Business owners of all sizes have one thing in common right now: they all need to form a view on what AI can do for them, where they should use it, how much it will cost, what benefits it brings, and whether they can move before their competitors do. It is an arms race, and not one anyone wants to lose, or so the media, influencers, gurus, vendors, and everyone else with a dog in the race would have you believe.

The fear of missing out with AI is real, and it is neither helpful nor going away. For small business owners in particular, the responsibility sits with them in a way it does not for larger organisations - there are no teams of specialists to defer to, no steering committees to hide behind. The decision is theirs and too often theirs alone, whether good or bad. But amongst all the hype, the new releases and the genuine capability, what rarely gets surfaced is just as important as everything else: where businesses probably should not deploy AI at all.

This is not a case against AI. The technology is here, the trajectory is upward, and the capability is real. If you ask a technologist what you should do with AI you will get a technical answer, if you ask a salesperson, you will get sold AI for everything, but as a small business owner you need to think in terms of bottom lines, customer impact, cost, and benefit. On that basis, AI is not needed everywhere, all the time, in every part of your business, regardless of how good it gets.

The question not being asked enough

The questions business owners tend to bring are the obvious ones: what can AI do for me, how does it work, what does it cost, what happens when it goes wrong? These are all valid and there is no shortage of voices willing to answer them. What far fewer voices are raising is the prior question: where should you not use it? That's a question that might matter more.

A useful starting point is not what AI can do, but what your customers expect from you and your business. In some parts of your business, the customer expectation is I'll encounter a human, and if your customers discover they have been getting a machine instead, the potential consequences are not just a missed sale - they are a broken relationship, a lost referral, and a review you did not see coming.

Sales: the worst place to outsource yourself

Sales is uncomfortable for most small business owners, the bigger the company the more this gets outsourced or hired in for. It is the part where rejection happens, where the uncomfortable questions get asked, and where the worry that you are not good enough can surface just when you need to perform. You want the business, you may genuinely need it, and so the idea of outsourcing it and removing the discomfort, handing it to a system that does not flinch, has a logical appeal. Some owners have also had bad experiences with sales hires who overpromised, misrepresented the business, or simply did not show up in the way the brand needed. The pain is real and the vendors pushing AI sales tools are pressing exactly that button.

But there is a litmus test worth applying before you do anything: how would you feel if every supplier you dealt with, every business partner you approached, every company you transacted with in your own personal life, was represented not by a human but by an AI? Would you trust them? Would you want to give them your money? The discovery call is the first impression of how you will show up when something goes wrong. If you cannot be present to find out whether a prospect is worth your time, you have already answered their question about whether you will be there when they need you.

Personal brand and video: where AI-generated content destroys what it was supposed to build

According to Daniel Priestley, personal brands achieve between 5 and 20 times the reach of business brands. This is not a fringe finding - you can test it yourself by looking at Elon Musk versus Tesla, or Richard Branson versus Virgin, on any social platform. The person consistently outperforms the brand. Customers want to know who they are doing business with. They want to connect, to get a sense of the person, to build enough familiarity that trust becomes possible. Google's own 7-11-4 framework for building that trust is built around content volume and consistency, and video sits at the centre of it because no other format closes the distance between a business and a prospective customer faster.

Most small business owners avoid video. They know it works, they know their competitors are not doing it, and they still find reasons not to. So, the temptation to solve the discomfort with AI-generated images and video is understandable, and the technology to do it is increasingly accessible. But before you take this route, consider what actually happens when a prospect or a customer sees your content and knows it is not you. The trust you may have spent months building through your own face and voice is undermined in a single post. If you've only ever posted content with AI avatars and then they meet you on a call, and you seem different, do you still have the same trust? Recent data across the US, UK and Canada from April 2026 shows declining trust in AI-generated content and a rising preference for genuine human interaction. Your customers are getting better at spotting it, not worse.

The data on what trust actually produces is worth keeping in front of you: a 25% increase in consumer spend when customers trust your brand, 53% more likely to try a new product even at a premium, 55% prepared to stay loyal and give you the benefit of the doubt when something goes wrong, and 63% more likely to recommend you. Word of mouth has always mattered - in an era of AI-generated noise and fake reviews, it matters more than it ever has.

Mapping where your customers actually expect you

Most small businesses have thought hard about how to get attention and land a prospect in a sales conversation, but fewer have mapped what happens to the customer after that. Enterprise organisations invest considerably in understanding and documenting every touchpoint in the customer journey. However, small businesses rarely if ever have that level of resources, but the irony is that small businesses often live or die by the quality of those touchpoints in a way enterprises do not. A large company can absorb a bad customer experience and barely notice, whereas a small business cannot.

The practical exercise is straightforward and involves mapping every single point where a customer could reasonably contact or engage with your business, and at each one ask a single question: "If we used AI here instead of a human, and the customer found out, would they still trust us?" This could be the personalised welcome video after onboarding, or the first time they get in touch with a problem, or the meeting where something important needs to be decided. At each touchpoint, the question is not can AI do this but instead should it.

Some touchpoints will pass that test more easily than others but will also be dependent upon who your customers are, so a simple status update, a booking confirmation, a simple FAQ answered can all be ok. However, others will not pass the test and genuinely damage the relationships that you've spent so long building. The big and easy mistake to avoid with answering this question is assuming without asking.

What the data says your customers actually want

A 6,000-person study across the US, UK and Canada conducted in April 2026 found declining trust in AI customer service and a growing preference for human interaction. Separately, 74% of consumers say they prefer a human for complaints, billing disputes, and any contact that involves real emotional weight, and 82% expect a clear and immediate route to a human when they want one. Perhaps most pointedly, 64% of customers say they would prefer companies did not use AI at all, with difficulty reaching a human cited as the top concern.

This does not mean AI has no place in customer interaction. 68% of those same consumers say they are comfortable with AI for simple status-style questions, and that is a meaningful opportunity to free up time for the interactions that require you and your team. The data is not making the case against AI, it is making the case for thinking about which specific interactions benefit from it and which do not.

Take AI adoption to its logical extreme as a small business owner and the problem becomes clear. An operation of one or a handful of people, served by armies of agents handling every customer touchpoint, looks from the outside identical to every other business doing the same thing. At that point you are being evaluated on features, price, and what strangers have said in reviews. There is no relationship to fall back on, no loyalty earned through the interactions you showed up for, and no reason for a customer to stay when something newer comes along. You have become a listing, not a business.

The assumption that AI belongs everywhere is the problem. Before you deploy it in any customer-facing part of your business, map where you are considering using it, be honest about whether those touchpoints are ones your customers would expect a human for, and then do something most businesses never do: ask your customers directly. Their answer will be more useful than any vendor's recommendation, and considerably more honest than your own assumptions.