Featured Solution - Watti

We all want to reduce our energy bills and carbon footprint. But getting the right advice is difficult.

Generic advice from flyers and websites often doesn’t apply to our specific homes. Expert, in-person advice is highly effective but expensive and impossible to deliver at scale.

This creates a major barrier, especially for those who face language or technology barriers. How can we provide personalised, expert advice to everyone who needs it, regardless of where they live or what language they speak?

This was the challenge we set out to solve.

The Problem: Scaling Personalised Energy Advice

 Delivering effective energy advice faces a fundamental conflict: scalability versus personalisation.

  • Generalised advice (like websites) can reach a large audience but is not specific to an individual’s circumstances.
  • Personalised advice (like in-person consultations) is highly effective but cannot be delivered at scale.

 

This gap leaves millions of people unable to access the specific information they need to make real changes.

The Solution: An Accessible, AI-Powered Assistant

Watti is our solution: an AI-powered digital energy assistant designed to provide the personalisation of an expert at the scale of the internet.

It uses a technology called Retrieval-Augmented Generation (RAG), which means it doesn’t just “invent” answers. Instead, it draws its information from a curated and trusted library of guidance. This allows Watti to provide accurate, specific advice and cite its sources, building user trust.

To overcome accessibility barriers, Watti:

  • Understands and responds in 243 different languages.
  • Interacts via both text and voice (Speech-To-Text and Text-To-Speech).
  • Provides clear links to external sources for further reading.

 

Watti can provide general guidance, run energy models based on a user’s home, and direct users to trusted local suppliers, all while keeping user data private and secure.

The Evidence: How Real Users Shaped Watti

 A solution is only as good as its real-world performance. We trialled Watti extensively, gathering 52 responses via in-app feedback, online questionnaires, and paper forms. This feedback provided a clear, actionable path for refinement.

Key Positive Findings:

  • Accessibility and Ease of Use: Users frequently described the tool as “helpful,” “easy to use,” and “a great, simple way to find assistance.”
  • Trust and Transparency: The inclusion of links and cited sources was highly valued, with one user noting, “Found the internet links helpful.”
  • Useful Tools: The EPC (Energy Performance Certificate) checker was highlighted as “interesting” and helped at least one user discover their eligibility for support.

Actionable Insights and Refinements:

The trial was crucial for identifying specific, user-focused improvements.

  1. Performance and UX:
    • The Issue: Several users noted a “slight delay” on the first use.
    • The Action: This feedback was prioritised, and the initial response time was improved.
  2. Interface and Accessibility:
    • The Issue: Users provided specific UI suggestions (e.g., changing button colours) and critical accessibility feedback, including requests for screen reader compatibility and an easy-read mode.
    • The Action: These UX and accessibility enhancements were prioritised for the next development sprint.
  3. Conversational Flow:
    • The Issue: Users wanted a more guided, conversational experience, such as “follow-up prompts” (e.g., “Would you like to know more about…?”) and “suggestions for related questions.”
    • The Action: Improvements to the conversational flow became a key priority to make the interaction feel more like an assistant.
  4. Content Gaps:
    • The Issue: Users requested more specific and localised information, such as links to actual energy tariffs and advice for renters.
    • The Action: This feedback informed our content strategy, leading to the library being updated with new, trusted documents (e.g., from Which?).

What's Next for Watti?

This feedback loop is central to Watti’s development. Our next steps are focused directly on what users asked for: continuing to improve the conversational flow, expanding our content library, and enhancing accessibility features.  

 

Watti is already proving to be a viable and effective solution for delivering personalised energy advice at scale, and we hope to deliver an open version of the platform in the near future.