Retail

Beyond the kiosk: Why AI in retail still needs a human purpose

We’re moving from technology that waits for the customer to figure out what to ask toward systems that can participate in the conversation, but the opportunity is not about replacing the associate.

Image: Adobe stock

September 8, 2026 | By Brett Beveridge, Founder and CEO, T-ROC

Summarize withChatGPTGrokPerplexityClaude

I've spent most of my career around retail stores, sales teams, customers, and technology.

One thing I've learned is that retail has never suffered from a shortage of new technology.

The harder question is whether that technology actually makes the shopping experience better.

We've all used kiosks that made a simple task harder. We've stood in front of screens that gave us plenty of information but no real help. We've watched retailers add technology to stores without solving the reason the customer needed help in the first place.

That's why I find the next stage of AI in physical retail so interesting.

We're moving from technology that waits for the customer to figure out what to ask toward systems that can participate in the conversation.

That is a meaningful shift.

But I don't think the opportunity is about replacing the associate.

I think it's about giving the customer access to the right expertise at the exact moment they need it.

The problem retail has been trying to solve for years

Think about a customer standing in front of a wall of products.

They may have ten choices.

They may know what two of the products do.

They may have three questions before they are comfortable buying anything.

What happens next?

If a knowledgeable associate is available, great.

If that associate is helping another customer, working somewhere else in the store, still learning the category, or simply unavailable, the shopper starts doing the work themselves.

They search on their phone.

They read packaging.

They compare products.

Sometimes they figure it out.

Sometimes they leave.

That gap between interest and confidence has existed in retail for a long time.

A traditional kiosk can provide information.

AI gives us the chance to provide guidance.

That difference matters.

From search box to conversation

Most self-service technology has historically required customers to know what they are looking for.

Pick a menu.

Select a category.

Type something into a search field.

Follow the prompts.

Agentic AI changes that interaction because the customer can begin with the problem rather than the product.

"I need a laptop for my daughter going to college."

"Which of these headphones works best for travel?"

"I have this phone. Which accessories are compatible?"

"I need something under $300."

That is much closer to the way people naturally shop with a good sales associate.

The system can ask another question, compare choices, explain a feature, present a relevant product, or help the shopper continue the purchase through another channel.

The technology becomes useful because the conversation becomes useful.

This is where VIBA fits into our thinking

At T-ROC, we have been working on this problem for years through VIBA, our Virtual Interactive Brand Ambassador.

VIBA began with a simple idea: expertise should not be limited by where the expert happens to be standing.

Our platform has combined remote human brand ambassadors with digital experiences so a shopper in one location can connect with an expert who may be supporting several stores. T-ROC has since extended VIBA with AI capabilities that allow shoppers to find, compare, configure, and purchase products across kiosks, screens, and their own devices.

That progression matters to me.

We didn't start with AI and ask where we could put it.

We started with a retail problem:

How do we get more customers the help they need, right at the point where a purchase decision is being made?

AI gives us another way to answer that question.

AI should support the associate, too

There is a lot of discussion about AI reducing labor.

I think that misses one of the better opportunities.

Retail associates spend a lot of time answering questions that are important to the customer but repetitive for the associate.

Is this compatible?

What colors are available?

What is the difference between these two models?

Is this item available in another location?

Does this promotion apply?

AI can handle some of that first-line assistance while the associate focuses on situations where human judgment, empathy, persuasion, or service matter more.

That is good for the customer.

It can also be good for the associate.

The associate gets another resource rather than another system competing for their attention.

T-ROC's current VIBA approach reflects that idea: the AI experience sits alongside retail teams rather than being positioned simply as their replacement.

Consistency matters, but perfection Is the wrong promise

One of the advantages AI can offer retailers is greater consistency.

Brands spend enormous amounts of time developing product messaging, training material, promotions, and selling standards.

Getting that information to every customer in every store is much harder.

An AI brand ambassador can be grounded in approved product information and brand guidance. That can help customers get consistent answers and give retailers another way to extend their selling strategy across locations.

But I would be careful with words like "perfect."

AI is still technology.

It needs good information.

It needs clear boundaries.

It needs monitoring.

It needs a way to hand a customer to a person when the question moves beyond what the system should answer.

That last point is especially important.

A strong AI experience should know when to help.

It should also know when to get out of the way.

The real opportunity is at the shelf

E-commerce has spent years getting better at learning from customer behavior.

Physical retail has had a harder time seeing what happens before the transaction.

Sales data tells you what somebody bought.

It does not always tell you what almost happened.

Why did they compare two products?

What question stopped the purchase?

Was the issue price?

Compatibility?

Availability?

Confusion about a feature?

A conversational AI experience can create another source of information about the questions customers are actually asking.

That information can be useful far beyond the interaction itself.

Repeated questions may point to unclear merchandising.

Frequent compatibility concerns may show where training needs improvement.

Requests for an unavailable product can identify assortment gaps.

Questions about a promotion may tell you that the signage is not doing its job.

This is where AI becomes much more interesting to me.

It stops being another screen.

It becomes another way to listen.

What retailers should ask before adding AI

I would not start by asking, "How quickly can we put an AI avatar in our stores?"

I would start with a few simpler questions.

Where are customers getting stuck?

Which questions are associates answering repeatedly?

Where are we losing sales because expertise is unavailable?

What happens when an item is out of stock?

Which parts of the customer experience require a person?

Where could technology give that person more support?

Once you know those answers, the use case becomes much clearer.

Maybe the answer is an AI brand ambassador.

Maybe it is a remote human expert.

Maybe it is both.

Maybe the right experience begins on a kiosk and continues on the customer's phone.

The channel matters less than solving the problem.

Retail has always been about people

That may sound strange in an article about AI, but I believe it.

People walk into stores because they need something.

Sometimes they know exactly what it is.

Sometimes they need help figuring it out.

Technology succeeds when it makes that interaction easier.

At T-ROC, our approach has always been to bring people and technology together around the customer. VIBA is another expression of that idea.

AI gives retailers the ability to make expertise available in more places, answer more questions, connect physical and digital shopping, and learn from interactions that were previously difficult to see.

That is exciting.

But the measure of success will not be how sophisticated the AI looks on a screen.

It will be much simpler.

Did the customer get the help they needed?

Did the associate have better support?

Did the brand represent itself well?

Did the experience make somebody more confident about buying?

If the answer is yes, then the technology did its job.

And if we keep that standard in mind, I think AI has a very meaningful role to play in what comes next for physical retail.