Why getting your processes, systems and customer data in order should be the first step in any AI transformation.
There's a lot of excitement about AI in customer service at the moment, and rightly so. The technology is developing quickly, and the opportunities to improve customer experience, reduce costs and make contact centres more efficient are significant.
But there's something I've been saying for the past three or four years, and it's just as relevant today.
Before introducing AI into your contact centre, make sure the basics are working properly.
It sounds obvious, but I still come across organisations that are keen to deploy conversational AI voice bots, a chatbot or automated customer journeys without fully understanding whether their existing processes and systems are ready.
And that's a problem, because if you automate a broken process, all you've really achieved is making that broken process happen faster.
Are businesses as AI-ready as they think?
In my experience, many organisations are less ready for AI than they believe.
That isn't necessarily a criticism. Most businesses have spent years building up different systems, processes and ways of working; technology has been added, departments have evolved, and customer journeys have become increasingly complicated.
The result is often a collection of systems that work reasonably well individually but don't talk to each other or work particularly well together.
A customer might contact a business about an existing enquiry, but the contact centre can't easily access the relevant information held in another system.
An agent might need to open five different applications to understand what's happening with a customer's account.
Or a straightforward refund request might require several manual steps behind the scenes before it can be completed.
These are exactly the sorts of problems businesses should be addressing before introducing AI.
Otherwise, you're simply putting a new interface in front of the same old problems.
Three things I'd look at before deploying AI
When we're assessing how ready an organisation is to introduce AI, there are three areas I particularly want to understand.
1. Do your processes actually work?
This is where I'd start, before looking at any technology.
Take a common customer enquiry and follow it all the way through the business.
What happens when a customer requests a refund, changes their address or books an appointment?
You need to ask:
Does the process work consistently?
Does it produce the correct outcome?
Are there manual interventions along the way?
And are there any unnecessary steps that could be removed?
Importantly, I'm not just interested in how the process is supposed to work on paper.
I want to understand what actually happens in reality, because there's often quite a difference between what’s been designed and what’s been delivered.
If a process relies on someone manually checking information, correcting errors or finding workarounds, introducing AI doesn't automatically solve those problems.
In fact, it can introduce further complications.
Before automating a process, I'd want to be 100% confident that it works reliably from beginning to end as it’s currently designed.
2. Can your systems talk to each other?
The next area is integration.
Your contact centre platform, CRM, back-office applications and knowledge systems all hold information that could help deliver a better customer experience.
But that information is only useful if it's accessible when you need it.
Let's take a simple example.
A customer calls your contact centre from their registered mobile number.
Your contact centre platform recognises that number, checks the CRM, identifies the customer and looks for any open cases or recent interactions.
Based on that information, the call can be routed to the most appropriate team, with the relevant customer details already available to the agent.
It's a relatively straightforward concept, but you'd be surprised how many organisations still aren't doing it and the crazy thing is you don't need AI to make that happen.
You need the right integrations, accurate customer information and a properly designed customer journey.
Some of the biggest customer experience improvements we've delivered have come from making existing systems work better together rather than introducing new technology.
3. Is your customer data and knowledge base reliable?
Even with perfectly connected systems, AI is only going to be as useful as the information it can access.
If your CRM contains outdated telephone numbers, incorrect addresses or incomplete customer cases, the AI may struggle to identify customers or understand their circumstances.
The same applies to your knowledge base.
Imagine deploying a chatbot to answer questions about your products, services or policies.
Where is it getting its information?
If your knowledge base is incomplete, outdated or poorly structured, you're increasing the risk of customers receiving incorrect or inconsistent answers.
And that presents an interesting question.
Would you allow a new member of your customer service team to start answering customer enquiries without providing them with accurate information and proper training?
Of course not.
So why do people let AI loose without providing the correct training and guardrails.
We should be treating AI agents much like human agents.
Give them access to the right information, train them on the processes they're expected to follow, establish clear boundaries and regularly review their performance.
