Jaco Cebula likes to joke that he is simply an "over-promoted software engineer", but in reality he has been asked by Ascot Lloyd, and its owner Nordic Capital, to oversee the digital transformation of the £9.6bn advice business.
As the company's first group technology director, Cebula is deeply involved in transforming the user experience of the consolidator's platform, turning it into a "seamless" experience, but he is also right at the front end of examining what artificial intelligence has to offer a business like Ascot Lloyd.
It is not "100 per cent blue sky", says Cebula, "but it's the use of technology and its application that's central to [Nordic's] vision. They're incredibly supportive; they see digital transformation as being critical in the wealth sector".
Such is the interest by Nordic – which as a private equity firm bought Ascot Lloyd 18 months ago from another private equity owner – that it is hosting a workshop in Stockholm for its portfolio of companies; businesses such as Max Matthiessen, a Nordic pension adviser, and Bilthouse, a German mortgage brokerage, to discuss ways of using and integrating AI into their business.
Anyone who's writing code and isn't using AI is missing a trick.But for Cebula, who is raring to try out AI in all sorts of contexts, it is simply a case of having a go and seeing what works – the choice is almost overwhelming.
"If you take a step back, definition of its uses is the critical thing at the moment. Having come from a technology background, making investment [typically] means there's large capital expenditure investment up front. All of these [AI] tools are just available. The opportunity is to do proof of concept.
"It's like using that kid in the sweet shop analogy, the potential to use these tools has never been higher."
He notes that for some people an LLM – a large language model that uses deep learning algorithms to process and generate new content – is the answer to every problem, but in his view it is a balancing act. For example, using ChatGPT to create suitability letters is not a sensible use.
"Rule-based systems are very precise in what they are doing. They are very deterministic.
"LLM is all probabilistic and it's much more difficult to ensure its performance every single time. With regulatory documents you need to be very clear. If you generate a suitability letter based on previous suitability letters you have to be really clear."
The truth is many people may have already been using forms of AI, as a form of it exists on excel spreadsheets to help generate formulae, for example. The difference now is that AI will be more deeply embedded into one's daily business life.
"Some of the more obvious areas where you can probably start adopting LLMs is giving people access to ChatGPT for writing meeting summaries. You can have a record and use LLM to summarise the discussions."
In addition, AI can be used to go through mountains of legal documents from an acquired firm.
"You can train a machine learning model to learn and go through them for you," helping spot any problems that would take a person longer to identify. "That's where you've got this huge productivity improvement."
Another area Cebula is experimenting with is getting a LLM to write some of the more tedious code that would take a human many hours, for example on the systems that create management information, drawing data from advisers' work with clients, and collating corporate data in one place.
"They're writing a lot of code for management information dashboards. Anyone who's writing code and isn't using AI is missing a trick. It's taking all the data we've got into our core client systems and providing that assistance."
He makes sure to check the AI-generated code for bugs after it has been written.
Cebula came to his role via a slightly circuitous route. He had been chief technology officer and chief operating officer at platform provider Multrees, then went into consultancy for a while and ended up working on Ascot Lloyd's new platform Hubwise.
This launched earlier this year, and Cebula was appointed group technology director in early September.
"It's well known in the industry that many traditional platforms are a bit of a black box, they don't have the API [application programming interface] capability – it's difficult to get to that underlying data.
"[With Hubwise], Ascot Lloyd can take control of that data and it's fully API. All of that data is relevant for management reporting."
The API capabilities mean users can link in from other programmes and prevents the need for re-keying in information from other programmes, for example from risk-profilers, which otherwise would be "a really inefficient use of time".
"If you have API, you can do an interrogation between the two of them and it fits very closely in terms of the way they work."
However, advisers at Ascot Lloyd will not be expected to automatically migrate their clients to Hubwise, with all the challenges that presents.
"The IFA makes the best choice. It's our job, having built the platform to make it the best choice for them. The idea is to make that proposition most attractive in terms of customer outcomes – there should be a natural migration across. It's not something that can be mandated, it needs to be better in terms of customer outcomes."
For Ascot Lloyd advisers, they at least have a champion trying to make the best technology available to them.
Melanie Tringham is features editor of FTAdviser
