Merchandising & Buying

Product attribute enrichment

Automatically extract and assign searchable product attributes — colour family, material, occasion, fit, aesthetic — from descriptions and images. Improves findability and filter performance on site.

AutomationPattern-matchingAccuracyTime saving

75–90%

reduction in manual attribute tagging time

Opportunity assessment

Business Impact
2

Minor improvement. Small efficiency gain with limited effect on overall turnover or bottom line.

Feasibility
4

Relatively straightforward. Off-the-shelf tools handle the core task. Implementation is well-understood with good examples to follow.

Data Readiness
4

Light data requirements. Uses straightforward inputs — documents, product descriptions, customer records — that are usually accessible with minimal prep.

Risk Exposure
4

Low risk. Limited external exposure. A human reviews output before it reaches anyone outside the team.

Change Complexity
5

Very easy to absorb. Runs quietly in the background or gives people a helpful new input. Makes everyday work easier with almost no friction.

Tooling required

Standard LLMComputer visionWorkflow automation

Things to consider

  • Walmart used generative AI to enrich 850M+ product attributes — CEO Doug McMillon stated it would have required 'nearly 100 times the current headcount' without AI (Modern Retail).

  • The attribute taxonomy must be agreed and locked before building — the AI assigns to your taxonomy, and changing it mid-project means reprocessing everything.

  • Image-based attribute extraction (colour, silhouette, styling) requires a computer vision model or multimodal LLM. Text-based extraction from descriptions is simpler and a good starting point.

  • Human QA on a sample of outputs each month catches taxonomy drift and keeps confidence levels honest — especially important where attribute accuracy drives filter conversion.

  • Poor attribute tagging directly suppresses filter usage and conversion. The downstream impact on search and navigation performance is often larger than teams realise, which makes this worth investing in properly.

  • Experiment starter: Select 100 products currently lacking full attribute coverage. Run automated attribute extraction and have the merchandising team QA the outputs. If 85%+ of attributes are correct without editing, proceed to automate for new product onboarding. Measure filter usage and conversion on those 100 products after re-tagging.

AI Transformation Playbook

Ready to assess your own opportunities?

The playbook gives you the full 150-opportunity directory, scoring tools, and 100+ templates for every stage of an AI transformation programme.