker: is the work repetitive, is the logic describable, and does it happen weekly? Three yeses and it pays for itself. One no and you're automating for the sake of it.

Where AI is the wrong tool in car finance

Anything the FCA would want a human to own. Affordability calls, suitability, the actual decision to lend. AI can gather the information and lay it out. It should never be the thing that decides, unchecked. In a regulated business the human's name has to be on the judgement.

A process your team can't describe. If three account managers handle a stuck deal three different ways, automating it just makes the confusion faster. Agree the process first.

A gimmick that doesn't touch a deal. Plenty of "AI features" look clever and change nothing about how many cars get financed. If it doesn't remove real, repeated work, it's a toy.

How to actually start

Not with a transformation project. Pick the single most annoying repetitive job on your floor — usually chasing or re-typing — give it to a tool for a month, and keep a human on the approve button. If it sticks, extend it. If it doesn't, you've spent little and learned exactly how your desk really runs.

That's how the portal I build grew: one tool for one sales floor, extended feature by feature because real account managers used it every day and asked for more. The technology is only worth anything when it quietly gives people their time back.

Wondering what AI would do for your desk?

Answer three questions and Emmanuel will reply himself: what's worth building for a brokerage, roughly what it takes, or an honest "you don't need that yet."

Diagnose the drag