Driving business change - the devil is in the data

Well it has been a while coming but I am now back on track! The first quarter of the year was full steam ahead on a range of interesting work, and my newsletter draft kept being amended as I came across new and interesting topics to discuss, but have now finished this one and lined up some great topics for future newsletters.

So much has happened since the last update, I don't know where to start, The Chat GPT speaker volume got turned up to loud if not blaring, blueprint 2 has had some more updates, core data record moves to a newer version, the Insurtech insights event came and went and finally Tech Nation announced its closure.

It has really been a busy first quarter for those involved in making insurance more digital.

In this episode we dig deeper into Synthetic data and its potential for the insurance industry below, but first ...

ChatGPT has been all the rage and interest has really piqued over the last couple of months and I must say some of the results have been impressive from questions posed at it, but like all AI related activities the need to be mindful of its potential for bias must never be ignored. However predictive it might appear it nearly always starts from a human inputting some parameters.

To find out more about ChatGPT and other large language data model capabilities my colleague Alex Whitmore added some great explanation below.

Blueprint 2 revolution or evolution? had updates in Feb and March, The report following a market survey showed slight increases in familiarity within the market comprehension level and as I have said before getting vendors much more engaged has been reflected in many more conversations from participants with service providers who may well be the good guys/ladies or could they become fall guys/ladies to this programme as it continues?

But whilst Blueprint 2 is essential for the market as a whole, talking as it does about "ambitious strategy and profound change" seems to be a bit of a stretch given it is more about EDI standards, table stakes really, but hopefully the foundation for something a bit bolder in the future.

CDR/MRC updates the latest version of the CDR v3.2 and MRC v3 have been released, these now align with ACORD standards

as we know the CDR's ultimate destination is to support a data first market, but the whole market must be on its guard to not allow a document first tolerance to creep in,

as this would be so 2019.

Insurtech Insights

Having attended this in early March much was to applaud with the great and good in attendance of those driving insurance digital.

The event was in many ways an overload on the senses, with mist and music and I would say almost too many people for a very large venue but it was energetic. Discussion around culture in the market and the challenge it presents on changing things for the better and a drive to go beyond risk protection and plugging the protection gap were some of the top line notes from some of the many sessions running.

Some interesting vendors on show, and whilst I take a great deal of interest in most things Policy Admin or 'Quote and buy' related, I am always amazed at the continual flurry of new entrants onto this, the 'busiest street' of the vendor market.

If you want to find out more about our ever growing 'layercake' of vendors in this space and what makes them different or a fit for your business, please get in touch.

UK innovation agenda at threat or not, my colleague Mark Huxley was quoted extensively in a great piece by Insurance Times, If you want to read more about Tech Nation and its recently announced closure, click below.

Is Synthetic Data the hoped-for game changer for Artificial Intelligence and what could it mean for the Insurance industry?

Artificial intelligence (AI) and machine learning (ML) are immensely data hungry in their training and development. Collecting and labelling these enormous datasets with thousands of objects is time consuming and costly. As far back as 2018 Gartner predicted that 85% of AI projects up to 2022 would deliver unsatisfactory outcomes due to bias in data, algorithms or the teams responsible for managing them.

Synthetic data offers a potential solution to the dilemmas of time, cost and bias as well as for insurance and other financial services organisations, privacy.

Synthetic data refers to a range of computer-generated data types that simulate original, real-life records. It can be created by stripping any personally identifiable information from a genuine dataset to fully anonymise it; or the original dataset can be used as the basis for a model that produces highly realistic data values and qualities. A 2018 Deloitte survey, for instance, found that “data issues” such as privacy, accessing and integrating were considered the biggest challenges in implementing AI initiatives. By generating non-identifiable datasets, however, synthetic-data generation can be a vital tool in protecting privacy of a large datasets. They also become easily reusable and speed up data related activities by having a number of synthetic data sets 'on the shelf'.

What is better than testing a new proposition than running it against a mirror of real customers who have purchased real products as a basis for the synthetic data set. However, as the architectures of AI become more complex, AI remains only as good as the data it is trained on. This is where we see the bottlenecks in AI: There is always a need for large and diverse datasets to advance AI models.

in the best scenario, real-life data is always the primary choice for any AI based solution. but given the constraints of privacy and cost, synthetic data is the best alternative, though generating realistic synthetic data is not an easy task. Synthetic data can be generated at large scale and it is much more cost effective.

As the elimination of bias is better achieved, synthetic data can deliver a better real-world simulation to develop insights and test hypotheses quickly. With low costs and speed of delivery, it is a game changer, enabling the full potential of AI and ML to be deployed across a wide range of industries.

Among the clearest industry use cases for synthetic data is financial services. Such data is already being leveraged to improve operations, with synthetic datasets generated from debit- and credit-card payment-transaction data, and identify fraudulent activity. The UK financial regulator, the Financial Conduct Authority (FCA), partnered with data specialist Synthesized to create synthetic payment data from five million records of real payment data to build a better fraud model without revealing individuals’ data.

So what does this mean for the insurance industry and Lloyd's in particular where RDS', realistic disaster scenarios are used widely for exposure and capital management, the use of available and refined synthetic models to run them would create much higher fidelity outputs and ensure even better market comparisons of portfolios could be achieved. Whilst I am not a deep dive expert on RDS' the combination of actual pricing of a carrier against a synthetic data set would when running a loss profile would potentially give a more real-life picture of potential exposure/impact.

So it is likely in the coming years we will see the Insurance industry increase in adopting the use of synthetic data models for a wide variety of uses to deliver performance and insights at scale for portfolios they currently hold and even those they might seek just to test hypothesis on, using much richer and effective data sets.

The growth in adoption of synthetic data sets by organisations in areas such as proposition development, portfolio management, system testing, fraud detection, claims and exposure management, is just around the corner.

The devil is in the data - creating the data strategy

At Altus we are often, and increasingly, asked by our clients to help with a range of data related topics, extracting value from data is nothing new but organisations are looking to us to support them in establishing a cohesive set of data capabilities and supporting technologies throughout their organisation. We are well placed given both our experience and independence when the data needs across an organisation often benefits from a referee in establishing a single organisational wide data strategy.

Below is our high level approach and steps to ensuring an insurance organisation can create a roadmap to an initial data strategy, which covers architecture, models and governance.

After understanding the businesses data requirements and capabilities in a discovery phase providing the foundation from which the data roadmap can be created. We then move into a targeted exercise with the key stakeholders to build a high-level data model. This data model will give insights into data usage throughout the organisation, with an emphasis on data quality, extraction and the opportunity for analytics and reporting.

After this is validated across the business and in addition to enabling data artefacts to be constructed, the information gathered will give a view on potential requirements for change and pain points.

Finally, we will deliver a deeper understanding of the data and a roadmap to show how the delivered assets can be used and enhanced going forward, in conjunction with a proposed target data architecture and data governance structure options.

We are responding to requests to support businesses really understand their data estate, remove rekeying and multiple inconsistent reports existing across the business and put in place appropriate data structures and rigour. Whilst considering the many technologies that now exist that can turn their data into knowledge.

To find out more how Altus Consulting are helping businesses change in 2023 please get in?touch

#insurance #specialtyinsurance #londonmarket #specialtymatters #Lloyd’soflondon #futureatlloyds #smarterinsurance #altusconsulting