Evaluate investment opportunities in early stage companies by crunching data on comparables, market factors and broader economic data. This helps investors narrow down their investment choices - but at this stage would continue to include significant margins of risk.

Optimise premiums and risks by granular understanding of the individuals and their propensity to claim.

Identifying new customer growth segments and opportunities - for example those that are underserved by the traditional operating approach. One example might be customers for whom there is limited credit history.

Use geo-data gathered from multiple image sources (satellites, drones etc) and use Machine learning to assess potential agricultural outcomes. AI enables frequent model updates based on new risk assessments as they occur.

Enhance customer underwriting risk predictions through machine learning on multiple approaches and vectors

Streamline insurance application processes - examples would include pre-populating forms or automatically linking with relevant information sources (e.g. public registers etc).

Automatically assess claim through accessing visuals of claimed damage. This might be ground or aerially sourced.

Use social media footprint to build predictive models for assessing individuals behaviour in the real world - specifically driving behaviour and therefore insurance risk. This raises potential data protection concerns in markets like Europe but is a key part of China's national AI development roll-out.

Create more personalised insurance pricing based on actual monitored customer behaviour. Examples would include 'black box' functionality in the car or wearable sensors on the body . Essentially the customer agrees to swap hitherto private data in return for what they expect to be lower pricing as the insurer can better measure real risk - and potentially the exact circumstances around an insurance claim.

Recommend lifestyle improvements to consumer to support insurance policy coverage (life, health, property) through “insurance­ robo-­advisor" (startup Clark). There will be cultural contexts in which this may work better than others.

Improve administrative productivity with Robotic Process Automation. Aviva for example claim a 15x improvements on their administrative productivity.

Lower the loss ratio for insurance companies through portfolio management. Note that the traditional insurance approach to portfolio management is under increased risk from data-heavy machine learning which will deliver increasingly accurate individualised risk pricing.

Identify and sell new mass market insurance consumer products - these may use the advanced data processing and risk pricing capabilities of AI to deliver low-cost niche products, e..g drone insurance, cracked mobile phone screens or specific use case products

Match patients with the right doctor for insurance purposes based on patient, supplier and insurance information

Enhance life insurance by improving life expectancy prediction and underwriting risk by analysing selfies (startup Lapetus). Selfies will provide key indicators - potential alcoholic intake for example. This will raise privacy concerns in markets like Europe,

Develop enhanced or new insurance proposition based on monitoring actual behaviours (e.g. telematics, driving behaviour, health sensors). Whilst this has consumer risks it does offer lower potential pricing.

Enhance insurance customer fraud detection by assessing behavioural data from multiple sources. Potential fraudsters can leave clues, including historic behaviour, that an automated system is better placed to trace, capture and analyse.

Accelerate claims processing by automating processes - this might including automated customer recognition or data capture to speed activity up

Use geo-data gathered from multiple image sources (satellites, drones, etc.) and use machine learning to assess potential flood risk on an individual building basis. AI enables frequent model updates based on new risk assessments as they occur.

Automate interaction between insurance agent with insurance company through chatbot functionality. This will enable firms relying on (often widely distributed) human salesforces to provide scale-able, up-to-date support and advice in dealing with customer issues. Staff supported may be in-house, 3rd party or independent contractors.

Assess building damage from weather or natural disasters with drones and satellite delivered imagery. Especially useful for reinsurance markets in the aftermath of major events.

Use social media footprint to build predictive models for assessing individuals behaviour in the real world - specifically lifestyle behaviour and therefore insurance risk. This raises potential data protection concerns in markets like Europe.

Recognise and categorise type of damage to predict insurance claim valuation through accessing photographs provided by on-site sources. Reduces need to deploy risk adjusters - and the chance that their assessments may be influenced by external factors (e.g. interaction with customers).

Search through data and find healthcare claims that were denied by insurance companies, discovering the causes and speeding up the resolution process, ensuring healthcare systems are accurately paid.

Identify most profitable consumers through data analytics

Robo-advisor automates investment and portfolio rebalancing and transactions based on defined client needs, responding to real-time market data

Compile aggregate analyst report sentiment score based on the balance of positive to negative words in written reports

Track (sometimes in or near real-time) risk exposures, ensuring capital requirements monitored and overall business risk reduced.

Predict how events like natural disasters will affect market prices based on data on similar events

Monitoring voice calls by traders to test for signs of wrong-doing Intelligent Voice have the potential to deliver this service.