You can build something really impressive for your customers and still watch them walk straight past it.
I'm a bit like that myself. I still buy most of my books in a bookshop. I know I could get them cheaper and faster online, but I like the process. Browsing the shelves, picking something up, holding it. Buying a book is part of the joy of reading it.
And I'm clearly not alone. The Booksellers Association counted 1,052 independent bookshops in the UK and Ireland at the end of 2024, close to a ten-year high and well up from a low of 867 in 2016. That's nearly 30 years after Amazon started selling books online.
When technology serves us, it gets adopted. When it detracts from life, it gets rejected, however impressive it is.
That matters as much for anyone building products, services or experiences. Your audience is running the same test every time they decide whether to use what you've made. With so many businesses putting AI in front of their customers right now, it's a test worth understanding before you build.
What we do when take-up is low
Organisations often come to us because they have built something, often investing huge amounts of time and money, only to find their customers don't want to use the end result.
When something new isn't being used, the usual response makes sense. Sometimes people really don't know it's there, so you run a launch campaign. You add a few more features, smooth out the onboarding and set usage targets. Some businesses go further and make the phone number harder to find, to nudge people into the new channel.
It's easy to judge something new on what it can do. And if it can do a lot, it's natural to assume the problem is that people haven't found it yet.
But I think most of these responses treat low adoption as an awareness or usability problem. Your audience is also weighing up what it takes away.
What people give up decides it
We don't want to outsource everything. Shopping is something people actually quite enjoy doing, and so is researching holidays and places to go. We might look for some support along the way, but we don't necessarily want the whole thing done for us.
That's why adoption of things like e-commerce can be quite slow. Online accounted for 27.3% of UK retail sales in April 2026, according to the ONS. That's a big number, but it's still below the 37.1% peak in January 2021, when the shops were shut. Given the choice, most of our spending still happens in person.
You only have to look at 3D TV. Around 2010 it was meant to be the future of home entertainment, and by 2017 Sony and LG, the last two manufacturers still making them, had stopped. There were technical reasons and the cost. But for me the bigger issue was what it took away. Everyone in the room needed their own pair of glasses, and the benefit of 3D wasn't worth losing the time with the people around you.
VR in the living room has a similar problem. Meta's Reality Labs has lost around $88 billion since it started reporting in late 2020, and this year Meta said it would cut VR investment significantly. Apple's $3,500 Vision Pro reportedly had its first-year target of one million units more than halved. Even when people are gaming, they want to be present in the room they're in.
What held them back was what people had to give up to use them. Audio, by contrast, adds to the experience without cutting you off from the room.
Social media is the tricky one. It's a brilliant way to stay in touch, and a majority of US teens tell Pew it helps them keep up with their friends' lives. But 48% now say it has a mostly negative effect on people their age, up from 32% in 2022. On the whole it was probably a plus initially. I question that now.
Run the test before you build
What's often missing is a simple step at the start. Look at what you're building from your audience's side, and decide on purpose what the technology should do and where a person stays involved.
This has been studied for decades. The Technology Acceptance Model, developed by Fred Davis in 1989, boils it down to two questions: is it useful, and is it easy to use?
That's a good start, but for me it comes down to a balance. Is there more benefit than negative? And the negative includes what we lose as well as what we spend.
Six questions your audience is asking
When I look at what's been adopted and what's been rejected, the same six things keep coming up:
- Convenience. Does it save time or effort when it matters?
- Joy. Does it take away something we enjoy doing?
- Connection. Does it bring us closer to people, or cut us off from them?
- Control. Do we still feel in charge, or is it making decisions for us?
- Confidence. Do we trust it enough to rely on it?
- Wellbeing. Does it leave us better or worse off over time?
Technology doesn't need to score well on all of them, but the benefit has to outweigh the negative.
It works as a test for anything you're building too. Look at it from your audience's side and go through the six. Where it wins tells you what to lead with, and what it takes away tells you what to fix, or leave alone.
Where AI passes and fails the test
This is why I don't think personal AI agents will be as widely used as some people predict, at least in our personal lives. An agent that plans my holiday or does my shopping? I'm not sure I want that. But difficult tasks, like group organisation and life admin? Yes, agents should do those for us, if they can do them really well.
Trust is a barrier too. Research from RTB House this year found 35% of consumers across the US, UK, France and Japan want a human to approve a purchase before an AI completes it.
Customer service bots show both sides. A bot can be great if you've got a quick, simple question, but it can be terrible if you need complex support. Gartner found 64% of customers would prefer companies didn't use AI for customer service, and their biggest worry was that it would make it harder to reach a person. So making the phone number harder to find takes away the very thing customers were worried about losing.
Then there are the bigger, more complex considered purchases, like buying a flat or a car. They involve a lot of research, comparison and uncertainty, and for most people that part isn't much of a joy. This is where AI will increasingly help, by researching options, comparing them, answering questions and narrowing things down.
But it won't remove the human touch entirely. You can now test drive a Tesla and buy one without seeing a human being, but that's still a minority. It's very rare someone would buy a car without trying it and speaking to someone about it. BYD opened its 125th UK retail site at the end of 2025 and is still investing in showrooms for that reason.
The bigger the decision, the more confidence we need before we act. And for a long time, I believe a human will still need to be involved to add the final layer of trust.
What this means if you're building something
For me the principle is pretty simple. Use technology to take away the effort people resent, and keep them in charge of the parts they enjoy or need to trust.
In practice, that means a few things:
- Start with the value. Be clear what your audience gets from it before deciding what the technology does. Often that's taking away the effort they'd happily hand over.
- Go through the six before you build. Then again before you launch, with real customers if you can.
- Decide where a person stays involved, and make them easy to reach. In a considered purchase, that's often the moment confidence is won or lost.
Get those right and the launch campaign has a much easier job.
Serve a motive or fail
The test for any piece of technology is pretty simple. Does it serve one of our motives?
When it does, it gets adopted. When it doesn't, it fails, however advanced it is.
Your audience is already running this test. Run it first, through their eyes. Is there more benefit than negative for them? If there is, they'll adopt it. If not, they'll reject it.
If you're working out where AI and technology should sit in your customer experience, we'd love to help.
