Why Fixing the Contact Centre Won’t Fix Your Customer Experience
The transformation trap
27 April 2026
In a digital world, the contact centre is no longer the only front door. The real CX failure is happening across your whole operation and the fix can’t start with the platform.
The contact centre is not the problem
You upgraded your contact centre platform. Maybe you deployed new CCaaS. Possibly an AI chatbot. And contact volumes haven’t fallen. Costs haven’t moved. Agents are still overwhelmed. What went wrong?
Nothing went wrong with the technology. The problem is where you looked for the solution.
Contact centre transformation is the most common response to a CX problem that, in most organisations, starts somewhere else entirely. To understand why, you need to see the full picture of what sits between your customers and the service they’re trying to reach.
The operating gap and the workflow tangle
There is a gap between the tasks customers, users and partners need to get done and the systems that run the business. Call it the operating gap. It’s where service happens and it’s rarely clean.
Inside that gap lives the workflow tangle: Siloed SaaS applications, email threads, spreadsheets, manual processes and disconnected data — none of which were designed to work together. They’ve accumulated over years of well-intentioned fixes. Each fix solved an immediate problem and created a longer-term one.
This is where the transformation trap closes. Any improvement attempt that doesn’t remove the tangle just adds to it. A new integration sits on top of an old one. A new exception path bypasses a broken workflow rather than fixing it. A new AI deployment inherits the same incomplete data as the manual process it replaced. Making meaningful change gets harder with every layer.
The contact centre absorbs the cost — it doesn’t cause it
The contact centre is no longer the only front door. Customers engage through web portals, apps, chatbots, email and social channels before they ever speak to a person. When those digital journeys fail — when a self-service form times out, when a payment isn’t confirmed, when an account update doesn’t propagate — the failure lands in the contact centre as a phone call.
John Seddon’s concept of failure demand is useful here: Contacts that exist not because customers want to engage, but because something went wrong upstream. A customer calls to ask why their payment hasn’t been confirmed. The contact centre handles it. The billing notification process caused it. That’s failure demand and, in most organisations, it accounts for a significant share of total contact volume.
Thinking of this as a contact centre problem is leading you into the transformation trap. It’s a whole-business CX architecture problem. The contact centre is just where the bill arrives.
What digital really means in an omnichannel world
Most organisations have added digital channels without changing the operating model underneath them. They’ve launched chatbots, built apps and opened web portals and then wondered why overall contact volumes haven’t dropped.
“Can every workflow in our business serve any entry point, with the right content, at the right moment?”
The answer is that digital transformation isn’t about adding channels. It’s about changing what happens when a customer uses any of them. The real question isn’t “Do we have a chatbot?”. It’s “Can every workflow in our business serve any entry point, with the right context, at the right moment?” For most organisations, the answer is no.
The tangle gets worse before it gets better
When digital fixes are layered over unchanged processes, each one creates new complexity. A new channel integration. A new exception path for customers who don’t fit the digital journey. A new hand-off point between systems that were never built to talk to each other. Every one of these is a future failure point and a new thread in the workflow tangle.
Agents and back-office teams feel this from the other side. Multiple systems open at once. Manual rekeying between platforms. No single view of what the customer has already tried.
A customer’s CX failure is often the direct result of an operational failure faced by the employee in parallel.
This is why omnichannel customer experience is harder than it sounds. Bolting channels together on top of disconnected data doesn’t create an omnichannel experience — it creates a multi-channel one with more seams. True omnichannel CX strategy only works when the orchestration layer beneath it is unified.
A journey in three acts
A customer notices an error on their account. They log a query through the app. No resolution, so they escalate to webchat. The agent asks them to repeat the issue — the app interaction isn’t visible. They’re transferred to a specialist, who asks again. By the time they call, they’re explaining the problem for the fourth time to someone with no record of the previous three.
No single channel failed. The architecture failed. Context was never shared, so every touchpoint was a cold start. The contact centre ends up holding a complaint that became harder — and more expensive — with every transition it made.
What AI actually means for contact centres – and where it goes wrong
The promise of AI in contact centres is real. Done well, it can reduce avoidable demand, improve first contact resolution and give agents capabilities that previously took years to develop. The efficiency gains on offer are significant.
But there’s a distinction that most AI deployment plans miss — and it determines whether AI reduces your contact volume or simply handles it more efficiently.
Where you put AI matters more than which AI you use
AI embedded where calls are created — in digital journeys, in back-office processes, in proactive customer communications — reduces call volume. It resolves demand before it reaches the contact centre.
