How to Manage the Implementation of AI in a Regulated Business and Realise the Rewards

Artificial Intelligence and Data science is everywhere!

There are some fantastic use cases using impressive AI technology comparing huge volumes of data against a reference sample to achieve results.

Some examples are:

• Removal of retail packaging - Encourage customers to scan an item with their phone, AI recognises the product and provides rich insight on its features – reduces cost and environmental impact

• Disease detection – identifies early diabetes from a scan of the eyeball

• Quality Control - Compare components in a production line against a quality level data set to highlight defective features and alert staff to take action

Social media and online shopping giants also demonstrate these capabilities, using customer data and their past experiences they predict intent and present customers with content and products likely to be useful. Whilst these are good examples of AI providing benefits for customers and patients, it is more difficult for businesses to deploy these capabilities when:

• Customer data is central to the product and can include sensitive details such as health, gender, ethnicity

• Regulated Industries where transparency of the customer service is mandated

• High brand recognition and value, that emphasises trust and performance

Protecting these values is core to the product and the success of the business, introducing AI can be seen as a risk that needs to be negotiated with caution.

How can Risk Adverse Businesses adopt AI

The benefits realised by AI and Machine Learning can be significant and there are many opportunities where it could be applied to help customers in risk adverse organisations. For example:

• Predicting customer’s unique needs and providing proactive engagement

• Explaining complex products and services specific to a customer’s profile

• Identifying vulnerable customers and providing them with the support they need

Balancing the risks and benefits and presenting them in a business case can be challenging when the capabilities are unproven and the risks are high. Regulators in Financial services have stated that firms need to have in-depth validation exercises to make sure complex machine learning works as intended. It is likely this will be the position for all regulated businesses but this should not be an obstacle to using AI but an opportunity to navigate to use cases where AI can be introduced with lower and understood risk.

Finding the Use Cases and the Best Solutions

There are always well-known points of customer friction, solutions will have been explored many times before, and there will also be areas for improvement that can be identified through customer journey mapping and service review. Bringing these together provides a holistic view of opportunity and benefit that can be compared against new digital solutions with low code implementations that can be delivered quickly and maintained by operational teams.

An Example Deployment

Matching opportunities with solutions in Financial Services – The table below was part of a much larger list that matched the challenges being seen with a digital solution and explains how benefits would be realised and success managed.

Each of these solutions improved customer experience and operational efficiency and by their marginal nature were considered a least risk approach to AI by using progressive deployments. The insight provided into customer intents and behaviours provided data to:

• Prioritise deployments

• Continuously refine solutions and content

• Shape a pathway to securing customer confidence and loyalty

This is now a strategic solution embedded in the business, can be scaled quickly and opens a window to the wider possibilities of AI in a risk adverse organisation. This approach avoids a leap into the AI unknowns, with perceived risks and regulatory challenges by providing entry level capabilities that can be scaled in a controlled manner, considers risks and can be adjusted to accommodate change. There are many capabilities that can be introduced in this way just requiring a shift in mindset and a flexible approach to delivery. Businesses that adopt these approaches will start their progression towards more intelligent solutions, achieve operational efficiency and greater benefits for their customers and staff.