So we sat down with Jamie Griffiths, the managing director of UK-based digital agency Reading Room (recently ranked 7th top UK B2B marketing communications agency), to get a better idea of what AI is—or could be—doing for businesses, the key areas of concern around adoption, how businesses should go about implementing it, and how long they should expect it to take.
Why should companies introduce AI into their business?
I think there is a lot of confusion around AI. People consider it to be quite new. When they think AI, they think of the big-machine learning projects companies like Amazon or Google are working on.
There are, however, practical applications for AI businesses can leverage to genuinely improve the user experience. But a lot of businesses don't realize AI is already very much in use—it just doesn't have the AI label attached to it.
Amazon, for instance, has seen year upon year of growth because its customers keep coming back and spending more. Why? It has the (almost spookily accurate) ability to understand who you are. It knows your likes and dislikes, where you are, the time of day and weather there, your spending habits, your shopping patterns, what you’re currently interested in, etc., and it responds by showing you content tailored specifically to you.
Similarly, by being more responsive to your customer’s demands and enhancing your customer micro moments on your digital real estate, you can boost engagement as you quickly learn and adapt to how your customer is using it, thereby improving conversions.
And many of these things can be achieved using content management systems (CMS), which enable you to create algorithms or rules that manage micro engagements for you. Some CMSs handle this really well.
Things start to get interesting using AI such as chatbots as you’re adding a digital channel for your customers to engage with you through. A channel using a communication interface universal to humans—speech. Going back to Amazon, their Alexa is a fantastically useful tool all round, as the company is listening to what the customers are asking and can respond very easily using that channel.
So there are lots of reasons for companies to introduce AI—the key ones being: improving customer engagement, reducing operating costs, and finding new ways of being wherever your customers need you today.
What should companies consider before even thinking about adopting AI into their business?
Businesses should consider the practical uses of it.
It wasn’t that long ago that smartphones became super-widespread and everyone and their dog wanted presence on Apple’s app store or on Google Play. Many businesses rushed out to build apps and then found that the cost of ownership and of maintaining those apps was quite high, especially as version upgrades meant you needed to then upgrade your applications to work with each new version. Often they also discovered that the take up of mobile app services could be quite low.
We are creatures of habit, and I believe businesses will invest too much time and effort in adopting AI technologies without actually understanding what the implications of it are. To quote Simon Sinek, you’ve got to start with the why. Why are we doing this? Why is this of benefit to our customers? And to us as an organization? Why is this going to help us? Why will this support operational efficiency? Why does it align with our offering and how we work as a business?
You need to be quite considerate about the practical uses of AI for you as an organization and how you're going to manage that. What are the operating costs going to be like? How are you going to support it ongoing?
It’s not dissimilar to the discussion we often have with our customers about what content management system is right for them—if you don’t have the capability to manage it, then it’s quickly going to become irrelevant or even a burden that you’re no longer invested in. If you are not invested, then the customers that have adopted your AI services, rather than becoming advocates of it, become naysayers. So it is really important that you consider the practical applications—how it’s going to help your customers and, just as importantly, how it’s going to help you.
What is a company’s typical resistance to implementing such a transformation and how do you get past it?
You know, it’s very easy to sit in a room and find 50 reasons for not doing something. But sometimes, it's the one or two reasons that you should do it that are the most important. You need to act on those, be bold and brave.
Something we like to do with our clients is rapid prototyping. We get something up really quickly that we can test within the organization or perhaps with their customers or a segment of their target market—we then get fast, accurate feedback quite early on as to whether this is the right direction for the company and what we could improve to make it better. Make it more useful.
This term does have a lot of negative connotations, but I say “let’s fail fast!” At least then we can learn what we can from that failure. Then we’re in a fantastic position to pivot and progress. It’s a really practical approach to take when adopting AI.
Test first. Start small. Select your test group carefully so as to ensure usable feedback rather than testing randomly with people who might not normally engage anyway. We often get our clients to give us access to a segment of their customers, we may recruit according to certain personas, or we can actually go out and test it with people that are using the services.
A good example of this is a Reading Room client—a regional council in the UK—who are looking to shift people away from using cars to using public transport. We literally went into the area in question to speak to people on buses, trains, and the local ferries. We even hung out in petrol stations to talk to car owners about their knowledge of transport in the area. This way we could work out the sorts of things they’d want to know if they were to be persuaded to convert to public transport.
What are the main concerns and pain points when companies are considering introducing AI?
When you start to talk about artificial intelligence, people think HAL from 2001: A Space Odyssey. Very big, very expensive, and very complex.
Cost is always a big concern and tends to be the main objection. The reality is, however, that if you use rapid prototyping, you can quickly get an understanding of what the ROI could be and then you can make a case for that investment. You may well discover it isn’t anywhere near as expensive or complex as you expected it to be. An AI service app doesn’t need to be any more expensive than a small- to medium-sized website. And if it’s well thought out, then you've actually got something that's tangible and meaningful. Another channel to your customer that will create better engagement.
The next big area of concern is around complexity. The key is to make it as simple as possible. You can always add complexity later. Start off with something simple and improve from there.
Then there’s quality. If you are providing information that is based on the data source, what's the quality of that data? Is that data in an acceptable format to present to your customer? Or do you need cleanse the data or add some dressing around it to make it more usable?
The only other significant area of concern that we encounter with clients is them not really knowing how to tackle AI. For this, we recommend: take baby steps. Test it. See whether it works or not and then build from there.
It’s taking baby steps that steers you clear of all these pain points. If you attack the project wholesale tooth and nail from the start, then the issues just become more acute because you're literally crashing into them one after another.
Get something to market that works and starts to build interest as well as measurable input from your customers. Scale as you learn. Remember, of course, you’ve got to maintain it. It's not a case of “set it and forget it”, you’ve got to manage it and keep improving it.
What would you say is the typical time scale is?
Using this alpha-beta approach you can literally get something up in a matter of weeks. We’ve launched chatbot services in three or four weeks from first kickoff meeting to prototype ready for testing with customers.
Something a bit more complex, perhaps more transactional, can take twice that. But there's no reason any business couldn’t expect a prototype AI service app in weeks rather than months (depending on complexity).
So according to Jamie Griffiths, AI is here and here to stay. And, powerful as it can be in customer experience and engagement, it’s nowhere near as inaccessible as it soundis. With a little know-how, small businesses can be implementing AI service apps to be managing micro engagements using their content management systems! But the key is to take baby steps. Start small, get something running quickly, select a test audience, listen to what you learn, and improve as you grow.
Jamie is the Managing Director at Reading Room, having previously carried out a number of different roles in his seven years in the Agency. Jamie has been involved in digital since 1996. As a result, his knowledge and experience have both breadth and depth. He is passionate about all things digital, and says he loves it because there is always something new to learn, or an interesting challenge to provide a solution for. Day to day, Jamie enjoys getting involved in the details of what we are delivering our clients. He is also a leading member of the Kentico Advisory Board in the UK.
Reading Room: "Kentico has been in our blood since 2012"
As a long standing Kentico Gold Partner, we’ve got a proven track record of delivering enterprise-level CMS systems. Our team of dynamic Kentico developers, creative thinkers, skilled strategists, and super support team are the experts when it comes to Kentico. And thanks to its customizable nature, it plays to our strengths. We’ve built many highly successful websites on Kentico for Education, Public Sector, Housing, Membership, Sport, and Commercial sectors.