From big data to smart data: intelligent customer engagement for insurers

The digital age presents enormous opportunities for insurers: it is particularly exciting to see how big data analyses can identify relevant customer needs even before contact is made – and deliver tailor-made offers at just the right time. With the right data and a good set of rules, leads can be generated automatically and tailored to individual customer companies. This can lead to an increase in cross-selling rates – through cross-divisional identification of needs and improved response and conversion rates in marketing campaigns through personalised and event-driven customer engagement. This blog post highlights how this can work, what prerequisites are necessary and what specific case studies support this approach.

From big data potential to smart data usage

Not all information is automatically useful. Insurers need to structure and refine their data and make it usable for hypotheses in order to truly turn ‘big data’ into ‘smart data’. The key to this is not only technology, but above all a data-driven vision and clear use cases.

Using machine learning and predictive analytics, a variety of internal (e.g. contract data, claims history) and external sources (e.g. weather data, mobility, social media) can be used to make accurate predictions, such as: Who might soon need liability, accident or long-term care insurance policies?

How insurers anticipate customer needs

1. Behaviour and life event analysis (life event triggers)

Insurers can identify needs through data from social networks, purchasing behaviour, online activities or customer interactions:

• Moving house → home contents or building insurance
• Wedding → life insurance, liability insurance
• Birth of a child → health insurance, education insurance
• Car purchase → motor insurance
• Job change → occupational disability or pension products

If, for example, driving behaviour changes via telematics, there may soon be a need for additional motor insurance. Tools recognise changes before customers take action.

Early warning of dissatisfaction: Social media feedback, chatbot interactions or CRM signals can be used to identify an impending desire to leave and address it, for example, with a better tariff offer – for proactive customer retention.

Example: A customer looks at an insurance policy in the customer portal without taking out a policy. Individual contact by an agent, advertising or a reference to an ongoing promotion could still lead to a policy being taken out.

2. Predictive analytics and AI models

AI models can recognise patterns based on millions of data records and predict insurance-relevant events, for example:

Risk of occupational disability depending on job, lifestyle and health data

Accident hotspots based on movement data (telematics)

Health history and need for long-term care insurance (via wearables, eHealth)

Example: An algorithm recognises that customers with a similar profile to customer John Doe will take out occupational disability insurance in two years' time – and proactively recommends it.

3. Personalised, context-sensitive approach (next best offer)

Modern IT systems combine:

• Demographic data (age, occupation, income)
• Portfolio data (which insurance policies have already been taken out)
• External data (e.g. weather, location, browsing behaviour)

This enables context-specific product recommendations.

Example: A corresponding approach could be, for example: ‘There have been an unusually high number of severe storms near your place of residence. Natural hazard insurance protects you in an emergency.’

Advantages and challenges

Advantages for insurers

When customer needs are known, they can be addressed in a targeted manner. In principle, conversion rates for marketing campaigns increase when only interested customers are targeted. But the negative case is also relevant: customers who are currently not interested in a particular product are also identified. By not sending this customer group advertising for unsuitable products, insurance companies save time and money, and individuals do not feel like they are being ‘spammed’.

At a time when comparison portals are playing an increasingly important role, it is becoming more difficult to place multiple products with a customer, as they always decide individually which provider is the most suitable. Through a proactive, targeted, cross-sector approach, gaps in customer needs can be identified and remedied – offering additional cross-selling and upselling opportunities before customers themselves look for cover and purchase the product from a competitor.

In addition to product placement, this creates stronger customer loyalty, as the customer feels well understood. The insurer is no longer a passive provider of insurance, but actively takes care of its customers – not only in the event of a claim. Proactive advice creates another positive touchpoint in the customer journey.

Challenges and limitations

Even if all the relevant information is available, it is still possible that a need has been analysed incorrectly. Not every move indicates a new insurance need, and even if the analysis is based on statistics, there are always cases that are not affected. Ultimately, the analysis is based on probabilities rather than facts.

Insurers are also often faced with the challenge of making all relevant data available. The available data is not always up to date or comprehensive, especially if customers have only taken out one insurance product to date. In such cases, the data must first be collected and documented, which incurs additional costs and effort.

Under certain circumstances, the customer relationship may also suffer if ‘suitable’ offers are sent out. Customers then feel that they are being ‘transparent’. To avoid creating an eerie or even intrusive feeling, explainable artificial intelligence (XAI) can be combined with clear communication: ‘We analyse your interests in order to suggest suitable insurance policies.’

Data protection and ethics

What needs to be considered:

• GDPR: Processing of personal data only with a legal basis (e.g. consent or contract fulfilment); personalisation must not violate applicable law; tracking and data analysis must be transparent
• Privacy by design: anonymisation, data minimisation
• Transparency: customers must know how data is used
• Protection of confidence: No ‘creepily accurate’ addressing (see above)

Further development: Adaptive, learning systems

With AI-supported systems, it is possible to address customers not only in a more needs-oriented manner, but also in a more personal way. With each interaction, the system becomes more intelligent (reinforcement learning) and customers receive not only personalised offers, but also:

• Timing optimisation (when is the right moment?)
• Channel optimisation (push, email, phone call?)
• Tone adjustment (e.g. young, factual, emotional)

Requirements and organisational implementation

• Data infrastructure and platforms: Modern data platforms such as data lakes break down silos and enable automated data analysis across all sources.
• Organisational transformation: Interdisciplinary teams with data scientists, specialist departments and IT – working in an agile manner – place customer focus at the centre of product development.
• AI-supported models: Predictive models identify relevant situations and trigger automated product recommendations before the customer takes action.

With in|sure Ecosphere, adesso provides the basis for lean processes and a good data foundation in software that adapts to market needs. With our expertise in campaign management and the implementation of CRM systems, insurance companies can prepare for the future – all from a single source.

Conclusion and outlook

A data-driven strategy enables insurers to identify needs before customers express them, offering a modern, intelligent and needs-oriented customer experience. This requires a solid database, smart models and an agile organisation.

Insurers who take this approach position themselves with a clear advantage: they are there before customers have even expressed their interest.