How are brands using GenAI behind the scenes? Case studies from Unilever, E.on Next & Amarra

Adobe’s AI and Digital Trends 2025 report, carried out in partnership with Econsultancy, found that 65% of senior executives believed AI and predictive analytics would drive growth in 2025, with respondents pointing to benefits in productivity, efficiency, content ideation and production, and decision-making.

While many organisations are still working to transition their generative AI implementation from pilot to proof, some are already seeing results in back-of-house applications of AI like scaling creative production, data insights, and optimising product listings.

Here’s how three brands in FMCG, energy, and retail are employing generative AI ‘behind the scenes’ to improve their business and creative operations – and the lessons that marketers can take away from their example.

Unilever: A GenAI design studio that scales creative production

Unilever has built an in-house graphic design studio, Sketch Pro, that employs generative AI in a bid to help the brand tap into trends and cultural moments faster.

Using Sketch Pro, which draws on the tool Pencil Pro (developed by The BrandTech Group), marketers can use prompts and insights about their key audiences to produce images and video based on 3D ‘digital twins’ of each product. Selina Sykes, Global VP and Head of Marketing Transformation, Beauty & Wellbeing, told Digiday that Unilever was using the system to produce creative assets 30% faster than its previous method.

“We used to send briefs off and get content back. Now it’s this agile, iterative approach,” she said.

This change will understandably cause some alarm for creative agencies, although Unilever has stressed that it is still committed to its agency partnerships, with Sykes calling GenAI “an aid, not a replacement”.

The initiative forms part of a wider approach to improve Unilever’s creative production as it moves towards social-first storytelling and ideas that resonate locally. In particular, by applying Sketch Pro with Unilever’s home care brands, the company is aiming to reshape brand perception, moving from more functional marketing to “emotionally resonant content and design”.

For example, the Jakarta team wanted to tap into Ramadan to create content for laundry detergent brands Rinso and Sunlight. When they identified the trend of lip-syncing through social listening, they were able to use Sketch Pro to rapidly create visual assets for TikTok and jump on the trend.

Within hours the content had driven 6 million organic views, upping the brand’s visibility by 22%.

Each brand in a particular market is assigned an AI tool that draws on brand guidelines and applicable regulations – what Unilever has called a “BrandDNAi” – in order to improve processes without impacting brand consistency.

Sketch Pro is currently running in seven cities, including London, Jakarta, and Istanbul, but is expected to be scaled across 21 markets by 2026.

Takeaway

Every second counts on social media, but jumping into the latest trends can be time-consuming for a major brand with a lot of internal sign-offs.

AI doesn’t have to substitute for human creativity, but it can speed up the process of generating those first iterations of creative assets. The goal shouldn’t be to jump on the GenAI trend for the sake of creating generic (or worse, low-quality) product images; but by training models in the right way, brands can enable their marketers to be more efficient while creating content at scale almost in real-time.

The best way to make it work for a global brand is to go local, test specific markets, and provide the right training and guidelines to maintain quality and consistency.

E.ON Next: AI-driven insights to improve customer experience

Energy company E.ON Next is all about data-driven decision-making – but found that the manual analysis of their customer insights was not scalable.

The brand collects customer feedback from a wide range of venues, including Trustpilot, social media, email, and app store reviews. However, this “significant” volume of feedback was “slow and inefficient” to analyse manually, as Mo Nuur, Head of Customer Product at E.ON Next, described:

“Good luck trying to sift through and make sense of hundreds of thousands of feedback entries manually – whether it’s understanding what customers are saying, determining their sentiment, or linking that feedback to business KPIs.”

E.ON Next partnered with Chattermill, a customer experience intelligence platform that uses AI to analyse customer sentiment. The goal was to listen and learn at scale to produce actionable insights. Instead of manually gathering data from third-party apps, social media, and owned channels, the brand could more quickly identify key pain points and product metrics they should focus on.

These insights led to two smart solutions that directly linked to customers’ pain points: Smarty, an app to help customers read their smart meters, which was created after noticing that many customers struggled with their meter readings; and Billie, an interactive digital bill that simplifies billing while answering questions and sharing tips to support customers with the process.

Thanks to the two tools, the Consumer Happiness Index (CHI) across the online billing journey increased by 48%, while E.ON Next’s Net Promoter Score rose from +9 to +22.

Takeaway

As the volume of customer data continues to increase, it’s becoming more challenging for companies to stay on top of it and gather the best insights. Using deep learning to organise and prioritise data can enable businesses to make better decisions at scale.

For E.ON Next, their data-driven focus on the voice of the customer led to new digital solutions that addressed real customer pain points.

Not all brands have the same needs, and every target audience is different, so it’s crucial to identify the opportunities that will lead to the highest impact.

Amarra: GenAI for product descriptions and inventory management

Formal dress wholesaler Amarra started using AI tools back in 2020 to improve efficiency and customer experience. The brand currently employs GenAI, specifically ChatGPT, to write the first drafts of product descriptions.

The necessary information about each product and its unique selling points are put in by the team, who also finalise the drafts to ensure they fit the brand’s voice.

According to Amarra co-founder Kunal Madan, the use of ChatGPT for content creation has cut down writing time by 60% without impacting product description quality.

The company is also using an AI-powered inventory management system that uses historical data and seasonal trends to analyse and predict stock needs at different times of year. The tool has now reduced overstocking by 40% while also helping to ensure consistent product availability for popular items.

By utilising AI to analyse customer review data, Amarra can also understand the motivations behind changes in buying behaviour.

Finally, Amarra uses AI chatbots in its customer service to speed up the process for common questions that need a quick response, and reports that AI chatbots are currently handling as many as 70% of customer inquiries.

Takeaway

According to Amarra’s Kunal Madan, businesses need to be prepared for teething issues. From challenges with their current tech stack to AI bias that can impact cultural nuances, there are several issues that highlight the need for cautious use of generative AI. Amarra’s customer service chatbot initially provided rote, impersonal responses to customers until its responses were fine-tuned.

There is also the ever-present question of whether using AI will lead to teams being downsized. For Amarra and Kunal Madan, “we are using AI to augment our employees’ capabilities, rather than replace their roles.” For a small business such as Amarra, implementing generative AI in the right areas arguably allows the brand to punch above its weight.

As with every new technology, adoption is a gradual process, and Madan’s recommendation applies to all the use cases explored here: be prepared for a learning curve.

Supporting skill-building among employees is one thing that can help organisations translate their visions for applying GenAI into tangible reality. According to Econsultancy’s ‘Bridging the Gap Between GenAI Ambition and Execution’ research, 33% of marketers see inadequate knowledge and skills as a significant barrier to effective GenAI implementation, while 33% also identified a lack of quality training.

Marketers whose organisations outperformed their competitors (i.e. market leaders) were also 2.5x more likely to have a highly effective GenAI skills and training offer.

While it may feel as if the pace of change is only gaining speed, organisations should still be deliberate and considered in their applications of generative AI, considering where the potential benefits are greatest and how to bring teams along on the transformation in order to reap the rewards.