AI has arrived at that moment which every new, hyped technology invariably reaches: everyone thinks they should be using it, but few have figured out exactly what for. When the opportunities for AI are almost infinite, where do you even start?
The first and most important point to make is, if you’re thinking about implementing the technology because it’s cool and everyone’s talking about it, then you’re starting from the wrong place. Wanting to ‘do AI’ is not a strategy.
We say, as we always do: focus on the customer. Start with the experience you want your customers to have, in a way that makes sense for the business. From there, you work your way back to the technology.
Think of it like a Venn diagram of customer needs, business needs, and technology features. By aiming for the sweet spot where all three intersect, you’ll deliver richer interactions for the customer and better outcomes for the business.
From the customer’s perspective, AI has the potential to deliver loads of benefits. One area where I see AI working well is in enabling customers to carry out some defined self-service tasks, like sending a WhatsApp to a utility provider to tell them you’re moving address, or letting your insurance company know you’re hiring a car while on holidays, and to transfer the policy.
Most self-care portals are built for all customer types, whereas with AI, companies can personalise the interaction to a much greater degree. Once a customer’s identity is already verified on the system, those are achievable use cases.
Another huge benefit from AI is empowerment through information. Let’s pick an obvious example: someone’s browsing the web and they’ve got questions about a particular product. Will it work out of the box? Do I need to make sure it’s compatible with a product I already own?
The example might be obvious but the solution is anything but routine. It’s a big deal to be able to answer a customer request immediately, 24/7. Most businesses today can't guarantee that; chances are, the customer experience will be closer to calling a helpline and waiting on hold for anywhere between 15 minutes to an hour.
By serving them relevant information, you can effectively short-circuit that messy middle in the sales cycle and bring that purchasing decision much further forward than it would otherwise be.
To give a simple example, shopping for clothes or shoes online, with AI there are a lot of solutions where you can use your smartphone camera to see what the garment would look like on you.
The great thing about exploring this part is, many businesses already have a wealth of useful data about their customers already, and they can use this when building an AI tool. They can populate web pages with real data that shows visitors ‘customers like you also looked at…’ and serve them a set of complementary products.
Let’s think of a B2B example now: if you’re a SaaS business, you know all about your customers’ usage of your product. What if you could proactively prompt your customer to move from one pricing tier to another that’s better suited to their usage patterns?
The recurring theme with all of the examples is optimising for a better customer experience. Investing in great UX is also an investment in the business, because not only does it increase sales but it means your customer is less likely to need to contact support.
This leads us to the second part of our Venn diagram: the business needs perspective. Here too, AI makes a lot of sense, because you can use it to enhance efficiency, reduce costs, or improve scalability.
If your business is going international, you can make your online presence multilingual with very little effort. You can train the models to stay within certain boundaries to the point where it's 100% on message, providing personalised customer experiences that an agent who’s answering several calls or chats in parallel can find difficult. When properly trained, AIs reduce the number of errors – and unlike humans, AIs don’t get tired or distracted.
But if you jump straight in without a clear plan, you won’t have clear benefits identified and you’ll have no way to measure them.
One of the most proven ways to identify real, valuable opportunities is by deep analysis of customer journeys. (If you identified a part of internal operations that you want to transform, then your ‘customer’, in this case, will be your internal users.)
In an afternoon’s workshop, we can flesh out concepts and see where they’re technically feasible. So you can plot a course to integrate AI services to optimise internal operations, enhance the customer lifecycle, or simplify acquisition paths.
During the workshop, we probe the challenges that customers might be experiencing. What are your biggest pain points and how can we overcome them?
Together with the help of our technology partners, we’ll look at technical solutions and tools based on the identified use cases, not just because some market research company said ‘this AI tool is trending’. It’s a systematic way of keeping the customer at the centre of your focus.
This is the same approach we take when a client engages us to roll out a website or to design their internal software: we look at what they have and we’ll advise them who’s the most relevant partner for their specific needs. So if someone has a lot of Microsoft-based services in their technology stack, we might recommend a particular partner.
It’s a flexible, bespoke approach; no cookie cutter solutions.
Whether you choose to apply AI to internal operations or out to external customers will depend on your needs and objectives.
It’s all part of a staged process that takes us from insight to testing and scaling. By design, the testing and the engagement are fast: pick an idea and run a test that gives a go/no go result. The first experiment is live in as little as four weeks. From there, we can either iterate the prototype or run another prototype on the next area.
Another point to make with AI is that you can get to a working model at a lower cost than with traditional software development, where you have to write code for every single use case, which gets very expensive to build and very expensive to test.
Generally, the level of integration with AI is nowhere near the level of development that would otherwise be needed. Overall, you save time and money in what you’re building and you save time and money in how it operates in the long run.
When it comes to AI, the best move you can make is to remove yourself from the hype cycle and stay grounded in practical, real-world applications. Brands can do great things with AI and it’s up to us to show them how.