At Recollective, we’re always making small product changes that can make a big improvement to the ease of use for participants and ultimately, the quality of the insight you collect.
Specifically, this month we’ve spent time on Conversational AI tasks and today we’re launching two upgrades: Talk Mode and Edit Last Post. Together, they make it possible for participants to respond with natural speech instead of being forced to type responses and enable them to correct mistakes in recent responses, keeping the conversation accurate from start to finish.
What are Conversation Tasks
Conversation is a new Task type in which the Recollective AI moderates a one-on-one interview with selected participants. The AI moderator, guided by your research objectives, asks questions and follow-ups based on what each person says, enabling you to collect deep, qualitative feedback at any scale.
What’s new
Talk Mode (voice input)
Participants can now choose to respond using voice input in addition to the existing Type mode.
How Talk Mode works:
- Participants choose Talk or Type when they first enter a Conversation Task
- They can switch between Talk and Type at any time during the session
- Talk Mode is designed to reduce friction on mobile and encourage verbose, natural responses
Edit Last Post
Mistakes sometimes happen. Whether that’s because a typed response didn’t really express what the person meant to say, or a transcribed audio response requires an edit for clarity, Recollective now makes it possible for Participants to edit their most recent message before continuing the conversation.
How Edit Last Post works:
- After posting a message in a Conversation, participants can edit it before moving forward with the interview
- When a participant edits a message, the AI moderator automatically updates its next reply to reflect the change
- This keeps the conversation context accurate and the probing questions relevant
Why it matters for qual researchers
Qualitative research depends on understanding nuance: the exact language people use, the emotion behind a reaction and the context that explains the “why.”
But two common issues can get in the way:
- Typing on mobile leads to short responses that miss the details
- Typos and dictated mis-hears create noise that slows down analysis
Talk Mode and Edit Last Post are built to remove those friction points, so you get clearer data you can trust.
The benefits you’ll actually feel
Richer responses, especially on mobile
Talk Mode removes the typing barrier. That means more top-of-mind detail, more storytelling and fewer one-line answers.
This is especially useful for:
- concept and message testing
- brand perception and positioning work
- customer journey and experience feedback
- moments-based research where participants are on the go
Higher integrity data with less clean-up
Edit Last Post lets participants correct typos, clarify phrasing or fix mis-transcripted words immediately.
That means:
- fewer confusing responses to interpret
- less time cleaning before analysis
- higher confidence in what you’re coding and reporting
AI follow-ups that stay accurate
Standard bots can get stuck when someone changes their mind or corrects a previous detail. Recollective’s AI moderator adapts in those instances. So, when a participant edits their last post, the next reply they see will immediately update to match the new context. Your conversation remains coherent and importantly, properly reflects the participant’s thoughts.
A quick example
A participant dictates: “This feels premium but confusing.”
Then they realise what they meant was: “This feels premium but the benefit is confusing” so they edit their last post. With this latest release, the AI moderator will update its next reply based on the edit and probe benefit clarity, not general confusion. That keeps the conversation accurate and the insight more actionable.
FAQ
Want to see it live?
Get in touch and our team will gladly help you explore where Conversation AI with Talk Mode and Edit Last Post can make a big impact on your research, whether that’s concept testing, message clarity, customer journey work or anything else!
.svg)