From start-up to AI-driven: how to build responsibly in the age of acceleration

I was invited to join a panel at this year's Sifted Summit, called From start-up to AI-driven: how to build with AI. What struck me most was the energy in the room – not just curiosity about AI, but urgency. Every founder, product leader and delegate seemed to be grappling with the same question: how do we move beyond hype and make AI genuinely useful for the people that matter most - customers? 

Answering that question - in a broader sense, not specific to the use of AI - has been our guide since day one. Our mission is to make car finance simpler, fairer and more transparent, and we see AI as a key enabler of that, but not the end goal. It’s easy to bolt AI onto existing systems to tick a box, but much harder, and far more valuable, to design AI around real customer problems. That is where the true transformation happens.

Always be automating – with purpose

One of Carmoola’s company values is “Always be automating”. To some, that might sound like we’re trying to replace people with machines. To us, it means something different: constantly looking for ways to simplify, streamline and remove the manual drudgery that slows people down, so that humans can focus on bigger, more creative challenges.

We even hire for that mindset. We look for people who are curious about making their own work obsolete – not because their role disappears, but because they are always evolving it. That kind of thinking helps us stay agile and continuously improve.

If you are unsure where to start with automation, my advice from the Sifted panel still stands: begin with customer support. It’s one of the best-trodden paths for safe, visible impact. 

Automating elements of customer support has enabled us to respond faster and more accurately, freeing up our team to deal with more sensitive cases where the human touch is still paramount. Our customer reviews consistently reflect that this improvement has worked – even when customers know they’re interacting with AI. The key is designing with empathy and oversight, not just efficiency.

Human in the loop – scaling the right way

The goal is not full automation; it’s better collaboration between humans and machines. AI can process huge amounts of data and spot patterns faster than any team could, but context, compassion and judgement still come from people.

That is why we have built Carmoola’s AI systems with a “human in the loop” design. In practice, that means automation handles routine or data-heavy tasks – such as routing customer queries or assessing affordability – while people stay involved wherever the stakes or nuances are higher.

The result is more balanced workflows and better customer experiences. Around half of our support requests are now resolved instantly through AI, freeing our experts to focus on complex enquiries and needs. We believe that AI will eventually be able to build trust and rapport, but it’s not quite there yet. People inspire confidence, so you need both.

Building on solid foundations

The biggest trap for start-ups right now is rushing into AI before they’re ready. You can’t build advanced intelligence on shaky data foundations. AI depends on clean, structured and labelled information, and on processes that are transparent enough for teams (and regulators) to trust the outcomes.

Before writing a single line of code, ask: is our data reliable? Are we solving a real problem? Do we have the right ethical checks in place? We approach AI with the same rigour we apply to lending – governance, explainability and continuous monitoring are non-negotiable. Every model that could affect customer outcomes goes through human review and bias testing.

Ethical AI is not a ‘nice to have’. In fintech, it’s a licence to operate.

Empowering people for the post-automation era

As automation reshapes the workplace, the role of humans will evolve, too. When repetitive work is handled by agents or algorithms, our value shifts to creativity, empathy and leadership. That is not something to fear, it’s an opportunity.

At the Sifted Summit, I said that every person in a company will need to think more entrepreneurially. AI can take care of execution, but it can’t replace imagination. It’s up to us to use that extra space to innovate, mentor and build more human-centred products.

Progress, not hype

The companies that win with AI will not be the loudest or the ones deploying the most tools. They will be the ones that integrate AI thoughtfully, ethically and in service of their customers.

We like to think of AI as a co-driver, helping us navigate faster and more accurately, but not yet taking the wheel. Because building with AI isn’t about replacing people; it’s about giving them better tools to do their best work.