In conversation with Tallia’s Co-Founders Jack Mander & Alex Bartlam

Tallia builds AI research and drafting tools for tax advisers. It combines a knowledge graph of tax legislation with the latest models to answer complex questions accurately, with citations. Outward recently led Tallia’s $4.5m Seed round, investing alongside ICAEW, Portfolio Ventures, and angels from across tax, legal and technology. You can read why we invested in Tallia here.

Jack Mander and Alex Bartlam met at a military sixth form, lived together through university and went on to very different careers. Jack as a Chartered Accountant & Chartered Tax Advisor in Big 4 tax, and Alex as a Captain and Technical Lead in the British Army. We sat down with them to shine more light onto the founders behind the company, from the influences of their connection to the military, to the rise of Tallia and decisions made along the way so far.

Alex: I come from a military family. Both my parents were in the military, and I went to a military sixth form, which is where I met Jack.

I’ve always had a technical streak. I studied electrical and electronic engineering, worked in cybersecurity in the US and then joined the army. I had some great jobs there, from ceremonial roles to UK ops, and left as a captain managing information and communication systems for UK operations supporting Ukraine.

After that I went to Cambridge to study technology policy, just as ChatGPT and Claude were becoming household names. That turned out to be quite relevant.

Jack: I knew Alex pretty intimately. We lived together at school and at university. I often joke that I’ve lived with Bartlam almost as long as I’ve lived with my wife.

I joined PwC’s transactions tax team working on large-cap private equity deals, then moved to Deloitte to do the same, working on the Blackstone account. Along the way I qualified as a Chartered Accountant & Chartered Tax Advisor.

At PwC, before LLMs were commonplace, I was building internal tools to research tax legislation. At Deloitte I was part of the effort to buy and onboard software for the tax practice. Those were the main insights I took into Tallia.

Jack: We were in different boarding houses, but the school only had about 100 people and everybody boarded full-time, so it was pretty intimate. Lots of camping… actually not camping – green exercises as they call them in the army.

Alex: We also had a lot of the same friends. It was more a shared group than seeing each other across a room and going over to say hello. And then we lived together all through uni.

Jack: I was still at Deloitte when I went to a YC event where Tom Blomfield was speaking. He said it’s really hard persuading customers to buy software from a new company, so your first challenge is to think of the smartest person you worked with at school or uni and pitch them on starting something with you. I rang Bartlam that night.

Jokes aside, he was the smartest person in our friendship group, hands down. And he knew something about engineering, which I didn’t.

Alex: Cambridge was the first time I came across the startup world properly. There’s a lot of romanticising of building a company there. Jack hadn’t romanticised it. His work at Deloitte was long hours and a decent slog. Mine was long hours, difficult and high stakes. It’s refreshing to start something with someone who knows it’s a serious endeavour and that you’re signing up to do something hard.

We’re still yet to have a huge argument. I’d semi-look at Jack while saying that. But we’re good at saying when something is a thing we actually care about versus a passing opinion we’re pushing to get something done.

Jack: Quitting my job. Neither of us had loads of savings, Bartlam was a student so he had even less than me, and my wife was pregnant at the time. It felt like a big risk. But I knew if I didn’t quit then, I never would.

Alex: Pretty much the same. If Jack had stayed at Deloitte he’d have made partner and had a very normal career. I had a pretty clear path in the army that would have sounded great as a full career. It’s about getting out before you get too comfortable.

Alex: We sing the national anthem every morning. Haha, no, to be fair, Sandhurst pushes hard on leadership, but it’s a very different style to what makes a good founder. A lot of what I took away is what doesn’t work with a high-performing team. My troop was quite different to the team Tallia. Our office has a lot of humour and we don’t take ourselves too seriously. Making sure everyone is approachable and in good spirits is front of mind.

Alex: Drive is the main thing. Then fitting into the culture and being able to talk to customers. Every engineer here goes and works in client offices. That isn’t just sitting around. Quite often an engineer spots something over someone’s shoulder, helps with a workflow, and it ends up in the product.

For engineering we care more about high agency than a tax background. Our first non-engineering hire came from tax and that’s going really well, so we’ll keep hiring from the people who’d be our customers.

On culture, we’ll need to write it up for the website at some point. But we try to be as approachable as possible. If someone’s doing something wrong or you disagree, you say so. At our stage it’s easy to go off on a tangent and sink loads of time into something that isn’t worth it.

