AI Conversational Research on WhatsApp - The New Qualitative Frontier
Case study: Energy Saver App User Feedback
Working with a pioneering approach on AI conversational research from Langwith Research for the Energy Saver App
The Challenge
For market research practitioners and utility providers, the challenge is always the same: how do you get meaningful qualitative insight at scale without incurring prohibitive costs or timeline delays? The recent user consultation for the Oxfordshire County Council’s Energy Saver App provided a fascinating, and potentially game-changing, answer involving AI conversational research.
Working with Langwith Research, we replaced traditional, low-level human interviews with a fully automated, AI-driven conversational research tool delivered entirely over WhatsApp. The results exceeded our expectations in both speed and depth, providing an extremely encouraging case study for this unique approach to AI-interview research.
AI Conversational Research on WhatsApp: The Breakthrough Methodology
The quality of the AI chats was truly impressive. We completed 129 in-depth interviews in under a week. Crucially, the transcripts read like thoughtful conversations between humans, with the AI successfully employing appropriate follow-up questions and genuine curiosity.
The methodology allowed users complete flexibility: they responded in their own time and could use text or voice notes. This comfort level drove remarkable engagement: we were surprised by the sheer length of some discussions, with several conversations involving over 100 back-and-forth responses.
We believe this approach unlocks a richer vein of consumer insight. Where the goal is to map user behaviour, expectations, and barriers, it’s hugely impressive and incredibly fast. Furthermore, we suspect some respondents felt more comfortable talking to an AI chatbot than a human, leading to more honest and open answers regarding financial worry, home improvements, and their willingness to change habits.
Mark Sutcliffe, Energy Efficiency Project Manager at Oxfordshire County Council, was a keen champion of this innovative approach:
“To improve the Energy Saver App, we needed genuine user insight, not just surface-level data. Langwith Research’s AI interview approach provided us with a robust, scalable way to have hundreds of detailed, high-quality conversations. This is allowing us to build a service that is truly shaped by the residents, ensuring we are on the cutting edge of delivering effective energy efficiency support.”
The Caveat: Human Insight Remains Essential
While the AI research data collection was revolutionary, the analysis stage offered a necessary reminder of the value of human expertise. Just feeding the transcripts into an automated AI analysis tool (as Holly and I tried when we first received them) was less impressive. We found that many subtle insights were missed and some quotes were mistakenly attributed or taken out of context.
Instead, the process still required considerable human input from the team at Langwith Research to code, verify, and synthesise the data – but it provided a much richer, cleaner starting point than a standard questionnaire ever could.
The clear takeaway is that this technology is a powerful data gathering accelerator, not a replacement for a skilled analyst.
Langwith Research: Conversational AI Research at Scale
Langwith Research specialises in bridging the gap between automated data collection and true qualitative depth. They achieve this by deploying conversational AI agents that mimic the curiosity and structure of a skilled human interviewer, delivering rich, verbatim data at the scale and speed of a survey. This allows organisations to move beyond simple questionnaires and capture genuine why behind consumer behaviours and motivations.
Speaking on the methodology’s value, Pablo Garfias Torrent, Co-Founder of Langwith Research, noted:
“The beauty of this system is that it delivers quality without compromise. We captured 129 rich user stories for the Energy Saver App in a fraction of the time and cost of traditional methods. For public sector bodies like Oxfordshire County Council, this means they can make crucial development decisions based on a depth of insight that was previously unattainable within standard project budgets and deadlines.”
Key Learnings
The scale of insight was impressive – here are a few highlights:
- High Energy Anxiety Confirmed: 61% of participants are worried about their energy bill, with 30% expressing a high degree of worry. This reinforces the need for utility and public sector tools focused on reducing financial stress.
- App Driving Behaviour Change: 65% said the app gave them a better understanding of their usage; 43% reported less worry; and 42% confirmed saving money (averaging over £100 per year). The app is demonstrably achieving its core goals.
- Advanced Technology Engagement: Over two thirds were aware of the unique solar + battery offer made available to some users as part of a government trial, highlighting that users are receptive to being offered advanced efficiency solutions.
- Primary User Improvement Requests: The most-requested enhancements included more actionable and personalised tips, more real-time data, and more detailed appliance-level insights.
- Strong Recommendation Index: Over 80% of users said they would recommend the Energy Saver App to a friend or colleague, indicating very high user satisfaction and product market fit.
- Credibility is King: The Council’s involvement was a huge driver of trust, with 79% of participants saying it made them more likely to try the app.
- Appliance-Level Detail: Users are highly motivated by seeing detailed appliance-level breakdowns of their energy consumption (27% requested this).
- Future Integration: The biggest improvement request was for the app to become a unified hub that integrates data from solar, batteries, and EVs (24% requested this).
AI Conversational Research on WhatsApp - Final Thoughts
This project proves that conversational AI, when delivered through skilled practitioners like Langwith Research, offers a high-quality, scalable alternative for qualitative research, fundamentally challenging the economics and timelines of traditional methods.
No, we don’t believe it can replace the more nuanced skills of a human researcher who is able to respond to tone of voice, body language etc, but it has unlocked the ability to understand consumer behaviour at speed and scale.
So while it might not replace the human interviewer in many situations, but in others like this, it definitely has a role to play.
Learn more about Langwith Research at https://langwith-research.com/