Smart Web Agency started about 7 years ago. I’m a software engineer by background, and initially I was simply building websites and software for businesses. One project led to another, the team grew and today we work across the UK, UAE, Australasia and Canada.

I have always been fascinated by where technology is going. I read a lot of books, but running a business has taught me there is often a big difference between what sounds exciting in a book and what a customer actually needs on Monday morning.

We have clients running software that is 10 or even 15 years old. They don’t necessarily need an AI transformation. Sometimes they just need better search, one repetitive job automated or their existing systems connected properly.

We see the same internally. We use AI every day for development, research, documentation and repetitive work but we still have developers, designers and project managers. Good people actually matter more now because someone has to question what AI produces and take responsibility for the outcome.

After 7 years, that is probably what I have learned most. Technology changes incredibly quickly but the problems customers bring to us are still very human.

Most of our new business still comes from referrals, existing clients and people who have worked with us before. After 7 years, that network becomes quite powerful. We have clients who first came to us for a website years ago and now use us for software, integrations, AI or just call when they have a technology problem they need help solving.

We also get business through Google, LinkedIn and partnerships but referrals are still the strongest because the trust is already there.

Where would I like more business to come from? Probably our own content and reputation. I spend a lot of time reading, experimenting with technology and talking to our team about what is actually changing, particularly around AI. I would like more people to discover us because they read something useful we shared or saw a real project we worked on, rather than because we were simply good at marketing ourselves.

I think that is a much healthier way for a technology business to grow.

One thing I have almost completely handed over to AI is meeting notes. I used to finish a call, write down the important points and then turn them into actions for the team. Now AI transcribes the meeting, summarises it and pulls out the actions. It sounds small but when you have several meetings a day, you really notice the time it gives back.

One thing I tried handing over was reviewing job applications. We get a lot of CVs, so using AI to shortlist them seemed obvious. I stopped doing it quite quickly. It was good at matching skills and keywords but sometimes the interesting candidate is the person who doesn’t have the perfect CV. There might be something in their journey, attitude or experience that makes you want to speak to them.

So AI can organise the information for me but I still want a human making that judgement.

Yes, definitely. The biggest change is that customers are much more informed before they speak to us. A few years ago, the first call might start with ‘what can you do?’ Now people often arrive having already researched the technology, compared approaches and sometimes even have a rough idea of what they think the solution should be.

We are also starting to see people find us through ChatGPT and other AI tools. We have had enquiries where someone has literally said, ‘ChatGPT recommended you.’ The numbers are still nowhere near Google or referrals but a couple of years ago that traffic basically didn’t exist.

What I find more interesting is how this changes trust. With Google, people might open five or ten agency websites and compare them. With an AI assistant, they may ask for three recommendations and arrive already believing those companies are worth speaking to.

That makes reputation, useful content, genuine reviews and having evidence of what you have actually done much more important than simply ranking for a keyword.

I would tell them to use AI every day but stop trying to use it for everything. Pick three or four boring jobs that take time every week and automate those first. Meeting notes, documentation, research, testing or searching internal information. The boring use cases often give you the quickest return.

I would also tell them to invest in their people. Give developers, designers and project managers access to the tools and let them experiment. We have found that AI works best when someone already understands the job. A good developer with AI is incredibly productive. Giving AI to someone who doesn’t understand the problem can just help them make mistakes faster.

For customers, I would focus on being genuinely useful online. Answer the questions customers actually ask, publish real experience and make sure AI assistants can understand who you are and what you have done. People are already changing how they search.

What would I ignore? Most predictions about how many jobs AI will replace by a particular year and the pressure to rebuild your whole business around the latest AI tool. The tools will change again in six months.

Learn the technology, use it heavily but don’t confuse adopting AI with having a strategy.