The Role of AI in Transforming Insurance Due Diligence: does it live up to the hype or does 'computer say no'?

Funding for AI startups hit nearly $50bn in 2023 according to Crunchbase Data.?OpenAI CEO Sam Altman said that part of his vision for AI was for a kind of super-assistant, one that would be a “super-competent colleague that knows absolutely everything about my whole life, every email, every conversation I’ve ever had, but doesn’t feel like an extension”.?Sounds grim to me.

However, there is now a genuine question as to whether the artificial intelligence bubble has already burst with billions wiped off the value of technology companies.?We have all seen this type of boom and bust before – to paraphrase Peter Kay – it’s the future I’ve seen it!

So is AI the “Garlic Bread” that we all crave to make our meal of work and life balance so much better?

In Hello World by Hannah Fry, well worth a read, she makes some excellent points about in-built bias.? If you give AI a whole dataset of previous sentencing decisions for example, it is hardly a surprise that instead of some sort of utopian unbiased approach, you get exactly what you had before.? A biased sentencing decision.?The old adage about rubbish in gives rubbish out is ever true.

I really like her “Magic” test when assessing whether claims about AI are bogus. It could also be applied to some Tech Information Memoranda I read. It goes like this “If you take out all the technical words and replace them with the word ‘magic’ and the sentence still makes grammatical sense, then you know that it is going to be bol&%@ks”.

There is no doubt that with sufficient accurate data, AI is set to bring great value to areas of science such as physics and biology where pattern recognition is so valuable.?Think of analysis of MRI scans to recognise abnormalities or to predict outcomes for differing types of cancer that can save lives.

There are plenty of firms out there telling me that if I pay them a chunk of money they can make my business more profitable, saving my key staff from undertaking mundane tasks and freeing them up to deliver even better customer service.?A laudable aim indeed and something I would welcome.?The first stage however appears to be paying them a chunk of money to find out how they can do it.?Call me a cynic but it does make me want to reach for my BS deflectors.

It is interesting that over the last twenty five years or so online insurance purchases have grown exponentially but predominately for personal insurance and simple micro businesses.? There remains a stubborn resistance from business owners to trust their entire livelihood to a computer via the internet.?They continue to value brokers, and the personal service they provide. Also, insurers have discovered that despite investing millions in technology the complex needs of a substantial business mean that a “one size fits all” approach just doesn’t work.

At Vista we have worked on more than 500 deals over the last ten years, from tech start-ups to multinationals with revenue in the billions. I can tell you that doing the job properly involves a huge amount of reading and analysis.

The integration of AI into insurance due diligence processes could offer benefits, including increased efficiency, improved accuracy, and enhanced decision-making capabilities. As entrepreneurs and insurance brokers we do understand that we cannot stand still.?It is important to understand and leverage AI to stay competitive and provide the best possible service to clients. However, it is equally important to be aware of the potential issues and challenges that come with adopting AI technologies.

Here are my thoughts on opportunities and challenges in more detail:

1. Streamlining Data Collection and Analysis

Insurance due diligence involves gathering and analysing vast amounts of data to assess risks, and determine appropriate coverage and pricing. Traditionally, this process has been time-consuming and labour-intensive. However, AI could significantly streamline some data collection and analysis.

Potential Issues:

Data Quality and Bias: AI systems are only as good as the data they are trained on. If the data is incomplete, inaccurate, or biased, the AI's outputs will be flawed. This can lead to incorrect risk assessments and poor decision-making.?As a business we often have to make recommendations based on flawed or limited data but our experience gives us the confidence to do so.

2. Enhancing Risk Assessment

Accurate risk assessment is the cornerstone of effective insurance due diligence. AI-powered tools can enhance this process by providing deeper insights into potential risks. For example, natural language processing (NLP) algorithms can analyse textual data, such as customer websites and news articles, to identify potential risks and assess the exposures of a business.

Predictive analytics, a subset of AI, can forecast future risks based on historical data. By analysing trends and correlations, AI models can predict the likelihood of certain events.?This could be used to help focus where to spend money on loss prevention and risk management measures.

Potential Issues:

Overreliance on Technology: There's a risk of becoming overly reliant on AI, neglecting our own expertise and intuition. AI should augment human decision-making, not replace it entirely.?Think “Computer says No”.

3. Automating Routine Tasks

Possibly one of the best benefits could be derived from Robotic Process Automation (RPA) which can be used to automate repetitive tasks, such as data entry and document verification. By reducing the need for manual intervention, RPA could minimise the risk of errors and ensure greater consistency in the due diligence process.

Potential Issues:

System Failures: Dependence on automated systems means that any technical failures or glitches can disrupt operations and cause delays.

4. Improving Fraud Detection

Machine learning algorithms can identify suspicious behaviours and flag potential fraud cases for further investigation.

For example, AI can detect inconsistencies in claims data, such as unusually high claim amounts or frequent claims from the same individual or business area. This would allow us to help the client to take proactive measures to prevent fraud and protect their interests.

Potential Issues:

False Positives and Negatives: AI systems are not infallible and can generate false positives (legitimate claims flagged as fraud) and false negatives (fraudulent claims that go undetected). This can result in unnecessary investigations or missed fraudulent activities.

6. Enhancing Decision-Making

AI could support analysis of data-driven insights that enhance decision-making. By leveraging AI tools, brokers could potentially provide more accurate and customized recommendations to their clients. For example, AI could analyse a client's risk profile and suggest the most suitable insurance products and coverage options.

The ability to make informed decisions based on real-time data could reinforce trust in recommendations made.

Potential Issues:

Bias in Decision-Making: If AI systems are trained on biased data, they can perpetuate and even exacerbate existing biases in decision-making. It is crucial to ensure that AI models are fair and unbiased.

Human Oversight: While AI can provide valuable insights, it is essential to maintain human oversight to interpret and validate AI-generated recommendations.

Conclusion

The integration of artificial intelligence into insurance due diligence processes is potentially transformative. From streamlining data collection and analysis to enhancing risk assessment, automating routine tasks, improving fraud detection, and enhancing decision-making.

However, it is important to understand that we are not there yet.? We must be cautious of the hype.? As a prudent business we are investing in the early stages of AI integration into the IDD process and will use it where we can to improve the quality of our output and the speed and consistency of our service.

Our team are some of the most experienced M&A specialists in the industry.? Hard-won, that experience and intuitive analysis is not easily replaced.?The variety and complexity of Private Equity deals means that there will remain a need for human intervention.

Garlic Bread?? Well maybe not yet.

Written by Peter Warburton Director, Vista Insurance Brokers