30x faster copywriting with gen AI
- 120K Product listings
- 2K Brands
- 45M Items delivered annually
The Very Group is a multi-brand, 100% pure-play digital retailer and flexible payments provider in the United Kingdom and Ireland, combining amazing brands and products with flexible ways to pay to help customers say yes when it matters most. The Very Group is a family of digital retail brands and a team of people passionate about helping families get more out of life.
“We have around 2,000 brands, ranging from the likes of Apple to Hugo Boss. We cover fashion, home, and electronics—basically, everything except grocery,” says Michelle Williams, data product owner at The Very Group. “Our mission is to help families get more out of life.”
The Very Group operates on three customer-focused pillars: ease, choice, and understanding. “Ease means making sure customers have a great digital customer experience when shopping with us,” explains Williams. “Choice refers to the 2,000 brands we offer and making sure we have what customers need when they need it. Understanding is about knowing who our customers are, what’s best for them, and how we can support them.”
The need to scale copywriting for product descriptions
Across its 2,000 brands, The Very Group has over 120,000 product listings on its website. Since The Very Group operates exclusively online, compelling product images and product descriptions increase purchases and reduce returns—ultimately leading to greater customer satisfaction. This requires effective copywriting for each product listed—a significant investment in human personnel time and a challenge at scale.
“We have a team of 10 copywriters who write descriptions for each product that goes on our website,” says Williams. “However, they can’t write 120,000 product descriptions. It’s just not possible. So, sometimes products go live on our website without a description. They have an image and a title but not enough information for customers to fully understand what they’re purchasing, the benefits, and why they should buy it.”
Integrating Teradata VantageCloud, Amazon Bedrock, Amazon Nova, and Anthropic for copywriting
Scaling this level of copywriting is perfectly suited for generative AI. The Very Group has deployed generative AI by integrating Teradata VantageCloud on AWS with Amazon Bedrock, Amazon Nova foundation models, and Anthropic large language models (LLMs) to generate product descriptions.
“We’ve implemented generative AI to help us write product descriptions,” says Williams.
Unstructured product image data stored in VantageCloud on AWS is accessed through open and connected application programming interfaces (APIs) and Amazon Bedrock. Bedrock then provides a serverless environment to access a catalog of the latest foundation and large language models—like Anthropic’s Claude and Amazon’s Titan and Nova models—to perform the visual question answering, generating natural language text that describes the product images. VantageCloud acts as model endpoints, serving as the compute and storage engine to access all the data and run generative AI models at scale.
“We’ve been using Amazon Bedrock and Amazon SageMaker,” says Williams. “Bedrock gives us the flexibility to choose models that are already prepared. They’re quick and easy to access. By teaming that with VantageCloud, which gives us the scalability of data, we can utilise the two together to become very powerful.”
Williams explains how solutions from Teradata and AWS work seamlessly together: “Once the models have been prepared and are working, we feed that back into VantageCloud for further analytics. The two are intertwined and work well together to give us scalability and flexibility.”
Generative AI contributes to greater customer satisfaction
The Very Group has already seen several benefits in leveraging generative AI to write product description copy. Copywriting productivity is exponentially increasing and freeing up copywriters to focus on more demanding, higher-value copy that requires more human involvement.
“Our copywriters used to take around 30 minutes to write a product description,” Williams says. “Now, that’s down to three minutes. We’ve become much quicker and more efficient in getting a product live on our website, while our copywriters are getting through much more volume.”
This results in a near 30x improvement in the time it takes to write and publish product descriptions. In addition to copywriter efficiencies, customer experience and satisfaction are improving.
Accurate product descriptions also help customers to better understand items that interest them, resulting in fewer returns. Since The Very Group is a pure digital retailer, shipping costs represent a significant portion of the cost of goods sold. Fewer returns can lead to significant cost savings.
“Product descriptions on our website aid the customer experience,” says Williams. “We see higher conversions, higher add-to-basket ratios, and a reduction in returns. This is a real business impact with coverage across all our products, improving customer experiences and sales.”
Trusted AI powers innovation at scale
The Very Group’s generative AI product descriptions use case is just the latest in its AI innovations. As an innovative digital retailer, The Very Group has been ahead of the AI curve.
“We’ve been doing AI at The Very Group in one form or another for a long time,” says Williams. “For example, we have forecasting models and pricing and promotional AI models that understand how we can optimise prices for our customers at the right time.”
The Very Group takes a measured approach to ensure transparency and responsibility. Williams emphasizes that Trusted AI starts with data.
“Data is key for any AI project,” she says. “Data at The Very Group helps customers, giving them the right item and making sure they can get what they want, when they want it. It’s always about improving the customer experience.”
To ensure the ethical and responsible use of data, The Very Group has created an AI Guild with a vision across the business to collaborate on a transparency framework.
“Very early on, from a data point of view, we were training our machine learning engineers and data scientists on generative AI,” says Williams. “However, we needed to make sure skills were widened across the business, so we’ve trained every single person in the business on AI.”
The AI Guild ensures that AI strategies add value and align to ethical practices.
“It’s not just about making sure we have ethical practices in place for AI,” says Williams. “It’s about doing the right thing for the customer and doing it in a trusted way.”
By seamlessly integrating and harmonising data across the organisation, Teradata VantageCloud provides The Very Group’s various business stakeholders with a foundation of reliable, accurate, and governed data—an essential foundation for every enterprise.
“Teradata enables us to have trust in the data and use it for any initiative—AI or otherwise,” Williams says. “Having unstructured data from Teradata is powerful. Our machine learning engineers and analysts trust Teradata. They don’t have to learn new skills to be able to use the power of the unstructured data that we’re unlocking.”
Deploying AI with careful oversight reinforces The Very Group’s ability to turn data and insights into actions that delight customers. With customer-focused AI and data initiatives, The Very Group fulfills its mission of helping families get more out of life.
Let’s connect
Find out how Teradata VantageCloud can help you to accelerate business outcomes and provide business agility you need.
Our sales representatives are here to help.