AI deployed only at the contact centre layer handles calls more efficiently. That matters. But it doesn’t prevent them. You’re optimising a symptom rather than closing the gap that causes it.
The most impactful AI deployments are mapped to specific call drivers. If your biggest source of failure demand is customers calling to chase case updates, the AI intervention belongs in the case management process — proactive updates, automated status notifications — not just in the agent desktop.
The risk of deploying AI into the tangle
As we explored in our previous article: Why the customer journey keeps breaking, agentic AI deployed on top of fragmented data doesn’t just fail to resolve demand — it can generate new demand. An AI agent will pursue its objective using whatever information it has. If that information is incomplete, it won’t pause to check. It will execute. Confidently, quickly and wrongly.
This is the transformation trap at speed. Workflow automation that was already unreliable doesn’t become reliable because an AI is running it. It becomes faster at being unreliable.
Effective AI deployment requires the underlying data and workflow to be connected and trustworthy first. Then AI doesn’t just handle contact — it removes the conditions that generate it. Agent-assist AI, given the right context, means a new advisor can operate at the level of someone with three years’ experience. That’s the compounding return. But it starts with the data, not the model.
AI readiness is a data and orchestration question. You can’t buy your way to AI-ready CX.
What good CX transformation looks like
The most common failure mode in CX transformation is starting with the tool. We’re implementing CCaaS. We’re rolling out a DXP (digital experience platform). We’re deploying an AI assistant. The technology decision is made first and the journey design follows — if it happens at all.
Journey-first transformation inverts this. You start by mapping what happens when a customer contacts you about a specific issue. You trace the journey end-to-end: Where context is lost, where workflows break, where hand-offs fail, where demand is generated that shouldn’t exist. Then you design the resolution. Then you choose or configure the technology to serve it.
The technology isn’t the transformation. It’s what the transformation runs on.
The consolidation discipline
The other discipline that separates lasting transformation from another layer in the tangle is consolidation. Most organisations have accumulated a fragmented estate of point tools — each bought to solve a specific problem, each now adding maintenance overhead, integration debt and governance risk.
Every new tool or AI deployment added on top of existing fragmentation is a future complexity problem. The transformation discipline is to remove as you add — to replace point tools with a unified platform rather than extend the estate. The unifying, composable orchestration layer is the platform. The point tools are what it replaces.
What that unified architecture enables: a single orchestration engine across front-of-house (CCaaS, DXP) and back-office (workflows, RPA, process mapping and improvement). Context that travels with the customer from first contact to resolution. AI that is grounded in complete data. Metrics that reveal journey performance, not just contact centre performance.
The result — as we identified in our previous blog — is a 1/5 to 1/3 reduction in cost-to-serve when organisations get connected data and orchestration right. That’s the outcome of journey-first transformation. Not a feature of any particular tool.
What 90 days can change – working together
The most common objection to CX transformation isn’t scepticism about the outcome. It’s capacity. “We can’t take on an 18-month programme right now.” That’s a reasonable position. Full transformation does take time.
The first 90 days don’t have to.
The right starting point is a focused diagnostic: Identify the two or three journeys generating the most failure demand and avoidable contact. Map the workflow and data gaps within those journeys. Deploy targeted orchestration or automation to close the highest-impact gaps. Establish the baseline metrics that will track progress from there.
That’s not a transformation programme. It’s the foundation one — and it delivers measurable results while it builds.
Proof in practice
Reduced processing time from days to hours by automating manual workflows with Liberty Create — without disrupting the systems already in place.
Saved £280,000+ by replacing a fragmented point-tool estate with a single platform.
Surfaced clinical context — mobility flags, transport arrangements, appointment types — directly to agents, with zero changes to legacy systems. Each of these started with a specific journey failure, not a technology decision.
Three principles to take away
Start with the customer journey, not the platform. Technology serves the journey design — it doesn’t define it.
Target efforts as to why calls are created, not just where they arrive. That’s how you reduce demand, not just handle it more efficiently.
90 days is enough to find the fracture points, close the highest-impact gaps and build the foundation for what comes next.
For more practical action, download the Future Ready CX Playbook for the detailed framework — including how to map your highest-impact journeys, assess your orchestration readiness and build the business case for transformation.
About the author
Richard Higginbotham
Product Marketing – Intelligent Automation
Richard and his team bring the Liberty platform to life – showing how people and AI can work better together. With a background in transformation, data and enterprise tech, he’s helped organisations across sectors modernise operations and reimagine service delivery, delivering human-centric solutions that make work smarter, safer and more effective.