---

Building Unilever's First In-House AI Creative Production Capability for B&W

The Challenge
The R&D division for Beauty & Wellbeing needed to explore how AI could support innovation storytelling through video while maintaining the quality expected of brands such as Dove, Vaseline and Nexxus.  Existing production relied heavily on external agency partners and there wasn't an internal workflow teams could confidently use.

My Role
I was responsible for designing and implementing the first internal AI-enabled production capability for the division.

Working across motion and video production, I identified where generative AI could genuinely add value within existing workflows. This required me to assess using things like existing stock vs AI and how that impacts on timelines. R&D was focused on identifying key success vs failure points and whether AI was useful for certain types of subject matter such as serums, materials, packaging and talent.

Through extensive testing and application, I developed repeatable production workflows, documented best practices, and created playbooks that enabled the wider team to confidently adopt AI within their day-to-day creative processes.

What I Built

• AI x Video workflow - 5 Phase SOP
• AI x Motion Key Visual Workflow
• Modular Video Moodboard Builder
• Cheat Sheets for team use
• AI Governance and education with leadership

Example: AI x Video workflow – 5 Phase SOP Map

Organisational Impact
·  Established hybrid AI production capability from scratch
·  Enabled faster iteration and closer collaboration with skin & hair category teams
·  Designed hybrid workflows across static imagery, video, and sound
·  Delivered playbooks with repeatable and scalable pipelines for team use
·  Created new ways of working with branded assets while protecting brand integrity
·  Aligned the leadership on AI capability, constraints and future applications

Learnings
AI is not a wholesale replacement for traditional production techniques, particularly when creating highly specialised visual content such as scientific and beauty imagery. Instead, its greatest value lies in being used strategically. Throughout this project, AI accelerated ideation, enabled seamless transitions between scenes, and generated bespoke visuals where suitable stock footage was unavailable or content needed to meet a specific creative brief.

When accuracy, consistency, and scientific credibility are non-negotiable, generative AI still has its limitations. Visuals depicting skin, collagen, or biological processes require a level of precision that AI cannot reliably guarantee, meaning existing assets or advanced production techniques, such as 3D visualisation, remain the benchmark for these types of content.

---

From AI Pilot to Production at Prose on Pixels, Havas

Challenge
AI existed as experimentation but there wasn't a proven production model yet. How do you convince creative teams that AI can actually produce work for clients? Likewise, how could we prove to clients that the tech was capable of producing superior results to traditional methods.

My Role
I was the first AI creative hire within the team, working on pilot projects and navigating how to develop new workflows for content production across a portfolio of brands. Working across healthcare, FMCG, and beverage sectors, delivering AI-powered video, animatics, and campaign content while improving workflows and team adoption.

Key Contributions

• Testing new workflows
• Using and providing feedback on prompt systems
• Leading multidisciplinary projects
• Expanded creativity for clients through rapid iteration

Projects

• Household
• Healthcare
• FMCG
• Beverages

Results

• 40–50% savings
• 10-day turnaround
• Greenlit campaigns
• Upskilled producers & editors