What do AI agents buy, which sectors will they lift?

I have previously written about what artificial intelligence can and cannot do, concluding that AI is really good at trawling through large amounts of digital and quantified data, as well as performing repeated tasks that do not involve creativity.

And what activity lines up nicely with this? Shopping online for the best product, at the best price, from the best merchant.

Thus has been born what is now called ‘AI agentic e-commerce’ – which basically involves the use of an empowered AI agent to find, compare and ultimately purchase goods online (within predefined parameters and limits).

Listen/read: Are you using AI for investing or financial management?

This function can serve everyday customers by, for example, managing standard recurring monthly grocery shopping carts (perhaps even communicating with fridges and cupboards that inform the AI agent what is currently ‘missing’ and therefore needs to be ordered).

However, this AI agentic e-commerce function can also very easily be plugged into business-to-business (B2B) purchasing – for instance, finding raw materials inputs for a manufacturing process within key quality, timing and price constraints. In this way, it could replace entire purchasing departments.

This aspect of AI is likely to see steady growth in the coming years (especially where it can replace expensive purchasing departments).

It is therefore worth considering some potential stock market ramifications.

E-commerce websites that use branding, loyalty (read: customer laziness) and subterfuge (read: fake specials) to generate their gross margins will be arbitraged away.

Read:

Amazon sues to stop perplexity from using AI tool to buy stuff
Ex-Google, Meta execs building emotionally aware AI sales agents

I would expect even Amazon to feel pressure from this as a range of smaller hyper-focused peers with some form of cost advantage pop up.

Perhaps they are vertically integrated since AI agents don’t need to shop on one website like Amazon, but can compare and buy from hundreds of small ones run by all the little suppliers to Amazon via Shopify spin-up stores.

I would be immensely cautious about e-commerce websites as investments, no matter how big.

ADVERTISEMENT

CONTINUE READING BELOW

I would be more interested in e-commerce-related logistics services offered by global and perhaps one or two local logistics operations.

Even party logistics (PL) players at the 3PL (core services, including order fulfilment) or 4PL (full-suite, including strategy) level may be a good bet.

Perhaps even the Shopifys of this world offer an opportunity as lots of small, vertically integrated e-commerce websites emerge.

Transaction volumes

All online transactions have to be digital and are therefore not consummated in old school ‘cash’.

As a result, the large transaction networks are the likely winners here.

This because they soak up the added volumes while providing customers, retailers and AI agents with the security, verification and infrastructure needed to transact easily and safely.

Scale and network matter here, and the big should get bigger.

Visa, Mastercard, American Express and to a lesser extent PayPal and a couple of others should win out of this tailwind in the digital space.

Interestingly, about a fifth of all retail transactions still happen in physical cash – even in the US.

The shifting of these into e-commerce should boost the volumes processed by Visa, Mastercard and American Express (whether it flows as fiat currency or stablecoins).

Read:

AI agents will become biggest stablecoin user, says Novogratz
Stablecoins are gaining ground as digital currency in Africa

Velocity of money

ADVERTISEMENT:

CONTINUE READING BELOW

A country’s GDP is little more than a sum of all its transactions over a specified period.

The velocity of money is an economic concept that measures how many times a single unit of currency circulates around an economy over a year or another set period.

AI agentic e-commerce should lower the amount of time taken to arrive at a purchase decision, thereby speeding up the velocity of money as transactions occur more quickly.

In this way, it could actually help lift GDP growth rates as it is adopted.

Mostly positive …

While plenty bad can be said of AI, this benefit should be considered as positive and pro-growth.

Being able to crunch more data from more merchants and arrive at better purchasing decisions that are executed even quicker is a good thing for the economy.

We are still in the early stages of global AI adoption, and it stands to reason that AI will be good at some things and bad at others.

E-commerce may well be one of those things it excels at, making it worthwhile to consider how it could shape the world going forward.

* Keith McLachlan is CEO of Element Investment Managers. 

* Portfolios managed by him may hold investments in Visa.

#agents #buy #sectors #lift