Initial Situation

The retailer had a range of generative capabilities being used across their Creative teams, but these were isolated and offering limited operational benefit. The client and vector8 had identified the need to consolidate these into a single, repeatable and scalable way of working; one which could deliver the creative quality and range they required as well as the operational benefit.

The Trading and Marketing teams’ vision was to build a best-in-class campaign process; one which would enable faster, higher quality campaigns at scale. That vision required a solution which would allow campaigns to be briefed and scoped much more efficiently, as well as an increase in the campaign media production rate.

An earlier phase of work with vector8 had identified the business case of building a bespoke solution. The opportunity to build a specific and differentiating capability was further enhanced by the avoidance of vendor dependencies and ongoing licence fees.

Our Contribution

• Agreed scope to deliver a full range of campaign management and media generation capability, with initial focus on two priority use cases: Paid Media campaigns and Product Detail Page imagery.
• Designed, built and deployed two bespoke applications for the retailer's Marketing and Trading operations, delivered collectively as the AI Marketing Suite and accessed through a single unified portal:
• Campaign Engine: a single system of record for briefing, budgeting and scoping marketing campaigns across the group.
• AI Content Studio: a multi-media generative capability spanning imagery, video and copy, producing channel-ready campaign and product assets.
• Gathered detailed requirements through structured collaboration across Marketing Operations, Creative Design, Trading, Digital Marketing, Studio and Production teams, ensuring the solution delivered capability and benefit across the business and established a foundation built for scale.
• Anchored the delivery approach in experimentation and evidenced decision-making, using this to resolve the core trade-offs inherent to generative solutions: responsiveness versus quality, precision versus variety, and control versus creativity.
• Designed a solution architecture in which user-facing applications remain agnostic of the underlying model vendor and version, enabling the retailer to adopt new models as the generative AI landscape evolves without re-engineering the application layer.
• Engineered application-level features to safeguard the quality and reliability of generated media, including structured and guided prompting for image and video generation, reference-imagery inputs for scene, human or product likeness matching, consistent application of brand guidelines to every generation, and stepwise generation journeys that build base imagery before applying product or merchandising overlays.
• Ran a structured handover and knowledge transfer phase, including detailed user training guides and upskilling sessions to embed the capability within the retailer's teams.

Value Delivered

• A comprehensive Campaign Engine now allows campaigns to be briefed, budgeted and scoped within a single system, removing the costly hand-offs and duplication of effort previously required across teams.
• The AI Content Studio delivers copy, imagery and video for a campaign from within that same system, with consistent user flows and consistent output quality across media types - replacing what had been a fragmented, tool-by-tool production process.
• The Digital Marketing team gained autonomy to produce brand-compliant, channel-ready media without dependency on central Operations and Creative teams, materially reducing time to market for topical, high-relevance Paid Media campaigns.
• The architecture de-risks the retailer's ongoing investment, positioning the Suite to incorporate future improvements without disruptive rebuilds.
• Embedded brand governance - guided prompting, reference-image matching and consistent brand-guideline enforcement - reduces compliance and quality-control overhead by design rather than through manual review.
• Structured training and handover leave the capability fully owned and operable by internal teams, supporting sustained adoption beyond the initial rollout.

‍