Thursday, June 20, 2024

EXCLUSIVE: “Making the case for AI” – Bastiaan de Goei, Instabase in ‘The Insurtech Magazine’

Bastiaan de Goei, Head of Insurance at Instabase, which combines deep learning and LLMs with low-code tools, explores how AI is changing roles in the industry for the better

By 2030, underwriting as we know it will have ceased to exist for most lines of insurance, reduced to a matter of seconds by automation, with machine and deep-learning models built within the technology stack. More than half of claims activities won’t be touched by a human hand.

Instead, a neural network of connected devices will handle everything from first notice of loss to settlement. And pricing will be available to customers in real time, as their behaviours and needs change, based on their usage and a dynamic, data-rich assessment of risk. Insurance will be embedded into every aspect of our professional and domestic lives, informing choices and effortlessly making good our misfortunes and mistakes.

That’s one vision of the industry at the turn of the next decade, as imagined by McKinsey. But the truth is, the technology exists to realise all of those things today and it’s companies like Instabase that are providing the artificial intelligence to achieve it. According to Instabase’s head of insurance, Bastiaan de Goei, its deep-learning and large language models (LLMs), and low-code tools can already ‘reduce the manual effort associated with claims processing by up to 83 per cent’. It does this by unlocking unstructured data, which exists everywhere across insurance, and extracting, digitising and validating that information to enable the automation of mission-critical processes. As de Goei explains it, a typical casualty claims process might involve an injury document, a medical record, and a description of what happened.

“Human effort is required to understand, interpret, and subsequently action anything related to that claim,” he says. “Instabase applies deep-learning and LLMs to understand that incoming information as if it were a human.”The resulting reduction in manual effort and the streamlining of processes promises wide-ranging benefits for clients, but potentially has even more of an impact on the organisations that deliver them and the individuals who work within them.

“Even over the last 24 months, the application of deep-learning model and LLMs has allowed us to do so much more with highly complex and varied documents, and other types of information and data that exist inside an insurance organisation,” de Goei says. “For the first time, technology can help the underwriter and the claims adjuster in their work, dramatically reducing or even eliminating the more mundane, data entry and simple processing type of tasks that humans ordinarily do.

”While we’ve not quite arrived at McKinsey’s 2030 vision yet, de Goei provides compelling real-world examples of how these tools are already being applied. For example, a short-term disability claim, a big product in the US, involves multiple types of documents, and the average pay-out time for a claim is up to two months.”Imagine if you have to [pay] one or two months of mortgage payments, and other regular bills. That’s painful for the end customer,” says de Goei. “Working with Instabase, one particular life insurer was able to reduce the claim cycle time from two months to two days.”

Using the short-term disability claim as an example, de Goei explains how the insurance industry also has an opportunity to use existing data that has been unlocked by companies like Instabase, to improve understanding of not just current claims, but claims going forward, too.

“Data can be used to understand whether that claim is going to develop into a long-term disability claim,” he says. “Our client’s data science teams use that same rich data to do predictive modelling that they couldn’t do before, because they didn’t have access to the data that sits inside those claims documents.

”Those teams can now intervene on a short-term disability claim that bears characteristics suggesting that it will turn into a long-term disability claim, and make recommendations to that customer on taking action to prevent it.

“Alternatively, the insurer can be ready for when that happens, and make sure they can pay out,” says de Goei.

“A company might tell a customer to take a picture of their driver’s licence or passport as part of a KYC process. But, in the background, that information is still processed manually”

It’s evidence of the transition from an era of ‘detect and repair’ to ‘predict and prevent’, which promises to transform every aspect of the industry. Whether or not customers – and regulators – are ready for it is another matter. But de Goei’s convinced that consumers will like the outcomes, which go well beyond the slick, but limited apps that many insurance companies will have used to onboard them.

“An insurance company might tell a customer to take a picture of their driver’s licence or passport, as part of a KYC process in order to process a claim. But, in the background, those pictures and that information are still processed manually,” he says. “So, in the end, although you have an app experience, it still takes a couple of days before the insurance company comes back to you.”

Using background AI, Instabase can help an insurer understand those documents, in a similar way to a human, and detect errors, omissions, or changes, and even fraud, adds de Goei.

