Artificial intelligence is already moving into everyday retail decisions, from product recommendations and customer service to campaign optimisation and demand planning. The practical question is no longer whether retailers will use AI, but whether the customer data feeding those tools is accurate, connected and governed well enough to trust. Experian’s Australian Black Friday analysis, drawing on 2024 holiday data, reported that AI influenced 19% of global holiday purchases. While that figure is global and originally sourced to Salesforce, it is a useful signal of how quickly AI-assisted discovery and decision-making are becoming part of retail journeys.
AI does not remove the need for strong data foundations. It increases the need for them. If customer data is incomplete, duplicated, outdated or fragmented, AI may surface or magnify existing data quality issues. That can make outputs harder to trust and may limit the value retailers can get from analytics, automation and personalisation.
For retailers, AI readiness should therefore start with practical questions about customer data quality, connection and governance. Can customer records be trusted? Can the same customer be recognised across systems? Is contact data accurate? Are consent and permitted-use rules understood? Can teams explain where customer data came from and how it is being used?
AI depends on the data behind it
AI and automation rely on inputs. In retail, those inputs may include customer records, transaction history, loyalty data, ecommerce behaviour, delivery information, contact details, service interactions, marketing engagement and location data. If these inputs are incomplete or inconsistent, the outputs may be less useful.
For example, a recommendation model may be less effective if purchase history is split across duplicate customer records. A campaign optimisation tool may be less reliable if response data cannot be linked back to the right customer. A service automation use case may create friction if contact or order data is incorrect. A personalisation program may become less relevant if the underlying customer segment is based on incomplete information.
What makes retail customer data AI-ready?
AI-ready customer data is not just “clean” in a narrow sense. It needs to be accurate, connected, complete enough for the intended use case, standardised, governed and monitored over time.
- Accurate: core contact, address, transaction and profile information is correct enough for the intended use.
- Connected: records can be linked across ecommerce, CRM, POS, loyalty, service and analytics systems.
- Complete: important fields are populated consistently enough to support analytics, segmentation or activation.
- Standardised: formats and definitions are consistent enough for matching, reporting and integration.
- Governed: consent, retention, permitted use and access controls are understood and managed.
- Monitored: data quality is measured over time, not treated as a one-off clean-up activity.
Where poor data can affect retail AI
Poor data can affect retail AI in several practical ways. Personalisation may be based on partial customer profiles. Product recommendations may miss important signals if online and offline purchases are not connected. Customer service automation may rely on incorrect contact or order data. Campaign optimisation may be based on weak or incomplete performance signals. Segmentation may become less reliable if customer value or lifecycle stage is calculated from fragmented records.
These examples do not mean retailers should avoid AI. They show why trusted data foundations matter before AI use cases are scaled. Retailers may get more value from AI when they first address the quality, connection and governance of the data feeding those use cases.
Governance matters as much as quality
AI readiness is also a governance issue. Retailers need to understand what customer data can be used, why it can be used, how long it should be retained and which teams or systems can access it. This is especially important where AI or automation supports personalisation, targeting, customer service or decisioning use cases.
Governance does not need to make innovation harder. Done well, it can help teams move faster with more confidence because the rules for using customer data are clearer. It can also support stronger internal alignment between marketing, data, IT, privacy, risk and customer experience teams.
Five questions retailers should ask before scaling AI
- Can we recognise the same customer across ecommerce, CRM, POS and loyalty systems?
- Are address, email and phone details validated at capture and maintained over time?
- Do we have duplicate or conflicting customer records that may affect analytics or personalisation?
- Can we track consent, retention and permitted use for customer data?
- Do business teams trust the customer data used for segmentation, reporting and activation?
How Experian helps
Experian can integrate data quality controls into existing systems and data pipelines, helping retailers validate, match, standardise and monitor customer data over time. Combined with customer insight and segmentation, this gives teams a practical way to assess whether their foundations are ready for analytics, personalisation and AI use cases.
Want to check whether your retail customer data id ready for analytics, personalisation and AI? Speak with an Experian representative today.