A bit like everyone, I spend time talking to our clients about AI, particularly how it can improve customer experience and the way customer service operations work. It’s something I’ve been involved in for a long time, and with the pace of innovation now, it’s amazing seeing what’s possible as the technology continues to develop.
But recently, I’ve found myself becoming a little frustrated with how some of those conversations are going.
A few weeks ago, I was involved in another AI demonstration. It was a good demo, the technology did what it was supposed to do and there were some obvious opportunities to improve its customer service. What interested me, though, was the conversation that followed.
Nobody really questioned whether AI could help. There seemed to be a general acceptance that the capabilities being demonstrated would have a role to play. Instead, the discussion quickly turned to which technology they should choose, whether there might be something better around the corner and how they could be confident they weren’t making a decision they’d regret in a couple of years, or months.
I completely understand those concerns, particularly when you’re making decisions that could have significant financial and operational implications. But it did get me thinking about how often I see organisations in a similar position.
I don’t think the biggest challenge for many organisations today is understanding what AI can do. In fact, I’d argue that most of the people I speak to have a pretty good understanding of where the opportunities are. They know which processes are inefficient, where customers are having poor experiences and which activities are consuming too much of their employees’ time.
The difficulty seems to be deciding what to do about it.
Part of that, I think, is down to just how overwhelming the market has become. A few years ago, organisations had a relatively small number of established options to consider. Today, almost every technology provider has an AI proposition, and the pace of development means that something which looked impressive six months ago can already feel outdated.
I find it difficult enough keeping up with everything that’s happening, and this is the area I work in every day. For someone responsible for running a customer service operation, who also has a hundred other priorities competing for their attention, I can understand why making a decision feels increasingly difficult.
There’s also a genuine fear of getting it wrong. Nobody wants to commit to a platform only to discover that another provider has developed something significantly better. Equally, nobody wants to invest in a particular use case and then find that it doesn’t fit with the wider direction of the business.
These are perfectly reasonable concerns, and I wouldn’t suggest organisations should ignore them. I’ve seen enough technology projects over the years to know that poor decisions, particularly around architecture and integration, can create problems that are difficult and expensive to resolve.
But I do wonder whether, in trying to avoid making the wrong decision, some organisations are inadvertently making another one – which is to do nothing.
And as I was always trained when working as an intra-day analyst, that indecision has consequences too.
The customer who has to wait twenty minutes to get a pretty straightforward question answered doesn’t care that an organisation is still evaluating its AI strategy. Neither does the advisor who spends a significant part of their day moving between systems or completing repetitive tasks that could potentially be automated.
Those problems continue to exist while the organisation debates its options. There is a cost associated with that, both financially and in terms of customer and employee experience, but it’s rarely considered in quite the same way as the cost of investing in a new technology.
I also think we sometimes approach these conversations in the wrong order.
One of the first questions I’m often asked is which AI platform I would recommend. It’s an understandable place to start, but for me, I need to turn the conversation around and spend more time talking about the problem we’re trying to solve.
What is it about the current customer experience that isn’t working particularly well? Where is the organisation spending money or effort that isn’t adding much value? And what does it actually want its customer service operation to look like in two or three years?
I think those discussions are far more useful than starting with a comparison of product capabilities. Once you understand what you’re trying to achieve, you can make much more informed decisions about the technology you’ll need to support it.
That doesn’t mean ignoring the longer-term implications of those decisions. If anything, I think having a clear direction becomes more important as the technology gets easier to deploy. The last thing an organisation needs is a collection of disconnected AI solutions that individually solve small problems but collectively create a much bigger one.
At the same time, I don’t believe you need to work out every part of your future AI architecture before you can start making progress.
Some of the most valuable work I’ve been involved in has started with a relatively straightforward problem and a willingness to test whether AI could make a measurable difference. Not every idea works, and sometimes what you learn is that the process itself needs attention before introducing any new technology. But at least you’re learning something based on what actually happens, rather than what a vendor presentation suggests might happen.
And IMO, I think there’s a lot to be said for that.
Of course, there needs to be a balance. Moving quickly without considering the implications can be just as damaging as spending too long deciding what to do. But I don’t think organisations should confuse having a sensible strategy with needing absolute certainty.
The technology will continue to change. New capabilities will emerge; existing platforms will improve and some of the things we’re excited about today will probably look ordinary in a year. That’s simply the reality of working in this space.
What shouldn’t change quite so quickly is the underlying reason for investing in it.
For me, that’s where the focus needs to be. Understand the customer and business problems, have a reasonably clear idea of where you want to get to and then make decisions that move you in that direction. You can adapt the technology choices along the way, provided you’ve thought carefully about how everything fits together.
I suppose my concern is that we’re reaching a point where the sheer amount of choice is becoming a barrier to the very improvements AI is supposed to enable. And while I understand why organisations want to be cautious, I think we need to be equally conscious of what that caution is costing them.
Because ultimately, customers aren’t waiting for us to work out our AI strategies. They’re just expecting better service.