AI agents are changing who makes buying decisions

Imagine you own a hotel in Accra. You have invested in comfortable rooms, reliable staff and an attractive website. A traveller in London asks an AI assistant to find a quiet hotel near a meeting, book three nights and stay within a budget. The assistant compares prices, availability, cancellation terms and reviews, then chooses somewhere else. You never had a chance to make your case. Not because the traveller disliked your hotel, but because the assistant could not establish what it needed to know.

The traveller remains your potential customer. But the first gatekeeper was a machine that never walked past your sign or admired your website. That distinction may turn out to be one of the biggest commercial changes of this decade.

Venture capitalist Chamath Palihapitiya made an arresting claim that intellectual property in software had become effectively worthless. His argument followed a discussion of “headless” AI services, including Grok Bot, and Salesforce’s move towards software that agents can operate directly. It was a dramatic verdict. Taken literally, I think it is wrong. Taken seriously, it should worry anyone who runs a business.

To understand why, forget the technical language for a moment. Most of us have spent our digital lives visiting software. We open an airline website to book a flight, a banking app to move money and an accounting system to prepare an invoice. Each company has built a doorway through which we must pass to reach its services. The design of that doorway matters because we are the ones opening it.

Now imagine simply telling an assistant: “Find the most reliable supplier, check the delivery dates, prepare the order and ask me before paying.” Rather than making you navigate six websites, it could communicate with the underlying systems, compare the answers and return with a decision. You would still set the priorities and, for sensitive actions, give approval. But you might never see most of the software involved.

This is what “headless” means. The machinery behind an application can be used separately from the screen in front of it. It does not necessarily mean there is no screen; it means the screen is no longer the only way in. A technical standard called Model Context Protocol (MCP) helps AI assistants discover and use outside tools. Think of it as a common plug, not a master key, a common plug for permissions, security and human consent still matter.

This is more than a futurist’s sketch. In April, Salesforce announced Headless 360, making major capabilities available through interfaces that AI agents can call. Google has introduced a common framework for AI-assisted shopping. And this month Shopify opened access to a catalogue of more than a billion listings so people can use their own AI assistants to shop, subject to approval before purchase. The infrastructure of commerce is being rearranged in plain sight.

The first consequence is that visibility could stop meaning what business owners think it means. For years, the objective was to attract attention: buy advertising, rank on Google, build a beautiful website, create an app. But what happens when the buyer’s assistant does the browsing? Your product may have excellent photographs and an inspiring story. If its price, stock, delivery terms or credentials cannot be checked reliably, it may be overlooked.

Consider a furniture maker in Kumasi whose craftsmanship rivals that of an international supplier. A corporate buyer asks an agent for 200 chairs, delivered by a particular date, with a warranty. The workshop’s problem might not be quality or even price. It could be the absence of dependable digital information about capacity, specifications and fulfilment. A business can be excellent in the physical world and almost invisible in the machine-readable one.

The second consequence is uncomfortable for businesses that profit from customers being too busy to compare. An assistant does not get tired after examining three insurance quotes. It can notice an expensive subscription renewal, test alternative suppliers or question a delivery surcharge. Convenience, long a source of profit for sellers, begins moving towards buyers. Where products are easily substituted, profit margins may face pressure.

Yet there is a paradox here. The easier it becomes to compare everything, the more valuable genuine trust may become. A customer might tell an agent: “Find the cheapest flight, but never with an airline I do not trust.” Or: “Buy the medication only from a properly licensed pharmacy.” A strong reputation could become a standing instruction rather than a fleeting impression. A brand that genuinely earns loyalty may become harder, not easier, to displace.

The third consequence may be the most unsettling. Who decides which businesses an agent considers? Its owner will set instructions, but the agent’s provider controls much of the technology, connections and presentation. An apparently neutral recommendation might reflect incomplete data or a commercial arrangement or a platform’s preferred partners. Google is already experimenting with offers appearing inside AI shopping experiences. The screen may shrink, yet the power of whoever controls the customer’s starting point could grow.

