By Barley Laing, the UK Managing Director at Melissa

Artificial intelligence (AI) is already transforming retail. It’s doing so by driving real time success in unlocking data-driven insights, boosting productivity, and elevating the customer experience.

It excels in personalising recommendations based on customer profiles and where they live, delivering segmentation for targeted marketing, aiding the automation of email campaigns and chatbot responses, and providing predictive analysis that helps to predict customer behaviour.

Declining data quality impacts on AI success

A major factor impacting on the effective implementation of AI is data decay. Data quality declines rapidly with customer contact data lacking regular intervention degrading at around 25 per cent a year as people move home, die and get divorced, based on our findings.

Furthermore, 20 per cent of addresses entered online contain errors, including spelling mistakes, incorrect house numbers, and invalid postcodes.

Prevent AI hallucinations

Those retailers with poor quality customer data experience the dreaded AI ‘hallucinations’, which leads to flawed results. Giving an AI tool access to inaccurate data on customers could lead to gibberish, or worse, biased and inaccurate outcomes.

For instance, poor personalisation could be delivered by AI having access to incorrect customer data, such as an inaccurate name or address, which will have a negative impact on sales and the customer experience.

Focus on data verification processes

To ensure AI has accurate data to work with, retailers can have data verification processes in place at the point of data capture, and regularly batch clean held data. This typically entails simple and cost-effective changes to the data quality process.

Address lookup or autocomplete

An address lookup or autocomplete service can work well at the customer onboarding stage. These tools automatically provide the correct address as the customer starts to enter theirs, allowing them to select an option that is accurate, easily recognisable, and correctly formatted for their country location.

Collecting accurate address data in this way ensures there’s no mistakes caused by fat finger syndrome, and there’s a reduction in the number of keystrokes required when typing an address by up to 81 per cent. Common positive outcomes include a faster path to checkout and a reduced risk of basket abandonment, supporting successful sales and a standout customer journey.

Comparable technology can be used to enable real time verification of email and phone data at first contact, helping many retailers to strengthen critical datasets and enhance AI performance.

Deduplicate data

Data duplication rates of 10 to 30 per cent on the databases of many retailers is a significant issue. Data duplication commonly occurs when errors in contact data collection take place at different touchpoints and when two departments merge their data. Duplication can confuse AI applications and add cost in terms of time and money, particularly with printed communications. For instance, not only is sending the same letter twice to a customer a waste of money in print and distribution costs, but it also risks damaging their reputation with customers.

Many retailers are looking at advanced fuzzy matching tools to help eliminate duplicate data. Such a service can merge and purge the most challenging records and create a ‘single user record’, which delivers an optimum single customer view (SCV) that AI can make learnings from.

Data cleansing

Data suppression or data cleansing is an important part of the data cleaning process, and therefore in supporting efforts with AI, because these services highlight those customers who have moved or are no longer at the address on file. A key part of this is having access to the National Change of Address (NCOA) database that’s available in the UK and US, and some other countries, because it highlights those who have moved, and provides their new address.

As well as removing incorrect addresses cleansing services can include deceased flagging to prevent the distribution of mail and other communications to those who have passed away, which can cause distress to their relatives and friends. Adopting suppression strategies enables retailers to cut costs, maintain trust, combat fraud, and enhance their AI activity.

SaaS data quality platform

Due to evolving software-as-a-service (SaaS) data quality technologies, retailers can more easily and cost-effectively collect accurate addresses and wider customer contact data in real time at the onboarding stage, as well as cleaning held data in batch, worldwide. Being SaaS they are easy to access and don’t require coding, integration, or training to use.

In summary

AI has the potential to drive revenue and provide a competitive edge over rivals. Retailers can maximise their AI efforts in 2026 through having access to high quality data. If not, they may experience AI ‘hallucinations’ with unreliable predictions and therefore bad outcomes. The application of best practice data quality procedures gives retailers the opportunity to supercharge their AI efforts, resulting in increased sales and an improved customer experience.

Explore another Melissa blog on how to fix your data here: How your data is letting you down – and how to fix it – IMRG

Published 13/02/26