“What that means for the customer is they’ll be prompted if something is missing straight away, and the claim can begin to be processed. Making validation, data extraction and checking for fraud easy for the agent, starts making the customer experience a lot easier, too.”


Such AI assistance in the back office is going to reduce certain multi-step processes, says de Goei. Take a typical large claim that has already been thoroughly investigated on site; it will then pass through multiple desk-based positions before a resolution is sought.

“Instabase can play a role in the evaluation of that information from the contractor or field adjuster, extract the claims amount estimate, as well as the liability decision, so that the desk-based loss adjuster’s job can, by and large, be automated. Rather than having multiple steps of evaluation, it goes straight towards a payment system.”

This goes way beyond simply replacing data entry and potentially changes the nature of the job. Is that a bad thing? Maybe not. De Goei cites another report by McKinsey from 2019 when it estimated that up to 40 per cent of an underwriter’s time was spent on routine tasks.

“New underwriters coming into the market are asking the insurance company, ‘what type of automation solutions do you have in place, because I do not want to spend half my time on stuff that is not of interest to me. I want to be focussing on underwriting’,” the authors said.

In a poll conducted this year by global insurance stock index ACORD among insurance professionals, asking about their perspectives on the future of the industry, 50 per cent said they anticipate that their greatest long-term source of competitive advantage will be the way in which they leverage technological capabilities. Auto insurers, such as Europe’s Covéa, in particular have shown leadership in this space. A survey from the National Association of Insurance Commissioners in the US revealed that 88 per cent of the 193 respondents in that vertical indicated that they currently use, plan to use, or plan to explore using AI/machine learning in insurance practice. Most saw it being deployed in claims, followed by marketing and fraud detection, with only a minority using AI/ML for underwriting, rating and loss prevention. All this will inevitably have an impact on the workforce. Traditionally skilled

teams and talent are still essential, but the requirements of key roles are changing. Underwriters may not become programmers, but they will have to work closely with an increasing number of data scientists in the industry on underwriting solutions. Which is why it’s important that low-code tools like Instabase’s are easily accessible and deliver data that’s clearly understandable to existing staff.


The ACORD survey noted that technology was seen as ‘the most critical capability/competency for the C-suite by 2040’, too. ACORD CEO Bill Pieroni said: “The need for human talent and expertise in our industry cannot be overstated. However, we will see the focus shift toward familiarity and proficiency with emerging tech capabilities. Industry professionals with this skill set will continue to drive innovation and advancement across the ecosystem.”

McKinsey also stressed the need for insurers to create the right talent and technology infrastructure over the next few years.

“Generating value from the AI use cases of the future will require carriers to integrate skills, technology, and insights from around the organisation to deliver unique, holistic customer experiences,” it said in its 2030 vision document.

“Doing so will require a conscious culture shift for most carriers that will rely on buy-in and leadership from the executive suite.”

It pointed out that while ‘the tectonic shifts in the industry will be tech-focussed, addressing them is not the domain of the IT team. Instead, board members and customer-experience teams should invest the time and resources to build a deep understanding of these AI-related technologies’.

“No one thing ever is the key,” concludes de Goei. “What Instabase – and other types of AI companies – are really enabling insurance companies to do is to truly focus on those things that require an art, rather than the ability to do data entry. “At the end of the day, it’s the whole organisation that has to come together, in order to enable extremely good customer experiences. Applied correctly, AI can be a great help with that.”


This article was published in The Insurtech Magazine Issue 10, Page 14-15

People In This Post

Companies In This Post

  1. Bank Execs See Attracting Gen Z As One Of The Biggest Challenges Of The Year Read more
  2. Pro Con Artist Cautions ‘No One Is Un-Scammable’ As Revolut Warns More Scams Reported Among Gen-Z And Millennials Than Boomers Read more
  3. Corpay to Acquire Cross-Border Payments Company Read more
  4. ZA Tech Rebrands as Peak3, Raises US$35M Series A led by EQT Read more
  5. UK’s Global Fintech Community on Track for Further Integrity and Ethics Skills Boost with Innovate Finance and CISI Certificate in Ethical AI Partnership Read more