By Anil Nair
The idea of shopping has now been elevated to a completely different level. E-commerce was a step-up, when shoppers didn’t have to visit a store to make purchases. ChatGPT has just launched a pilot, “Instant Checkout”, which allows buyers to make purchases directly within the app, without having to visit other websites, enabling the payment process as well. You just use conversational artificial intelligence (AI) to convey what you want, and the AI engine then takes over.
Instant Checkout operates on your behalf and relieves you of all the intermediate work, including research, comparisons, and the final checkout. Earlier, you scanned various sites to figure your options, assessed what’s available, perused reviews and comparatives, added the item to a buying cart, clicked the buy button when you were (reasonably) sure that’s what you wanted, and then paid up.
What is happening now will be a huge shift from the buyer being at the centre of the transaction to just setting goals and leveraging automation to achieve them. This alters the market dynamics substantially, defying buyer behaviour concepts, and challenging all current models of commerce. While traditional bots played the role of virtual assistants, offering recommendations, operating within the realms of human oversight, agentic buying will involve gauging the need, perceiving the context, figuring outcomes, and executing through multi-step processes, entirely autonomously.
For example, if flight tickets were primarily about best prices earlier, it will now be about the airline, routing, layovers, seat choice, meal and entertainment preferences, the accumulation and use of mileage points-and price. All this while making the experience friction-free and dramatically lowering the cognitive burden on the buyer.
Many elements of the value chain will change. With the metamorphosis of the buyer from an involved browser to just stating intent, the relationship between man, machine, and markets will also change. Like the buying journey, especially for those who savour the process more than outcomes. Subliminal preferences won’t ever come into play anymore.
Customer psychology will thus change dramatically and forever, with buying behaviour now being contingent on the logical aspects of buying-availability, accredited reviews, and pricing, maybe even ESG (environmental, social, and governance) scores.
It is likely that the ad world will transform. Creative aspects like branding and website design could be overlooked and banners and videos will have limited impact. Instead, every product detail will need to be machine-readable, must convey comparative superiority, and the entire promotional effort will be to become the singular default choice.
Category pages and search optimisation will be less relevant. New advertising formats will have to focus on agent engagement and unambiguous hand-offs. With algorithmic mediation increasing the bias towards hyper-rationality, brand loyalty would be sacrificed, ignoring emotive hooks embedded in campaigns. AI agents will tend to switch brands, to avail dynamic pricing and opportunistic spot-benefits to maximise utility.
Agentic commerce will change traditional supply-demand equations too. The ordering of predictable, routine items will be entirely system-driven. Monitoring smart devices will smooth the ordering of consumables. But the ability of the AI agent to hold till optimal terms are secured could contribute to volatility. The ability to split orders logically and place orders on multiple retailers will increase fragmentation and change the supply chain rationale.
Also, retailers will have to expose their inventory levels and fulfilment schedules in real time. Models taught at business schools will have to recognise these changes. While the process will speed up end-to-end and hyper-personalisation will be in play, issues like data security and cybersecurity related vulnerabilities will have to be addressed.
Agentic complications and controls
On the lighter side, if the AI agent focuses only on terms, and decides on a large purchase, without due consideration to space and duration of storage, an unsuspecting buyer may suddenly be stuck with barrels of cleaning liquid or pickle, obtained at jaw-dropping discounts, with nowhere to store it and years to consume it. Or if many such autonomous agents stumble upon an item, extremely low priced by oversight, during a festival shopping spree, it could result in a digital stampede.
Clearly, there is a need for agentic negotiation protocols that go well beyond reading the price, to deciphering all the terms and their implications, and they must encompass bidding, escrow, and deal closure aspects more thoroughly. An agent trust registry that shows the dynamic score of trustworthiness, including agentic commitment to pay, will calm nerves.
There are legal and ethical issues to consider too. With total surveillance over the buyer’s habits, financials,
calendar, status, and health, agentic decisioning will have to be embedded within secure frameworks to avoid dark patterns and anti-competitive practices, all happening at lightning speeds.
What will be most intriguing is when the seller is also agentic, would have figured the criteria most often deployed by buying agents, when offers are rapidly customised to close deals faster, when selling agents create urgency with simulated demand, and AI buying agents comprehend all of this, and respond appropriately.
Anil Nair is the founder of ThinkStreet
Disclaimer: Views expressed are personal and do not reflect the official position or policy of FinancialExpress.com. Reproducing this content without permission is prohibited.