Don't overlook the technology you've already got
Something else I'm seeing is businesses jumping straight towards AI when there are simpler opportunities to improve efficiency.
Robotic Process Automation (RPA), for example, has been around for years and can still be incredibly effective.
If you've got a repetitive, rules-based process that requires information to be moved between systems or a series of predictable actions to be completed, traditional automation may be perfectly adequate.
It doesn't always need an intelligent agent making decisions.
In fact, a sensible approach might be to simplify the process first, automate the predictable elements using existing technology and then introduce AI where it’s actually needed.
There's also the question of what capabilities your current contact centre platform already provides.
Modern cloud contact centre solutions have evolved significantly, and many organisations aren't using all the functionality available to them.
Before investing in another platform or additional AI services, it's worth asking whether your existing technology could already solve some of the problems you're experiencing.
Start narrow and go deep
Once your processes, data and systems are in a good place, then it's time to look seriously at where AI can make a difference.
My advice would be to start with a specific use case rather than trying to automate everything.
For example, you might identify that customers regularly contact you to check the status of an order.
If you have reliable order information, accessible systems and a well-defined process, that's potentially a great opportunity for automation.
Give the AI one job and make sure it does that job exceptionally well.
Then measure the outcome:
Did the customer get the information they needed?
Was the enquiry genuinely resolved?
Did they have to contact you again?
Once you've proved the approach works, you can expand into other areas.
That's a much more sensible strategy than deploying conversational AI across an entire contact centre and hoping it can handle whatever customers throw at it.
The biggest opportunity might not involve AI at all
There's another question I think businesses should be asking before investing heavily in customer-facing AI.
Why are customers contacting us in the first place?
How many enquiries are genuinely necessary, and how many are happening because something elsewhere in the business hasn't worked properly?
Perhaps customers are calling because they haven't received an update.
Perhaps an online process is confusing.
Perhaps they're chasing an action that should have already happened.
These enquiries are often described as failure demand, and they can consume a significant amount of contact centre capacity.
AI can certainly help businesses handle this demand more efficiently.
But wouldn't it be better to eliminate the reason for the contact altogether?
If you can identify and fix a process that's generating hundreds of unnecessary enquiries every week, you've improved the customer experience and reduced operational costs without needing to automate those interactions.
That, to me, is a much better outcome.
So, where should you start?
If you're considering introducing AI into your contact centre, I'd suggest asking yourself five questions:
Do we fully understand our most common customer journeys and where they break down?
Are our existing processes consistent, reliable and working as intended?
Can our contact centre, CRM and back-office systems access and share the information they need?
Is our customer data accurate, and is our knowledge base complete and up to date?
Have we explored what our existing technology can already do before investing in something new?
If you can't confidently answer those questions, I'd be spending some time addressing them before moving ahead with a major AI deployment.
Get the foundations right, then build on them
I'm certainly not against AI. Quite the opposite. I think it presents some fantastic opportunities for contact centres and the wider customer experience industry, but technology alone doesn't deliver a great customer experience.
Good processes, accurate information, connected systems and well-designed customer journeys do much of the hard work.
AI can then build on those foundations and help organisations deliver something even better.
So before asking what AI could do for your contact centre, perhaps start by asking a simpler question.
What's not working properly today, and why haven't we fixed it?
You might be surprised how much you can improve before introducing any AI at all.
How NextGenCX can help
At NextGenCX, we help organisations understand what's working in their contact centres, what's creating unnecessary effort for customers and agents, and where technology could deliver genuine improvements.
Our independent, vendor-agnostic contact centre reviews and future-state blueprints look at your existing processes, customer journeys, systems and opportunities for automation.
The aim isn't to recommend more technology for the sake of it. It's to identify the changes that will make the biggest difference to your customers and your business.
Before investing in your next AI solution, let's make sure you're getting the most from what you've already got.
