Building a startup is tough, but Roshan Tamil Sellvan is working smarter and harder.

Alongside his co-founder Alan, he launched AdvisoryAI to transform the world of wealth management. Instead of building another AI chatbot, they developed a system that learns from every interaction, adapts to financial advisors’ unique styles, and stays in line with tough regulatory standards – saving hours of manual work.

We spoke with Roshan about the surprising way he found his perfect co-founder, how embracing a ‘lazy’ mindset became a secret weapon, and what’s next for AdvisoryAI.

---

What specific problem is your startup solving, and why does it matter now more than ever?

AI is no longer some futuristic concept. Tools like ChatGPT have made everyone realise that AI can write your emails, summarise your documents, and even help plan your next holiday. So why not bring that same magic into financial advising?

But wealth management is a whole different beast. It's full of strict regulations, complex documentation, and professionals who’ve been doing things the same way for decades. Many of them are brilliant at what they do, but not exactly thrilled about learning new tech.

That’s where AdvisoryAI comes in. We’re not just throwing a chatbot into the mix and calling it innovation. We're building an AI that actually understands the advisor’s workflow, captures the nuances of client conversations, and generates fully compliant reports. All without the back-and-forth prompt wrangling.

My co-founder Alan ran a financial advisory firm for years, so he saw the pain points up close. Writing suitability reports, annual reviews, meeting notes... It's all incredibly time-consuming. Sure, AI can do it, but we’ve gone a step further. We’re training it to do it right.

How did you and your co-founder, Alan, meet? 

We met at Entrepreneur First, which is a three-month programme with global cohorts in places like Berlin, London, San Francisco, and Singapore. It brings together technical and non-technical people to build startups. About 200 people join each cohort, and I met Alan in London.

If you’ve watched Love Island, it’s actually a very similar concept. There are nights on the rooftop where people sit around brainstorming and discussing ideas. In that process, everything happens so fast because you have such limited time. You might start working with someone, and realise it’s not the right fit. If it doesn’t feel right, you have to move on immediately.

It was quite brutal, to be honest. I was working 16, 17 hours every single day (even on weekends) for three months.

So in that process, did you explore a few different ideas? Or did you know this was the one for you?

Each co-founder brings a distinct area of expertise. My background is in Computer Science (I did my master’s in Machine Learning) so I can pretty much build anything in any space. Alan had expertise in financial management and we had both worked with a few different people before we teamed up. We both brought significant entrepreneurial experience to the table. I’m a third time founder; I'd exited two startups prior to this. And Alan had sold his financial firm as well.

They say you should invest in the founder rather than the idea, because the idea can always pivot. It involved asking a lot of uncomfortable questions, so you know that you’re prepared to tackle anything together. That’s what happened, and after three months, Alan was the co-founder I ended up with!

Wow, that sounds exciting! Was it a tough application process to get onto the programme?

Yes, it’s pretty tough – they only accept 3% of applications worldwide. It’s an intense process with three rounds of interviews where they assess your ability, mindset, and how quickly you solve problems. If the average person takes three months to solve a problem, how can you solve it faster?

They want you to have a founder mindset coming into the programme. Throughout my life, I always ask, what’s the quickest way to do this? because I’m a bit lazy. I always want to find the quickest solution, and that mindset really helps.

Is there something most people misunderstand about AI?

One of the biggest problems is fear. We found that older professionals have more barriers than young people. They think: I’ve been doing this job for 20 years and AI is going to come in and take over my job.

But AI is not here to replace people; it’s here to aid you in doing your job in a more productive way. If you’re currently spending eight hours writing a report, AI can generate the bulk of it in five-to-ten minutes, freeing you up to do something else. The real challenge is shifting perspectives away from the fear that AI is taking over the world.

What’s the most exciting technical challenge that your team has solved so far?

One of the hardest challenges we’ve solved – something that our competitors still struggle with – is perfecting the feedback loop while simultaneously reducing hallucinations and maintaining creative capabilities.

Think of it like ChatGPT generating a report. You might have to keep inputting further prompts to get it right. We've built a system that learns from those edits automatically. After editing five or so reports, that sixth report will be exactly what you want it to be, the way you would write it. Striking that balance of personalisation, accuracy, and controlled creativity is one of the hardest things to crack in AI, but we’ve made major progress.

There’s a lot of discussion around AI ethics. How do you ensure that your product is deployed responsibly?

It’s important to set clear ethical boundaries, especially in the finance industry. That’s why we’ve built safeguards into our model architecture.

For example, we train the AI to filter out inappropriate language and proactively analyse outputs against current FCA regulatory frameworks before deployment. Our reasoning pipeline verification process employs multiple specialised models working to validate accuracy and compliance at each step. So if a report doesn’t meet a certain quality threshold, the system prompts a rewrite. These layered measures help to ensure compliance and trust in the technology.

How about scalability? What’s been the biggest challenge in scaling your product?

