Like it or not, AI is now embedded into almost every business plan for 2025.
It’s in your tools, your inbox, and your daily workflows. It’s being explored by leadership teams, championed by innovation leads, and trialled by product owners. But when it comes to applying AI meaningfully, not just using it, but using it well.
Do you build your own custom AI solution? Buy something off the shelf? Bring in a consultant with a stack of whiteboards and acronyms? Or just quietly close the tab and hope this whole AI thing blows over?
If you’re stuck in decision limbo, you’re not alone. Here’s a no-nonsense look at the three main approaches to AI in business.
You don’t need a robot. You need a plan
This isn’t about building sentient assistants. It’s about identifying the manual, repetitive, and time-consuming tasks within your business and determining if they can be improved. AI is one potential solution, but without a clear strategy, you might end up with a tool that doesn't deliver the desired results.
This is where AI consultation starts making sense. Rather than jumping straight into tools and tech, it’s about making sure everyone’s on the same page. That means stripping away the jargon, getting clear on where AI fits (and where it doesn’t), and building a plan that’s actually relevant to your organisation. Every step is focused on purpose and practicality, so you’re not just adding AI for the sake of it, but because it genuinely supports your goals.
Option 1: Build it yourself (or with a little help)
When building makes sense
Building your own AI solution can be a good shout when your business needs something tailored, scalable, or integrated with your existing systems. If your goal is specific, and ready-made tools are too generic, then a custom build might be the answer.
Think: custom mobile app development services, a bespoke recommendation engine, or something that handles niche workflows in your organisation.
If that’s the case, you’ll want a development partner who doesn’t just write code, but helps you work out what should be built in the first place. (We happen to know one of those - hello.)
But proceed with caution
Building is the long game. It takes time, planning, budget, and ongoing support. And if you haven’t done your homework, there’s a real risk you’ll invest in something shiny and smart that no one actually uses.
User research, prototyping, and testing are your friends here. It’s why we bundle them into our Discovery Sprints and Prototype Design and Development work. Because building the wrong thing well is still the wrong thing.
Option 2: Buy it off the shelf
When buying makes sense
There are loads of solid, ready-made tools that do clever things with AI. Sentiment analysis? Check. Image recognition? Absolutely. Chatbots that don’t want to murder your brand reputation? Finally, yes.
Buying is usually faster, cheaper (at least up front), and gets you something functional without hiring a team of machine learning engineers. Perfect for experimentation or solving common, well-defined problems.
The trade-offs
Off-the-shelf tools are great until they’re not. Limited customisation, monthly fees that quietly balloon, and the ever-looming threat of vendor lock-in can make things tricky in the long run.
Also, just because a tool claims to use AI doesn’t mean it’ll work well for your users or your sector. A chatbot that works brilliantly for eCommerce might completely flop in higher education.
And then there’s the “integration tax”. Getting tools to play nicely with your existing systems is rarely as straightforward as the sales demo suggests.
Calling in a consultant
When consultancy makes sense
Sometimes you don’t need to build or buy, you need to figure out what you actually need in the first place.
That’s where consulting can be useful. A good AI consultant will help you define the problem, explore potential solutions, assess feasibility, and map a sensible route forward. All before a single line of code is written.
Just be careful who you call
Not all consultants are created equal. If they only offer high-level advice with no delivery experience, or suggest generic solutions without digging into your business, walk away slowly.
We believe strategy should be rooted in real-life delivery, which is why our AI consultation services sit alongside development, UX, and support under one roof.
Still unsure? Ask yourself this
If you’re on the fence, these questions might help you nudge yourself off it:
- Is my business problem unique, or can it be solved with an existing tool?
- Do I need something now, or can I invest time to build it right?
- What’s my risk appetite?
- Do we have internal skills, or will we rely on external partners?
- Are we experimenting, or committing to long-term change?
What we’ve learned from years of helping businesses figure this stuff out
Plenty of organisations have hit the usual AI crossroads, whether it's forcing GPT into things it was never meant for, or missing out on small, valuable wins because they thought AI had to mean building a robot.
The truth is, there’s rarely a single right answer. It’s often a blend.
Maybe you start with an off-the-shelf tool, see where it adds value, and build something custom down the line. Maybe you just need clarity and strategy first, not code.
Whatever the case, we’re not here to push one approach. We’re here to help you figure out what works for your users, your goals, and your budget.
In short
- Build if you need control and have a clear, unique use case.
- Buy if your problem is common and you need to move quickly.
- Consult if you’re not sure what the problem even is yet.
Need help deciding? Let’s chat. No scripts, no fluff, just a straightforward conversation about what makes sense for your business.