The Future of Surveys Is a Conversation, Not a Form

By Caplena

  • article
  • AI
  • Artificial Intelligence
  • Customer Experience (CX) Feedback
  • Data Analytics
  • Dashboards
  • Survey Analysis
  • Conversational Surveys

Summarise with AI

ChatGPT
Claude Logo
Gemini Logo
Perplexity Logo

I’m often asked where surveys are heading next, and my answer tends to surprise people. Surveys are here to stay, because companies will always need to ask people questions and listen to their answers. What’s disappearing is the long form, with boxes to tick and a bar at the top telling you how much further you have to go.

In its place, I see two shifts coming, and below I’ll walk through each one and what it means for how teams design feedback programmes.

Voice Becomes the Default Way People Respond

The first shift is in how people actually answer. On mobile, voice will become the first and default mode of giving feedback, much closer to having a conversation than filling in a dull form. That matters because the quality of open-ended responses has always depended on how much effort it takes someone to give them. 

A form with text boxes asks people to stop, think and type, which is exactly the kind of friction that produces short, low-effort answers. Speaking into a phone removes most of that friction. People talk more naturally, say more and give you richer material to work with, which is good news for anyone trying to extract genuine insight.

This matters more than it might sound, because open-ended text has always been the richest form of feedback available. The real value, though, has never come from the open text alone, but from combining it with the structured data around it: who the person is, when they responded and what product, route or service they’re talking about. 

For example, for an airline, that combination lets you move from a vague statement like “people are unhappy with the dessert selection” to something actionable, such as noticing that senators flying from Berlin to JFK have recently become unhappy with the desserts specifically. Voice input, properly analysed alongside that metadata, should only make that kind of precision easier to reach.

From One-Way Feedback To a Two-Sided Conversation

The second, and in my view more important, shift is that feedback will stop being one-way. Today, when someone fills out a survey, their answer lands in a pile of data, and they have no real sense of whether anyone looked at it. 

As AI makes analysis far easier and faster, I expect many more cases where a company reviews the responses coming in, spots that a particular segment has said something interesting or that a point raised deserves more depth, and then goes back to ask about it directly.

Concretely, that means a business could notice that 500 people in a specific sub-segment mentioned a particular issue without enough detail, then run a targeted, interactive follow-up aimed at exactly that group, asking a custom question built around what they already said. 

That turns feedback into something closer to an ongoing dialogue between a company and its customers, rather than a single exchange that ends the moment the form is submitted. It also gives people a reason to keep participating, because they can see that what they said actually led somewhere.

Closing the Loop Between Collection and Analysis

I think this is part of what’s behind Salesforce’s acquisition of Listen Labs, though I want to be careful here since I haven’t seen any public statement from them about it. Looking at the product, I can see the first half of this picture: an interesting approach to interactive collection. 

What I haven’t seen yet is the second half; the interactive follow-up built on a proper analysis that’s already sorted responses into meaningful segments. You need that analysis step to work out which segments are saying what, before a follow-up conversation can be targeted. If Salesforce is building towards that, it isn’t public yet.

It’s also the direction we’re moving in at Caplena. We started as an analytics platform, focused purely on making sense of open-ended text, and we’re now building out the collection side as well. The reasoning is straightforward: a strong survey collection tool and a strong analysis engine can reinforce each other. 

Once collection and analysis sit together, you can drive the kind of interactive, probing feedback loop I’ve described, rather than treating collection and analysis as two separate jobs handled by two separate tools.


What This Means for How Teams Design Feedback Programmes

For research and CX teams, I think the practical implication is to stop designing feedback programmes as single, static instruments that get sent out and then left alone until the next wave. If voice becomes a normal input channel, programmes need to be built to handle more natural, less structured responses from the start. 

And if follow-up conversations with specific segments become standard practice, teams need a way to act on an analysis quickly. To spot a segment worth probing further, design a short follow-up aimed specifically at them and get it out while the issue is still live instead of discovering it three months later in a quarterly report.

None of this means the discipline behind good research disappears. If anything, the need for reliable, reproducible analysis becomes more important, because it’s that analysis that decides which segments get a follow-up conversation and what that conversation should ask. 

But the shape of the programme itself, a single form sent out and then set aside, is very much on its way out. What replaces it looks much more like an ongoing dialogue, carried out mostly in natural language, between an organisation and the people it serves.


Author

Learn more about

Scroll to Top