No surprise that at Davos AI is a dominant topic but what may surprise many is that Mistral a French AI startup has emerged as the toast of Davos. Three CEOs of three dominant tech companies told the FT that Mistral is one of the best available products from objective benchmarking. 

Enterprises must choose between open and closed-source models — or between systems whose technical details are transparent to third parties, compared with those that remain proprietary to a single company. Mistral deploys the former model and its founders helped build Meta's open-source model LLAMA. Open-source models are particularly attractive to highly regulated entities, such as insurance and banks, who wanted to experiment with generative AI but could not do it with proprietary software because of compliance reasons. 

Whilst Mistral competes with well-funded rivals OpenAI and Googe, Microsoft gains the best of both worlds. Whilst it invests heavily in OpenAI, Mistral builds its offering on Azure. Enterprises will not want to pin all their hopes on one supplier so Mistral appears to be in a promising position given the benchmarking results. 

Insurers are no different than other enterprises; they fear missing out but are also aware of the potential pitfalls of GenerativeAI (GenAI) and Large Language Models (LLM). They all need to experiment.

On the other hand, as a result of the M&A histories that have resulted in today's insurers inheriting multiple technology stacks they suffer from data silos of unconnected data formats, and, circa 80% of unstructured data. This means that few understand what data they have never mind leverage enterprise-wide insights from it. 

That's what makes a particular segment of companies vital for insurers that wish to experiment with and leverage GenAI and LLMs to find and apply to the optimal use cases.  Two such companies are Aimii in the UK and Palantir internationally. Both offer vast data scientist and data engineer resources, proven platforms, and secure guard-railed experimental environments to run secure PoCs and discovery boot camps.

Aimii, interestingly, positions itself as an enabling AI and data management partner that will help insurers understand their data faster, experiment to validate use cases faster, and thus achieve competitive advantage faster. It has many blue-chip customers as clients for insurers seeking scalability and proven capability.

Palantir has a proven background in defense industries, global enterrises and recently won a major contract with the UK's NHS and so has a strong pedigree.

Insurance CEOs generally wish it to be known that they are experimenting with GenAI and LLMs and wish to assuage the fear of missing out whilst not having a clear vision of which use cases will deliver a competitive advantage and a decent ROI. Many are seduced by the vast subsidies offered by Microsoft, OpenAI, and Google to experiment but may come to rue those inducements once the full bills arrive for production roll-outs.

Hence the significance of Mistral as another option and partners like Aimii and Palantir to experiment. And don't forget that other AI technologies e.g. extractive AI may prove better tools than GenAI for specific use cases. Insurers need practical help and not just technology.

A brave new world needs competent guidance and resources to make the right decisions on leveraging AI.