How to decide where AI can be used in your organization

In a world where even LEGO now comes with Bluetooth capabilities, it’s safe to say that technology is everywhere and it’s widely accepted. However, when it comes to  artificial intelligence (AI) – specifically its role in customer experience – it does well to remain cautious about the opportunities that are arising on a near weekly basis, particularly as many organizations are yet to cement or clarify how AI can be best used to enhance CX. 

Over the course of his career, Michael Anderson, VP of client management at Davies, has supported dozens of businesses to automate repetitive processes, improve outcomes and streamline operations, first through robotic process automation (RPA) and later, AI. Yet he’s seen stark disparities in how different companies, and often industries, are understanding and utilizing AI – and it’s just part of why he advises “cautious excitement” around this fast-moving area of technology. 

“Don't put every AI capability you have in front of the customer right now,” Anderson says. “By all means, use some of it where it's appropriate, but not until you understand your processes and your customers’ journeys.” 

Michael’s advice to clients on AI for CX is that it shouldn’t be expected to solve all the challenges that exist. “You shouldn't be expecting AI to totally solve your customer experience problems, or put it to customers in place of your organization at this moment in time.

"AI is developing really fast and we urge a level of caution. It's brilliant and there are lots of applications where you can use it – and, obviously you've got to get on the train at some point – but don't think it's going to solve all your problems,” Anderson says.

However, he does see many successful use cases for AI in CX; whether it’s enhanced self-service through AI-powered FAQs, voice assistants that can understand regional dialects and accents, or employee support through summarization, data analysis and real-time performance analysis. 

“There is a bit of a holy trinity here in how the employee experience, customer experience and operational efficiency and effectiveness can all be supported with automation and AI. If you can use automation to support that holy trinity, then I believe, you’re almost there,” Anderson says.

“Looking ahead, we see the money Meta is putting into augmented reality (AR) and virtual reality (VR) and beyond gaming there aren’t many use cases yet, but I can't help but think that its ability to change retail is just around the corner.”

Introducing AI to the organization 

That’s the corporate side of things. But much like with the debut of Windows 95, individuals are using AI – specifically generative AI – outside of work, and sometimes more than they are using it in their work.

It’s a dynamic that can build a lot of expectation around a certain technology, but it can also see capabilities being introduced to an organization through the backdoor, without the necessary and critical foundations in place. 

Anderson recalls: “Windows 95 first launched in the home. People were going to work using the black screen computers with green text and we were all thinking the computers we have at home, are way better than the ones that we use in work.”

He adds: “In many ways, it's the same with AI. Because we've all been provided with ChatGPT and other products and tools and we’re probably using them. Yet the employees are saying, well, when can we have it at work? When can we have it in our organization?”

The CEOs Michael works with are asking similar questions. Anderson says one CEO was new to their industry and asked what they should do about AI. “And I think that's the general question on everybody’s minds at the moment; what do we do about AI? How do I use it? Some organizations are a bit reticent, whereas other organizations obviously are starting to use AI and adopt it quite a lot at the moment,” Anderson says. 

Whether it’s for organizations or individuals, the message his clear: do your due diligence and ensure you understand the technology. 

Another crucial step for an organization to take is to first automate with less sophisticated tools, such as RPA. 

“I am aware of areas of the financial services sector who still haven't thought about how to use RPA. How are they going to go from being in an analogue world to an AI world? It’s clear there are some who still have quite a journey to make,” Anderson says. 

Optimize or automate? 

Tthe list of ways AI can support CX is quickly stacking up, however, a central piece of the puzzle is still missing: the why behind AI. 

“We do quite a lot of work with organizations to get them to think about whether or not they actually have a CX vision? Not just a company vision and a company mission, but a CX vision. And most haven't,” Anderson discloses.

Recalling one financial services organization which predominantly operated out of the US and Canada, Anderson’s team at Davies implemented an approach called “life moments”, which worked to understand the specific life moments that would prompt a consumer to become a customer of the organization. In total, 23 moments were identified and then analyzed for whether or not the customer’s needs could be met with automated processes with a “value irritant analysis”.

“When you looked at those life moments, you recognized that although they had the brand and they had these products, they actually needed to understand how, in each of those product areas – and across personas for those life moments – that they could provide a service,” Anderson says. “Then, when there are interactions, what interactions were of value to the persona and what were of value to the organization? What was not of value to the persona and not of value to the organization?”

He adds: “You look at some of those interactions and you recognize that you've got to put a white glove treatment around some of them; you would never offshore those interactions or put automation in front of them.”

Despite the clear-cut approach irritants remained for both customers and organization. Anderson says: “At that point we would ask, do we need to do this anymore? Can we optimize the process or can we automate it? Then you start to say ‘OK, so what are the things we need to best serve customers in those moments that we're needed?’ Human centered design is incredibly important, especially when we look at how some organizations are set up, and the complexities that exist as a result in their processes and customer journeys.”

Generally speaking, an organization can reach a point of clarity by reflecting on its principles and therefore, how AI can be used without undermining those principles. 

“One principle could be, are we digital first or digital only? Are we always available? Therefore, if I'm stuck in a journey somewhere, can I pop out and be given a real person any time of day? That then determines how you are going to organize your processes, automation and everything else, and then how that is going to be governed. You don’t want a project that is at odds with your CX vision or any of those principles,” Anderson says. 

In short, anchor everything around the CX vision, from journey to data collection. On that point, Anderson says: “In terms of data governance, ask what data are we collecting? How are we collecting it and what are we doing with it? How are we making sure there's no bias in our data?”

AI also requires governance in its own right. AI governance is a fast-moving and incredibly unforgiving area where seemingly small oversights can lead to mistakes so big they become global headlines. However, when CX Network conducted its 2025 research into the Global State of CX, as many as 48 percent of 353 respondents said their organization does not have governance models in place for generative AI. 

“First and foremost, give somebody the responsibility for AI,” Anderson says to those who are yet to get this crucial piece into place. Davies announced the appointment of its first chief AI officer, Paul O’Brien, in September 2024. “You have to really think about the governance up front and maybe look to some of the regulated industries for guidance, even if the regulators will be different entities,” Anderson says.

In short, while the capabilities are there for anybody to unbox – in the right circumstances – there are five critical steps that are needed to support the successful use of AI for CX: data, journeys, process, vision and governance. 

All Access: The Revolution of AI in CX 2025

Join us to hear from industry leaders, innovators, and CX experts as they share insights, strategies, and tools for harnessing customer feedback to drive meaningful change. 

Register Now