Agentic commerce: Why B2B haggling is ripe for disruption

“Agentic AI does for shopping what the washing machine did for laundry.”

That’s the bold claim from futurologist Magnus Lindkvist, who was speaking at Mastercard’s Future of Shopping event at the Royal Exchange in London, near the Bank of England.

But this is a revolution that won’t just affect consumer retail, he reckons. In fact, it will arguably have a deeper impact on how B2B firms conduct business.

Think about how peculiar the word ‘shopping’ actually is, Lindkvist pointed out. We do not describe watching films as “cinema-ing” or going on holiday as “hotelling”.

Shopping is an activity invested with a social meaning that is different to simply buying. And that experience is rapidly evolving thanks to AI, he argues. Indeed, we may need a whole new vocabulary to describe it.

To illustrate this historic change, an impressive tapestry was unveiled at the event, a nod to the Bayeux Tapestry, which is currently on loan to the British Museum. It’s been given the playful name of the ‘Payeux’ tapestry, showing off the history of financial transactions up to the modern day.

It was an impressive backdrop for the unveiling of some fresh data. Researchers working with Lindkvist and his team of futurologists for Mastercard spoke to 26,000 people in households with teens, across 13 countries and found:

• A third (33 per cent) of teens already use AI weekly
• One in three (36 per cent) of those teens would happily hand over to a fully AI-run shopping assistant that chooses and pays for products

The futurologists predict that the teens’ vision will come to pass by 2030:

• AI agents will negotiate at scale for shoppers
• One in 10 consumers will use a trusted AI assistant to shop and pay on their behalf, starting with routine purchases, like groceries, toiletries and subscriptions

Of course, people will hand over this ‘shopping’ to bots because they believe they will save time and money.

Payments reconfigured

This rise of the bots has important practical implications for Mastercard as well as many others in business.

An important security factor will be ‘tokenisation’, according to Agnes Woolrich, Senior Vice President, Product and Customer Solutions at Mastercard. This involves the substitution of sensitive card information during transactions so that criminals can’t get access to it. “This will be critical to agentic commerce,” she says.

All parties in these new kinds of transactions will need visibility of what permissions and instructions the agent has been given.

Pre-authorised payment will also become necessary, and Mastercard has already begun working on this with developers.

Moreover, the proof of a human’s intent to pay will become as important as the processed payment itself (since that will be initiated and handled by a bot), the company reckons.

And in terms of compliance, multi-factor authentication designed for human actors on websites will need to evolve to suit agentic commerce.

We should expect legislation that will make agents authorised, monitored and accountable.

The key word here is ‘trust’. Everyone involved in the transfer of money, goods and services must have faith in this new system.

Can a bot be mis-sold?

A transaction may still go awry, even if a bot has executed it. It may have been given inaccurate information by a website, or perhaps another bot. So every transaction will need a machine-readable warranty embedded in it, ensuring all parties know their rights.

This will help to establish who is accountable in case of a contested refund, or a question over mis-selling.

But bots will really come into their own when it comes to finding the best prices, because unlike humans they can be infinitely patient.

Customers who have wised up to dynamic pricing online may increasingly enlist autonomous agents to make purchases at the optimal moment.

So the shelf price will become the retailer's opening offer as AI agents negotiate at scale on shoppers' behalf.

Automatic haggling

So how does all this affect you if you lead a scaling mid-sized business?

You may think all these insights mainly apply to companies that sell a retail consumer product. They need to consider how future customers will find them. How do they stand out in a world of bots rather than people searching SEO terms?

“I would argue almost the exact opposite,” says Lindkvist.

“It’s actually B2B companies that have the most to gain here,” he argues. “Their prices too are often dynamic, in a different way. In B2B there is usually some negotiation leeway, haggling, ‘how do you get the deal? It’s selling on RFPs [Request For Proposals from vendors] and so on. That's usually where you find strange discounts.”

Now AI agents can do this B2B haggling between themselves.

Though some may think this is a risky strategy, Lindkvist again disagrees. You can control agents with whatever rules you want, including caps on what they can spend or asking them to give you sign-off, he points out.

“Whereas an SME now is entrusting a divisional head or someone lower down with a budget,” he says, “on which you might only follow up every quarter. And at that point it could be too late. It’s going to be the ultimate time-saving, killer app.”

Right now, Lindkvist, who works with many companies around the world, sees only “a mild interest in this”, because most companies are using AI as “a kind of rapid-fire, automated research tool.”

But this is a future that could escalate “exponentially”, he predicts.

So get ready for the bots online, an invisible revolution that might free you up to go social shopping without buying.