Site icon Digital Commerce 360

Generative AI in retail is too ‘immature’ for implementation, Stanley Black & Decker says

Stanley Black & Decker is looking for a generative AI tool to generate product descriptions and speed up its product detail product optimization.

After a three-month pilot of two different generative AI product page tools based on ChatGPT, tool manufacturer The Stanley Black & Decker Co. decided to continue its hunt for an AI product description tool generator with other vendors.

“These tools are great — but, caveat — we found they don’t quite meet our needs for a number of reasons,” says Dean McElwee, director, global ecommerce collaboration at The Stanley Black & Decker Company.

Stanley is looking for ways to speed up creation and optimization of product detail page content for its 200,000 SKUs. For a human to do this work, it takes about one hour per product detail page, McElwee says. The goal would be to have this reduced to about 10 minutes, which would mostly be a human validating the content and checking with the brand’s legal department for any flags.

Stanley Black & Decker’s AI product description generator pilot

The pilots looked at generative AI vendors Jasper.AI and Copy.AI side-by-side in creating product descriptions and detail page content for 10 different categories across a mix of its brands in English, Spanish, French and Portuguese. Stanley measured the performance across a few factors including accuracy, readability, tone of voice, style guidelines, legal parameters and usability. Neither tool scored perfectly, and each had a few errors across all the measured categories, with many errors in the tone of voice and style guidelines categories.

For example, Stanley Black & Decker has several brands, each with its own audience, including its do-it-yourself-focused brand Black & Decker, and its professional power tools brand Dewalt. The Dewalt shopper wants information quickly about how the tool will work and a quick demonstration, while the Black & Decker shopper needs more information, education and inspiration. The AI tool created content that did not match the tone of voice of each specific audience, such as using exclamation points for Dewalt shoppers, McElwee says.

“These tools don’t really understand nuance and tone of voice,” McElwee says. “Even if we were to use them, we need to go in and check tone of voice and in some cases, rewriting a large portion, which negates the use case of speeding everything up.”

AI product page generates needs to create country-specific content

How the tool translates content is also important to the global company. Even if the language is the same, countries have specific terms for words that are important when describing how products work. For example, in the U.S. the term for the concrete where pedestrians walk is “sidewalk,” and in the U.K. the term is “pavement.” When describing how to edge grass, the merchant needs to use the correct term in the product description.

Using the correct term also greatly impacts SEO. A U.S. shopper looking for a weedwacker, will not search for a “string trimmer” or a “strimmer” which is what shoppers in the U.K. and Ireland call this product.

If the term “weedwacker” is not consistently used on the page, the product will not show up high in search results in Google or on a retailer’s website.

“When people search for things on the internet, it’s unbranded normally with a category-based descriptor,” McElwee says.

Country-specific content is also important for shoppers when it comes to units of measure, including voltage, which matters when shopping for tools. The AI tool would need to learn that even if the product is the same, the language would have to be different based on the website.

The manufacturer also ran into issues with what it would call “claims.” For example, the tool would generate content such as, “These are the best ear pods and they will last longer.” That type of content would be considered a claim, and Stanley’s legal department would not approve that content, McElwee says.

Stanley Black & Decker continues to look for a generative AI product page tool

Because of all these issues, the manufacturer is going to continue looking for a generative AI vendor that could help it efficiency optimize its product detail pages. An optimized product detail page (PDP) will ensure that the keywords on the page have high ranking for search engine optimization, and the text provides shoppers with a clarity of how the product works, McElwee.

Stanley Black & Decker is among the 5% of retailers that are using AI and it is not meeting their expectations, according to a Digital Commerce 360 retailer survey of 97 retail marketers in May 2023.

Of Stanley Black & Decker’s 200,000 SKUs, McElwee estimates only 5-10% of them are optimized. While that sounds like a small portion, that small number of SKUs is responsible for roughly 80% of its revenue, he says.

Generative AI would help free up more if its copy writers to oversee the tool and get more pages optimized. If more of its product detail pages are optimized, that translates to more revenue, he says. He does not believe it would mean a reduction in headcount.

“We still need a human to check for tone of voice, put it in through legal,” McElwee says. “It’s quite specific and nuanced to our brand and where they are sold in the world. … I hope it might work faster, hopefully get more done, more efficiently.”

There is not a set budget for implementing a generative AI tool. It’s too early what a tool like this would mean for conversion rate impact, he says.

Stanley Black & Decker’s generative AI future

One generative AI vendor the Stanley Black & Decker is looking at piloting next creates and generates results that are more based on that brand’s website content. In contrast, the previous vendors were fed website content, but also relied on other publicly available web content that OpenAI fed into the ChatGPT language model.

While McElwee is hopeful about the technology being useful for its business, he is realistic about its potential.

“We’re going through the Hype Cycle,” McElwee says about research firm Gartner’s methodology about evaluating new technology.

He says generative AI is somewhere between peak of inflated expectations and trough of disillusionment. Meaning, many are excited about generative AI, but soon executives will realize that it can’t solve every imagined use case for their business.

“It’s still immature,” he says.

In a recently published blog, Gartner places generative AI for sales at the peak of inflated expectations.

Gartner places generative AI for a sale application at the peak of inflated expectations on its Hype Cycle for revenue and sales technology. Image courtesy of Gartner.

McElwee estimates that it will be 18 months before Stanley Black & Decker is actually use generative AI in its business.

Do you rank in our databases?

Submit your data and we’ll see where you fit in our next ranking update.

Sign up

Stay on top of the latest developments in the ecommerce industry. Sign up for a complimentary subscription to Digital Commerce 360 Retail News.

Follow us on LinkedInTwitter and Facebook. Be the first to know when Digital Commerce 360 publishes news content.

Favorite
Exit mobile version