From manual to intelligent — an AI proof of concept for sustainability

From manual to intelligent — an AI proof of concept for sustainability

YEAR


2025

From manual to intelligent — an AI proof of concept for sustainability

COMPANY

Anthesis Group

ROLE

UX Lead

Ai Thinking

Facilitation

Vision

Systems

Leadership

INDUSTRY

Sustainability Consultancy

SUMMARY

In just 2.5 weeks, the team designed a 52-screen high-fidelity proof of concept showing how AI could transform the way sustainability clients manage compliance data, generate insights, and collect supplier information. The prototype aligned senior leadership around a shared AI vision and positioned Anthesis as a competitive force in sustainability tech.

device

Desktop

PLATFORM

Web

💡 Overview

To design a proof of concept demonstrating how AI could solve sustainability clients' most pressing data challenges — from compliance tracking to supplier data collection — within a three-week sprint.

Results:

  • 52-screen high-fidelity prototype delivered in 2.5 weeks

  • AI proof of concept aligned senior leadership around a shared product vision

  • Design system expanded to support an immersive AI-powered experience

  • Positioned Anthesis to leverage AI as a competitive advantage for new client acquisition

  1. The Challenge

The problem

Sustainability clients were drowning in data — managing compliance tracking, data submissions, review processes, insight generation, and supplier data collection across complex, largely manual workflows. The opportunity was to show how AI could automate these processes, reduce the burden on teams, and make actionable insight generation fast.

Goal

The problem

Design a credible, high-fidelity proof of concept within three weeks that demonstrated the potential of AI across four critical areas: uploading data, managing data, generating insights, and collecting supplier data.

Sustainability clients were drowning in data — managing compliance tracking, data submissions, review processes, insight generation, and supplier data collection across complex, largely manual workflows. The opportunity was to show how AI could automate these processes, reduce the burden on teams, and make actionable insight generation fast.

Goal

Design a credible, high-fidelity proof of concept within three weeks that demonstrated the potential of AI across four critical areas: uploading data, managing data, generating insights, and collecting supplier data.

My Role

As UX Lead, I led the end-to-end design process from problem definition through to stakeholder presentation, directing a team of five Product Designers. I facilitated discovery and ideation sessions, made strategic scope decisions under significant time pressure, and maintained close collaboration with the Head of Software Development to ensure technical feasibility throughout.

  1. Discovery and definition

Research

With a condesed timeline, we worked with existing research rather than commissioning new studies. I compiled all internal personas into a single document and distributed it across Design, Product, Sales, and Marketing — giving everyone a shared understanding of client needs before the problem definition session.

I then facilitated a cross-functional whiteboard session with stakeholders from Product, Marketing, Sales, and Design to map client pain points and consolidate them into the four problem areas that framed every design decision that followed.

In parallel, the Design Team conducted competitive research into how AI is applied across B2B SaaS products — surfacing interaction models and design patterns that informed our user flows.

  1. Key design decisions

Before design work began, I led the team in mapping key user journeys - making a deliberate scope decision to focus on the flows that would deliver the most value within the POC. When exploring AI-powered data upload, for example, we chose to centre the prototype around a Google Drive integration — a clear, tangible flow that demonstrated the AI experience while keeping scope manageable and technically credible.

I held daily team sessions and one-to-one reviews throughout, iterating on designs against the Design System and managing the team's assumptions about AI functionality openly — keeping the work credible without slowing it down.

To bring the AI concept to life and stretch my Junior Designer, I delegated the AI chatbot experience to her — a component designed to guide users through gathering, reviewing, and generating sustainability insights. It contributed directly to the prototype and gave her a high-visibility piece of work to own.

  1. Delivery and impact

Once the team completed the high-fidelity designs, I brought them together into a 52-screen interactive prototype and presented it to Senior Leadership and the wider Digital Team. The prototype was well received — stakeholders praised both the user experience and the strategic application of AI. The next step was client presentation to gather feedback and inform future iterations.

The project also left a lasting operational asset — the design system was expanded to support the new AI components, ensuring visual consistency and scalability beyond the POC itself.

  1. Learnings

This project proved that high-quality, strategic design is possible at speed — if the process is right. Investing upfront in problem definition and journey scoping meant the team could move fast in delivery without losing direction. The most rewarding outcome was watching the team rise to the challenge and deliver something that genuinely shifted how the business thought about AI's role in the product.

This project proved that high-quality, strategic design is possible at speed — if the process is right. Investing upfront in problem definition and journey scoping meant the team could move fast in delivery without losing direction. The most rewarding outcome was watching the team rise to the challenge and deliver something that genuinely shifted how the business thought about AI's role in the product.