June 19, 2026
The Future of CX: Where VoC Goes from Here
In this interview, CXM sat down with Carlos Espinosa, Executive Chairman of Clientship, to talk about the future of Voice of the Customer (VoC) and his company’s plans to approach things differently.
The State of VOC Today: An Honest Assessment
The honest picture of VOC maturity across most industries is one of sophisticated tooling and underdeveloped practice. Organisations have access to more customer data than at any point in history. The ability to capture feedback across channels — digital, physical, voice, social — is now largely commoditised. And yet the gap between what customers expect and what businesses deliver continues to widen.
Carlos Espinosa, Executive Chairman at Clientship, thinks the most underutilised signal is the unstructured one: the open-ended comment, the voice recording, the social mention that doesn’t fit neatly into a rating scale.
Espinosa commented “Organisations have become expert at counting. They are still learning how to understand. That is where the next wave of CX maturity will be decided — not in the dashboards, but in what happens downstream of them.”
AI Is Changing Everything — and Nothing
Artificial intelligence has arrived in CX with enormous promise and, in some quarters, enormous hype. Cutting through the vendor noise is challenging, but I asked Carlos directly what impact he is actually seeing due to AI adoption.
Espinosa told me, “Let me be direct about what AI actually changes and what it does not.
What AI genuinely transforms is the speed and depth of signal interpretation. At Clientship, Zeta processes customer feedback using purpose-built neural networks — not generic large language models — because customer sentiment has nuance that off-the-shelf models consistently miss. Industry-specific vocabulary, cultural context, the difference between frustration and resignation in a piece of text: these require models trained on the right data, not repurposed conversational AI.”
He continued, “Fourteen years of building and refining these models has produced something qualitatively different from anything you can obtain by pointing a general-purpose AI at a survey dataset.”
Espinosa added. ” What AI does not automatically change is organisational will. A model that predicts customer attrition before the customer signals it overtly is only powerful if the company has a retention process ready to activate. AI can tell you that a customer is moving toward the exit. It cannot make the business care enough to stop them. That still requires leadership.
From Reactive to Predictive: A Genuine Shift
The traditional VOC paradigm has been reactive. Something happens, the customer gives feedback, the organisation responds. But Espinosa says that model has a fundamental flaw — by the time feedback arrives, much of the damage is already done.
“Pre-abandonment detection changes this entirely. When you can model the behavioural patterns that precede a customer decision to leave — and flag those customers before they signal it overtly — you shift from damage control to prevention. This is not a marginal improvement in CX capability. It is a philosophical reorientation. VOC is no longer a measurement instrument. It becomes a strategic early warning system.
“NPS and CSAT forecasting — generating three-month projections based on current trajectory — extend this further. CX leaders can now enter board conversations with forward-looking data, not just retrospective reporting. That changes the nature of the function entirely.”
The Democratisation of VOC
One of the most important shifts happening in this space is not technical — it is about access. For most of the last decade, VOC technology has been the preserve of specialists, people who understand survey logic, dashboard configuration, and data modelling. The insight has been locked behind a wall of technical complexity that most organisations can’t scale.
Espinosa commented, “The next frontier is making customer intelligence genuinely accessible to everyone who needs it — the general manager asking why his region’s scores dropped last quarter, the product team wanting to understand which features drive attrition, the CEO preparing for a board meeting. They do not want a course in BI tools. They want answers. And they want them now.
“Conversational AI is the mechanism by which this happens. When a platform allows any user to ask a natural language question and receive an expert-level analytical response — synthesising survey data, operational metrics, social signals, and predictive models in real time — the entire relationship between an organisation and its customer data changes.”
This is the exact approach Clientship are taking with Zeta, their Voice of Customer engine within MiyabAI — the integrated platform Clientship has built for end-to-end customer interaction. Zeta has been engineered around the moment when every person in an organisation — regardless of their technical background — can hear the customer clearly and act with confidence.
Co-Creation: The Natural Endpoint of Listening
If the history of VOC is a journey from measuring satisfaction to predicting behaviour to democratising insight, then the next destination is co-creation, according to Espinosa. But is this one step too far when many customers don’t understand the technology?
Espinosa commented, “Inviting customers into the design process — not just to validate decisions already made, but to genuinely shape what is built — is the logical conclusion of a truly customer-centric philosophy. Customers who help design an experience do not leave it lightly.
Where do customers go next, and how do they make this transition to co-creation and abandon previous thinking about VOC?
Espinosa says, “Stop treating VOC as a measurement system and start treating it as a decision-making infrastructure. Every significant business decision — product investment, service redesign, operational change — should have a customer signal attached to it. Not because it is good practice, but because organisations that make decisions with the customer’s voice embedded into them systematically outperform those that do not.”