Jack: At PwC, without LLMs, you could almost map mechanical legislation as a decision tree. With lots of time and engineering support you could almost get to a compelling research tool. But without an LLM in the process it didn’t quite work. That project would have been knocked to the side by LLMs.

Then I sat on Deloitte’s buying committees, where we’d talk about what was in the market. Legal had taken off with Harvey, then Legora and lots of follow-ons. Accounting is essentially the same work: knowledge workers looking at precedent and rules and writing reports. And there was nothing for them. The few tools that existed were bolt-ons from publishing houses, and the product was poor.

So you had something that was almost there, and suddenly the technology to make it work in practice, saving people several hours a day. You could see what happened in legal and think, there are hundreds of thousands of people in tax. Surely it comes to tax next.

Alex: Legislation is incredibly intertwined, within the UK and even more so once you open up to multiple jurisdictions. Working out which rates apply to which actions and assets is complex. So mapping the metadata and the relationships between those data points, so a model can bring the right context into its answer, is fundamental.

Accuracy is everything. You can’t be okay and sometimes get it wrong. You need to be in the high 90s on complex queries, otherwise advisers can’t use it in practice.

The second piece is being model agnostic. As models improve, our product improves with them. We’re happy when a new model comes out rather than scared.

And because tax law never sits still, we don’t map it once and sit on it. We refresh it daily. The simple answer is we throw a lot of compute at it and it’s worth it for the results.

Jack: The Budget is a really nice example. A general model, even one doing agentic web search, hasn’t been trained on a Budget that happened yesterday. So it goes off and brings back outdated, superseded sources.

We embed the Budget day updates on the day and give the model re-ranking guidance, so it doesn’t just retrieve the new material but understands the hierarchy of sources.

Advisers also need to horizon scan. If I’m doing a tax structure, it might need to last 10 years. So how do we look ahead at what’s in the news and build in alerts? And how do we get it right the day after the Budget, which is exactly when everyone rings their adviser?

Jack: We took the classic advice: go deep and narrow. It doesn’t quite work in tax. Advisers have specialisms, but they have to know about all areas of tax. We started with one area of tax structuring and quickly realised we had to cover all of it. It’s not good enough if I can tell you your inheritance tax position but miss the stamp duty on the transfer.

Alex: We thought a research agent this accurate was something we’d build much later down the line. It felt like boiling the ocean. But to do even one vertical in tax well, you need an agent that understands the whole environment it’s operating in. And once you’ve got that, you might as well sell it to all the adjacent teams too.

Jack: There are lots of tech sceptics worried about what AI means for the profession, for jobs and for how people are trained. We see a net gain. So it’s great to have the profession’s own body endorse us and send a signal that says: this is something we want you to embrace and use to grow the profession.

We also work in a very regulated industry. ICAEW using Tallia helps them learn how firms are using these systems and how to regulate them well, so firms can grow but use AI safely. That feeds into our product design all the time. Clients are buying trust, and having ICAEW on board helps us deliver it. And while they’re UK focused, they have members all around the world.

Jack: We’re deep believers that the tax industry will grow much faster with AI. There’ll be more work for advisers, not less. Loads of people never took advice because it cost as much as the penalty would have. As advice gets cheaper, many more people will take it and pay for it. Tallia will give advisers the tools to meet that demand.

Alex: There’s a lot of “everyone will just use Claude or ChatGPT”. Tax advisers don’t really buy that, and our competitive pilots show you need a specialist tool. Firms that have rolled out frontier models across the business are still talking to us.

Frontier labs all leapfrog each other every month and there are data residency questions, so there are plenty of reasons not to lock into one ecosystem. Meanwhile we’re built on top of all of them, so our clients benefit from every bit of their progress, just with the accuracy and citations that tax work demands.

Jack: At the moment we’re getting a lot of warm referrals and we’re growing in the UK. We want to be the default choice here, then use the member firm networks to grow internationally. In any country where you need trusted tax advice, whether you’re an adviser or an in-house team, Tallia is where you turn. Whether it’s our product or APIs into it, we want to be the tax reasoning engine that sits behind the tax advice.

Thank you to Jack and Alex for sitting down with us. If you’re a tax adviser or in-house team who can’t afford to be roughly right, book a demo with Tallia. And if building the tax reasoning engine behind the advice sounds like your kind of hard problem, Tallia is always on the lookout for more talent – take a look at their careers page.