This is why the claim that software intellectual property has become worthless needs challenging. Headless access does not abolish copyright, patents or the value of useful technology. Nor does an open connection mean that a rival can reproduce years of proprietary data, customer relationships, compliance systems or hard-won operational knowledge. Indeed, Salesforce is exposing its capabilities to agents precisely because it believes the underlying platform remains valuable. What is threatened is not all software value, but the assumption that making customers enter your interface guarantees your relevance.

Some screens will survive for excellent reasons. A surgeon needs to examine an image. A finance director needs to interrogate a report. A customer signing a mortgage needs to understand the terms. The more serious the decision, the more important it may be to show evidence and retain human judgement. The future is not necessarily a world without interfaces; it is one where interfaces appear when they help rather than because a vendor requires another login.

There is a further risk for ordinary people. An agent given access to money, confidential files and commercial relationships can make mistakes at extraordinary speed. Fraudsters will try to manipulate it. Businesses may dispute who authorised an order or who bears responsibility for an error. A customer should not have to surrender control in exchange for convenience. Consent, audit trails, accountability and clear limits on automated decisions will become commercial necessities, not minor technical details.

So what should a business owner do? Start with an unusually simple test. Could a competent assistant establish what you sell, what it costs, whether it is available, when it can be delivered and what happens if something goes wrong? Could it contact you or transact through a trustworthy digital process? Do your records match what you promise? You do not need to build an elaborate AI system tomorrow. You do need information, processes and service standards that withstand inspection.

For smaller businesses, especially across emerging markets, this is both a threat and an opening. The firm with the largest advertising budget need not always win if a machine can compare actual performance. But the small firm that has never digitised its inventory or documented its reliability may be excluded before the competition even begins. In an age obsessed with artificial intelligence, accurate ordinary information could become a surprisingly powerful asset.

Entrepreneurs should ask one more question: if an agent can reproduce the convenience of my product, what remains difficult to copy? It might be exceptional service, dependable fulfilment, a specialist dataset, regulatory trust, a community or simply a promise consistently kept. Those are less glamorous than a dazzling app. They are also closer to what customers have always paid for.

Palihapitiya’s prophecy that intellectual property in software has become effectively worthless may be exaggerated but it points to a profound shift. The important contest is no longer just for people’s attention. Increasingly, it will be for the confidence of the systems acting on their behalf. The businesses that endure will combine two qualities that sound contradictory: they will be easy for machines to understand and worth trusting by humans.

For generations, business owners have asked how to get customers through the door. The next question is harder: will the customer’s agent even know that your door exists? Think about that for a bit.

Thank you for reading. I welcome your reflections, questions, and suggestions for future topics. Subscribe to the ‘Entrepreneur In You’ newsletter here: https://lnkd.in/d-hgCVPy, follow me on all social platforms at @thisisthemax, or get weekly updates via my official WhatsApp channel: www.bit.ly/whatsappthemax.

Wishing you a purposeful and successful week ahead!

'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'-'

The author, Dr. Maxwell Ampong, is an entrepreneur, investment and strategy executive & CEO of Maxwell Investments Group. His work spans finance, technology, sustainability, trade, enterprise development and institutional partnerships. He serves in executive and advisory capacities across business, academia, culture, agriculture and social development. He is Nana Kwadwo Marfo Ampong I, Aboafuohene of Kona in Asanteman. Dr. Ampong holds postgraduate qualifications in finance, law, business administration, international relations, artificial intelligence and machine learning. He writes on economics, investment, entrepreneurship, technology, social and sustainable development. Entrepreneur In You operates under the auspices of the Africa School of Entrepreneurship, an initiative of Maxwell Investments Group.

Disclaimer: The views, thoughts, and opinions expressed in this article are solely those of the author, Dr. Maxwell Ampong, and do not necessarily reflect the official policy, position, or beliefs of Maxwell Investments Group or any of its affiliates. Any references to policy or regulation reflect the author’s interpretation and are not intended to represent the formal stance of Maxwell Investments Group. This content is provided for informational purposes only and does not constitute legal, financial, or investment advice. Readers should seek independent advice before making any decisions based on this material. Maxwell Investments Group assumes no responsibility or liability for any errors or omissions in the content or for any actions taken based on the information provided.