How AI Can Transform CX in Retail

There’s a lot of pressure on artificial intelligence to be the savior of Silicon Valley. According to Forbes, 64 percent of businesses expect AI to increase productivity, and 25 percent are adopting it to accommodate labor shortages. Simply put, everyone from big tech to small startups is banking on AI to help them build and scale.

AI won’t drive the same results for everyone, however. It must be tailored to a company’s use cases and implemented where it can deliver better outcomes than its human counterparts.

Here are a few tips for companies integrating AI into an existing product or building with it from the ground up.

Let AI Help Your Business

The idea that AI is here to replace all our jobs is pretty far-fetched. However, some tasks can (and should) be enhanced by AI. For example, in e-commerce, product recommendations used to be configured manually. Companies have automated this process using pattern recognition and AI over the past five years. A/B testing has confirmed that AI is better at recommending products to an online shopper than a human is. In this case, AI delivers better results, improves customer experience, and saves humans from mundane tasks. Companies that continue to do tasks like this manually will be quickly outpaced by competitors that leverage AI.

The Benefits of AI

In 2016, Facebook launched chatbots for conversational commerce. Although the product saw some success, it didn’t revolutionize e-commerce in the way Facebook had hoped. Here’s why: The buying experience online, especially in fashion, combines visual browsing, in-the-moment inspiration, and search. People weren’t used to talking to an online sales rep when shopping for shoes and struggled to verbalize their wants. It was still much easier to simply navigate to a collection and browse dozens of shoes at once. The hype around conversational commerce eventually died down.

The takeaway is that businesses must consider how their customers interact with the brand — in this case when shopping online. Only then can they successfully replace or improve that interaction with AI.

Now that we have large language models (LLM) and ChatGPT, will this bring back the bot? The jury’s still out, but for now, visual purchasing will likely remain a dominant discovery method in online shopping.

How to Leverage AI to Outsmart the Competition

Companies that integrate AI cleverly will grow, and those that don’t will wither. That doesn’t mean that teams need to adopt every shiny AI tool that comes their way — they should intentionally leverage it in ways that will enhance customer and employee satisfaction. Here are three things to consider:

1. Build a solid engineering framework.

In AI, there’s the model — the algorithm that uses a data set to arrive at a decision — and everything else that the model needs to be successful. AI models are becoming a commodity; it’s “everything else'' that's the hard part. Teams must build a solid engineering framework to collect the right data, sort and organize it, and feed it to the model. If companies do this well, they can simply implement the latest model while their competitors scramble to make their data work with it. Companies that leverage third-party AI tools should ensure that vendors have a solid framework in place to continue delivering the best results.

2. Understand the limitations of AI for your use cases.

A single application of AI can’t solve every problem for every audience. Conversational commerce illustrates this well — conversational AI didn’t perform well for online shopping, but it’s seen massive success in the travel industry. Airlines use it to share flight updates, assist in check-in, and accommodate passenger upgrades — all things that they previously would have done over the phone or in person at a ticket counter.

Companies should consider how AI can support or improve their customers' unique needs and experiences. Once they’ve determined how to use AI, they must ensure that they have enough data to feed the model to learn and ultimately deliver good results.

3. Give AI a second chance.

Timing is everything. Applications of AI that felt impossible three years ago are generally available to the masses today. Consider how generative AI alone has impacted nearly every industry over the past few months. Chatbots augmented with generative AI produce better results today than Facebook bots did in 2016. Now, chatbots are responsible for customer interactions across the internet.

Companies that tried to leverage AI a few years ago and were met with roadblocks and poor performance shouldn’t write it off. They should identify areas where AI can improve customer experience and find the right model to meet their unique use cases. A solid engineering framework and dataset will allow them to adopt the latest models as they become available. The companies that do this well will be the ones we’ll be writing about five years from now.

Zohar Gilad is the co-founder and CEO of Fast Simon, a site search and merchandising platform for e-commerce.