Right now, we’ve tapped into the UK financial and wealth management market – we’re generating £100k per month, close to £1.2M ARR. But we know there’s room for much more growth.

Interestingly, we started out building an assistant, but we realised we were essentially developing a highly effective report-writing tool. And that’s something that can be adapted for different industries. We’re currently testing with various law firms, where the response has been really promising.

Initially, we weren’t as bold in expanding beyond finance. But that’s changing. We’re also scaling internationally – expanding into the US, Australia, and Canada. Of course, each market has its own regulations and nuances, but that’s part of the challenge we’re tackling as we grow.

With that in mind, have you thought about how your technology might evolve over the next 3-5 years?

We're building technology that will transform how financial advice works for everyday people! Within three years, our system will act like a smart virtual assistant that creates perfect reports after client meetings and also helps advisors with scheduling, follow-ups, and spotting important financial opportunities they might otherwise miss. Think of it as giving each advisor a dedicated support team, without the extra cost.

The real excitement is in how this will make quality financial advice affordable for everyone. Today, getting personalised financial guidance often requires having significant wealth. Our technology will dramatically reduce the time advisors spend on paperwork and administrative tasks, allowing them to help more people at lower costs.

Soon, the same quality of financial planning that currently requires a six-figure portfolio will be available to teachers, nurses, and delivery drivers trying to make smart decisions with their first retirement accounts.

Do you see it transforming into a self-serve tool that doesn’t involve financial advisors?

Yes and no. Think of it like a gym – you can use the equipment and train yourself (self-serve), but you also have the option to hire a personal trainer (financial advisor). So, it's a hybrid system, not an either-or choice.

Many clients will gravitate toward self-serve digital tools for basic financial tasks and education, while advisors will likely focus on providing higher-value services like complex planning, behavioural coaching, and relationship management.

You’ve currently got a team of 13 based in London. How have you found talent attraction?

One of the biggest challenges in hiring isn’t technical skills. Rather, it’s a mindset and culture fit. We need people with a startup mentality because we move fast, and some struggle with that pace.

We’ve created a handbook for potential new hires to give them a flavour of the startup life. But ultimately, we filter candidates early on. Our job listings are transparent and that helps to refine applications. If someone wants a standard 9-5 with great work-life balance, this isn’t the place for them.

The second filter is our coding test. It’s tough, but it’s not necessarily about whether they complete it. We’re assessing how they work under pressure: Are they coachable? Do they ask the right questions? Problem-solving is one thing, but the right attitude is what we’re really looking for.

We think the opportunity in this space is huge, but the personality has to be the right fit.

That sounds intense! When it comes to AI trends, what excites you, and what do you think is overhyped?

The idea of AI taking over the world… it’s a bit much. There’s a lot of talk about AI replacing software engineers, but did computers replace mathematicians? No. AI was created by humans, and humans will always have the advantage because of our ability to think and reason. You still need the creative brain around the technology to make it happen.

AI should help people. I read that the concept of an eight-hour workday was popularised in the early 1900s – the Ford Motor Company adopted it because Henry Ford believed it was the right approach. But who says four hours isn’t productive? If AI handles the repetitive, mundane tasks, we can focus on deeper thinking.

It also gives people space to be better parents, better people, to be present – and that’s something AI can’t do.

If you weren’t working on this startup, what other problem would you like to solve with AI?

I don’t have just one single idea, I have a list! I’m always writing them down. I believe in tech for good: AI should be used for the benefit of people, not just to make money for a small group.

If I weren’t doing this, I’d love to work on something in FemTech. Workplace harassment, sexual violence – these are issues disproportionately affecting women. I watched my own mother struggle through a career filled with invisible barriers and blatant discrimination. So I’d love to contribute to missions in that space.

That’s amazing! Outside of AdvisoryAI, what are you into? How do you unwind?

On weekends, I love visiting parks. I got into birdwatching when I was very small; nature has always drawn me in. Hiking, camping… just escaping the noise and busyness of the city.

So most weekends, you’ll find me in the park, camera in hand, on the lookout for birds.

Finally, is there anything else we haven’t covered that you’d like to discuss?

I'm still fairly new to the UK and constantly learning about the culture. One thing that stands out is how vital young people are – not just as future leaders and workers, but as the next generation of everything.

If we don't invest in them well, guide them, listen to them… honestly, all this innovation we're pouring ourselves into right now could end up meaning very little.

I don't pretend to have all the answers – definitely not a parenting expert here! But I firmly believe that the world we're trying to build with technology only matters if we’re also nurturing those who will inherit it. They're the ones who'll take what we've started and run with it in directions we can't even imagine yet.

---

📥 Roshan is one of the 54 innovative technical founders recently featured in our 'Faces of AI 2025' report, which is available to download for free now.

At Generative, we work with AI-driven startups to scale innovation, refine strategy, and accelerate growth. If you're building in AI and looking for the right talent to take your company to the next level, let's talk.