Communicating AI impact

Overview

Over six months as an embedded UX designer, I helped Genesys make the value of its AI-powered Predictive Routing feature easier for customers to understand. After the model had been running in production, clients needed a clear way to see whether predicted benefits were actually showing up in their KPIs, without exposing sensitive model logic. The work focused on turning complex performance data into a transparent, actionable story that reinforced trust in the AI over time.

Organisation

Genesys

My responsibilities

UX/UI design, Research & usability testing, User flows, Prototyping, Insight synthesis

Challenge

Predictive Routing uses machine learning to match customer interactions (calls, emails, chats) to the best-suited agents based on historical performance. After 30 days live, customers needed an intuitive way to understand: is this working, and where is it making a difference? The challenge was to visualise AI-driven value clearly, without revealing the proprietary model behind it.

Research playback insights 1
Research playback insights 1

Approach

Understanding and framing → Reviewed requirements and documentation, mapped workflows, and synthesised past research to pinpoint where customers struggled to interpret AI decisions and outcomes. Storytelling through flows and visuals → Built user flows and clear narratives that translated model behaviour into business insight, using plain-language feature names (not technical IDs) to match how customers think and talk about the feature. Co-design and alignment → Partnered closely with Product, Engineering, and Data Science, iterating from low-to-high fidelity prototypes to build shared understanding and keep decisions moving, including how to represent ‘Quality’ and other non-traditional KPIs. Testing and validation → Co-facilitated usability sessions to confirm the experience was clear and intuitive, and to validate how results should be shared. This led to exportable views for stakeholders.

Approach

Understanding and framing → Reviewed requirements and documentation, mapped workflows, and synthesised past research to pinpoint where customers struggled to interpret AI decisions and outcomes. Storytelling through flows and visuals → Built user flows and clear narratives that translated model behaviour into business insight, using plain-language feature names (not technical IDs) to match how customers think and talk about the feature. Co-design and alignment → Partnered closely with Product, Engineering, and Data Science, iterating from low-to-high fidelity prototypes to build shared understanding and keep decisions moving, including how to represent ‘Quality’ and other non-traditional KPIs. Testing and validation → Co-facilitated usability sessions to confirm the experience was clear and intuitive, and to validate how results should be shared. This led to exportable views for stakeholders.

Logica UI mock-ups
Logica UI mock-ups

Impact + Outcomes

Approach

The work delivered an analytics experience that made AI results easier to read, trust, and act on after launch. It helped customers track the sustained business impact of Predictive Routing, and it fed practical insights back into the product roadmap through prioritised recommendations. The project also established a repeatable way of testing and communicating AI-driven value across future products. A key learning for me was that customers don’t need to see how the model works but they do need a clear, human explanation of why it matters and what’s changing over time.

Understanding and framing → Reviewed requirements and documentation, mapped workflows, and synthesised past research to pinpoint where customers struggled to interpret AI decisions and outcomes. Storytelling through flows and visuals → Built user flows and clear narratives that translated model behaviour into business insight, using plain-language feature names (not technical IDs) to match how customers think and talk about the feature. Co-design and alignment → Partnered closely with Product, Engineering, and Data Science, iterating from low-to-high fidelity prototypes to build shared understanding and keep decisions moving, including how to represent ‘Quality’ and other non-traditional KPIs. Testing and validation → Co-facilitated usability sessions to confirm the experience was clear and intuitive, and to validate how results should be shared. This led to exportable views for stakeholders.

Impact

The work delivered an analytics experience that made AI results easier to read, trust, and act on after launch. It helped customers track the sustained business impact of Predictive Routing, and it fed practical insights back into the product roadmap through prioritised recommendations. The project also established a repeatable way of testing and communicating AI-driven value across future products. A key learning for me was that customers don’t need to see how the model works but they do need a clear, human explanation of why it matters and what’s changing over time.

Previous — AI Eligibility Checker

Previous — AI Eligibility Checker

Previous — AI Eligibility Checker

Previous — AI Eligibility Checker

Previous — AI Eligibility Checker

Previous — AI Eligibility Checker

Enter password to view work