Alex Molinero, insurance lead at Inawisdom, explains how removing complexity and focusing on quick wins will ensure results in insurers’ digitalisation journeys
Most business operation leaders in insurance will be familiar with the concepts of artificial intelligence (AI) and machine learning (ML) - they will even be keen to implement them. But there’s often a big gap between awareness of these tools and action to introduce them.
Although many business leaders recognise the value that AI can unlock, knowing where to begin can be a challenge.
There are also often common obstacles to overcome, like legacy systems, a lack of specialised technical skills and questions around data quantity and quality.
The good news is that getting started with AI isn’t as difficult as it may seem. Here are some key considerations that make the path to success a lot smoother.
Start with the business value
We’ve all seen businesses take the plunge with AI just for it to fall flat. The problem usually lies in the underlying approach - a focus on the tech itself rather than the problem it’s trying to solve.
That’s why the first step to AI success is identifying an achievable use case that is tied to a clear business outcome.
Start by going after the low-hanging fruit – use cases with low complexity, but high business value. In insurance, we’ve found that intelligent document processing (IDP) often fits the bill here.
For example, one of our customers focused on reducing the amount of time underwriters spent manually inputting data from customer documents.
By implementing IDP, document processing times were greatly reduced, which translated into the faster delivery of quotes and higher customer satisfaction.
Lay the right foundations
It’s no secret that data is critical to AI. But it’s not just about having data - how you store and manage it is also key.
Without a centralised data lake, businesses will struggle to put ML models into production. This is the essential foundation for an AI strategy.
If a firm doesn’t have this foundation in place, it doesn’t have to put its AI journey on hold - just building a data lake must be part of the first steps.
Address the skills gap
ML is a specialised field and requires a range of disciplines – not just data scientists, but data engineers, business analysts, data architects and so on.
Many insurers won’t have all of these skills in-house - and that’s ok. In our experience, a hybrid approach works well - upskilling internal resources in some disciplines and bringing in external expertise in others.
What this looks like in each business will depend on existing capabilities, the project requirements and the firm’s overall strategy.
A popular approach across our customers is to rely more on specialised, external expertise in the early stages of a project, then gradually upskill their internal team as things progress so that they’re ready to take on the day-to-day management once the solution is productionised.