May 05, 2026
When Your Customer Is a Machine, Who Answers the Survey?
Voice of the Customer (VoC) software companies have enjoyed a comfortable arrangement over the last few years. Brands pay six- and seven-figure annual licences to collect feedback, run sentiment analysis, and produce dashboards that tell senior leaders what their customers think. Qualtrics, Medallia, Sprinklr, InMoment, and their peers built recurring-revenue empires on this model. Gartner kept publishing Magic Quadrants, procurement kept renewing contracts, and over time, the insight layer came to feel like a permanent fixture of the enterprise CX stack. It may not be.
A Forrester blog post published last month, titled Insight Was Never The Point: Arise, Systems Of Action, made one of the most direct arguments yet that AI is eating the VoC value proposition from two directions at once.
The first is easy enough to grasp — if a large language model can synthesise every support ticket, review, social media mention, and chat transcript into a coherent picture of customer sentiment without a survey, a lengthy implementation, or a six-figure licence, the commercial logic of traditional VoC software starts to weaken.
Meanwhile, as AI agents increasingly mediate the customer relationship by planning purchases, filing complaints, and switching providers, there may soon be fewer humans on the other end of the post-interaction survey.
The Feedback Loop Is Thinning
Recent research found that 14% of UK consumers already use AI agents to interact with brands and make purchases, and projects that figure will more than double to 37% by the end of this year. Separately, Gartner forecasts that conversational assistants will resolve 70% of customer service journeys by 2028, and that AI systems will autonomously conduct up to 80% of common service conversations by 2029. When the customer reaching your contact centre is an algorithm, and the agent handling the query is also an algorithm, the traditional survey becomes a bystander in its own process.
The survey model was already under strain before AI entered the conversation. Customers are reluctant to answer email surveys, with average response rates sitting at just five to 15%. The vast majority of customers never respond at all, leaving organisations to make decisions based on a small, self-selecting sample of the angriest and the most enthusiastic. Qualtrics itself acknowledged the problem by launching Conversational Feedback, an AI feature that prompts respondents to elaborate on shallow answers.
If surveys were delivering the depth that CX teams need, why does an AI need to coax more out of every response?
Wall Street Is Already Pricing the Risk
In March, JPMorgan and a consortium of banks pulled a $5.3 billion debt deal that was supposed to finance Qualtrics’ $6.75 billion acquisition of Press Ganey Forsta, a healthcare data analytics firm. The reason cited by investors was concern that AI advances would make many software offerings redundant. At the same time, software debt in collateralised loan obligations had recorded the worst total returns of any sector in early 2026. Wall Street, in other words, is pricing in the possibility that the insight layer as a standalone business may not justify the leverage being stacked on top of it.
The M&A activity around VoC only reinforces the sense of an industry in transition. InMoment was acquired by Press Ganey Forsta, which was then targeted by Qualtrics. Forrester analysts expect Qualtrics to integrate components of both companies and eventually sunset overlapping platforms. In the 2026 Gartner VoC Magic Quadrant, Verint and Concentrix both dropped into the Niche Player quadrant from previous challenger and visionary spots. Alida disappeared from the report entirely in the prior cycle.
Outcomes as the New Product
Meanwhile, a new generation of companies is building around the premise that insight without action is a depreciating asset. Decagon, which tripled its valuation to $4.5 billion in a recent funding round after signing more than 100 enterprise deals, handles three-quarters of customer contacts autonomously and charges nothing for those it fails to resolve. TheyDo raised $50 million on the proposition that the customer journey is the operating model, not a slide deck for a quarterly business review. Unwrap is built on the premise that zero-shot AI, insight generation without pre-defined tagging, can replace the taxonomy-building overhead that traditional VoC platforms demand.
The incumbents know the ground is moving. Medallia, Qualtrics, and Forsta have all signalled their intention to move from “insight to action.” Salesforce presented the same idea in February as an architectural principle, describing the migration from systems of record to systems of action, where agents execute decisions inside governed parameters rather than surfacing feedback.
Bill Staikos, Managing Partner at Be Customer Led, recently weighed in on the topic: “This has been the problem with too much of the VoC world for years. SaaS built a whole category around collecting, tagging, sorting, scoring, and circulating feedback, then acted like the job was done. But it really wasn’t. It’s just a more expensive way to admire a problem.”
None of this means Qualtrics or Medallia will vanish overnight. Both retain enormous enterprise footprints, deep regulatory expertise in sectors like healthcare, and data assets that are difficult to replicate. More than a third of Qualtrics customers have upgraded to its AI capabilities. The question is whether the category they defined retains its current scale and pricing power.
