Software Alliance: Best-in-class modelling for insurers
Guy Shepherd, chief executive of Software Alliance, discusses the future of insurance modelling, why he is optimistic about the year ahead and possible new areas for the Mo.net platform
What insurance technology trends have stood out for you over the last year?
Clearly the last year has been quite challenging for everybody. My overriding feeling is that most insurers have been taking stock of their enterprise modelling technology to ensure they really do offer the necessary levels of robustness and flexibility required in these unusual times.
I think the real benefits of cloud computing have finally started to become apparent over the last 18 months, driven in part by the need for new working patterns. It’s good to see insurers starting to unlock benefits beyond the relatively obvious applications of IaaS or PaaS, although many are now starting to see the limitations of their first-generation legacy modelling software when running in the cloud. We’ve had numerous enquiries regarding our own Modelling as a Service offerings, so it’s encouraging to hear people challenging what has been the status quo for so long.How can Software Alliance help insurers that are overwhelmed with data?
The key thing for us is data integration with the Mo.net platform. For me, data is only as good as its quality, cleanliness and accuracy.
Part of the problem with data is you are in danger of having too much and that can bring complacency. The key is to unlock the value in data, rather than the volume.
From a product viewpoint, Mo.net is typically a cog in a bigger data machine. My view is that Mo.net is better served working in concert with something else, whether it is the programming languages R or Python, for example.
I am a strong believer in using the best tools for a particular challenge or job, rather than trying to do everything, and Software Alliance is not about trying to do everything. We are focused on modelling and serving the outcomes from modelling exercises.
How can Software Alliance help insurers that are overwhelmed with data?
I do worry when I hear insurers talk with pride about the volumes of data they have in their data warehouses, or these days data lakes. Simply having an ever-expanding volume of data doesn’t in itself provide any particular value and can lead to over-confidence and even complacency in the quality of insight derived from this data.
Even with relatively modest volumes of data, there are still numerous challenges associated with the quality, completeness and indeed understanding and application of data, especially for life insurance contracts written many years ago. In my experience, more data only exacerbates these challenges.
At the end of the day, the business objective is surely to derive knowledge and actionable insight from data. Our focus is therefore focused on making that journey as simple and effective as possible. Mo.net has always provided best-in-class data integration capability, for source data, assumptions or results. We recognise the common challenges associated with modelling data and have developed functionality and integration points to address these.
How do you expect risk modelling to change?
The life insurance modelling community is likely to face a range of conflicting challenges over the coming years.
For a start, the industry still needs to get to grips with IFRS 17. While there are benefits on offer to those who embrace the opportunity to do things differently, it does feel like some insurers are simply bolting IFRS 17 functionality onto an already overcomplex and inefficient operational machine. Part of the problem is that people have never had time to reflect and consolidate what they’ve ended-up with after each programme of regulatory change.
Technology will clearly play an important role in any future state IFRS 17 solution, but it’s not clear whether the bets already placed on new tooling will ultimately meet industry expectations.
Alongside the regulatory hurdles, I imagine we’ll hear more about machine learning, artificial intelligence and RPA. Each of these isolated technologies opens some interesting possibilities, but there is a shortage of personnel with the necessary domain and technology expertise to understand the merits and shortcomings of these new paradigms, and develop appropriate use-cases for the industry, beyond the obvious ones.
Have there been any major updates to Mo.net this year and what new is coming up?
The last year has mostly involved supporting our major clients who have been making significant changes and enhancements to their modelling ecosystems. These initiatives have been driven by a combination of IFRS 17 readiness and business process optimisation and streamlining.
We have also been working with several new overseas clients who are in the process of re-platforming from legacy modelling platforms, which is very exciting for us. To this end we’ve been busy working on more tools and model templates to help accelerate the move to Mo.net, including enhancements to the IFRS Assess Enterprise offering.
Finally, we’re starting to look at the next generation of the Mo.net platform and will shortly be prototyping several new features which should extend the potential reach of the Mo.net into areas of financial modelling.
