skip to Main Content
Deep Learning : AI for InsurTech Insurance Innovation

Meet the Founders

Eighteen months ago Steve Tuff and Stephen Anderson decided to commit all their efforts to helping businesses involved in Insurance and Reinsurance explore and exploit Artificial Intelligence, using Deep Learning in particular. It was a logical decision based upon the fact that they both knew that AI is already, and will continue to revolutionise the (re)insurance sector. Both Steve and Stephen have been involved in the (re)insurance market for many years, in fact they met whilst working at Willis Faber Dumas in Trinity Square, becoming firm friends with their paths crisscrossing over the years. 

Steve Tuff, Partner

I had been involved during the last 5 years in the Intelligent Search sector, using AI, but focused on Investment and Banking organisations primarily to better understand derivative documentation. 

It was blindingly obvious that this area of application was barely scratching the surface in terms of the power of AI and I became more and more frustrated by such a narrow focus. Particularly when you would hear more and more about the increasing reality of driverless cars, computers beating humans at complex games, and so on.

My background is in marketing and sales, working for the likes of Oracle, BEA, Wang and a natural language startup, so I have seen many promising technologies and their associated hype cycles. However, I have always remained excited by technology and have seen huge benefits delivered to early adopters.

I passionately believe that Deep Learning is for all businesses whether they are big or small, global or local. Deep Learning should not be the preserve of BigTech with deep pockets. We want to democratise Deep Learning. The re(insurance) market provides a unique opportunity in that it combines a wealth of historic data and processes heavily reliant on people.

Deep Learning can expose the ‘Corporate Intelligence’ within an enterprise, for example, from the consumption of historical structured and unstructured data. Therefore, there is a risk in businesses giving third parties such as BigTech or InsurTech access to their data and absorbing what in effect is their Intellectual Property (IP) often to build a market solution or engine. Here at direct|affinity we have no interest in our own solution and absolutely no interest in any of our client’s IP we instantiate say in a neural network.

+44 7770 661034

Stephen Anderson, Partner

I have been involved with the Financial Services sector for well over 25 years. Much of this has been in Insurance and Reinsurance, an area I really enjoy. It is an opportunity rich complex environment with lots of moving parts and involving different types of organisations, and also great people.

My involvement has typically been in delivering actual change and managing it operationally, and I have a personal legacy in that over 10% of London and Bermudan market insurance and reinsurance policies and claims continue to be transacted on a platform that I conceived and which my team built and has sustained since 1998.

I have worked in commercial and personal lines within (Re)Insurer, Broker and Intermediary organisations in senior managerial and strategic roles, as well as in the management consultancy sector. 

For example, I set up a joint venture with Stephen Catlin that I ran, growing and maturing the Catlin Group (XL Catlin) information technology function from an initial 2 legacy staff to over 250. I was exceptionally fortunate to have the support of Stephen and other key board members, as well as a substantial budget, to develop and operate theframe, an enterprise-wide (re)insurance solution encompassing underwriting, claims, and accounting & settlement.

About 2 years ago I felt I really needed a complete refresh and started contemplating some form of academic research. Having a background in theoretical physics & maths I found myself gravitating towards AI, and since I had previously done some work in cellular logic image processing with Mike Duff at UCL, headed straight for neural networks and Deep Learning.

It is obvious that the Deep Learning’s potential is enormous, not just in identifying say that an image is ‘a cat is playing with a red ball’, but for to real day-to-day business activities.

Rather than do research I decided to help (re)insurance businesses explore and deploy Deep Learning

One of the things that excite me is that, through Deep Learning, businesses can turn their thinking upside down and inside out. You can start to identify and feasibly perform activities that have perhaps never been conceived or if they have, considered economically or practically impossible.

Additionally, with Deep Learning you can challenge how you perform what you already do. You can seriously consider the cognitive activities performed by your personnel and how they relate to a series of “automated”, i.e. manual data entry, activities that are executed as part of a business process. Ultimately, Deep Learning can help you remove one or more manual steps that support a process freeing up your personnel for more value add activities.

With Deep Learning in mind, you can remove constraints on your thinking.

+44 7887 767722 

Back To Top
×Close search