CASE STUDY
Digital Wastewater Management Twin
Transforming wastewater treatment with AI-enabled insights and operator-first design
%
Lower resource consumption
Reduced energy and chemical usage, aligning with cost and sustainability goals.
%
Improved compliance
Fewer non-compliant events, decreasing the need for manual audits.
%
Less Downtime
Predictive alerts shortened equipment outages and boosted productivity.
The project delivers measurable efficiency while empowering frontline staff with trust and clarity.
Design choices that drove results
Explainable AI Insights
Algorithms provided transparent reasoning, not just black-box outputs, building trust with cautious operators.
Shift Handover Tools
Designed communication features that simplified transitions, ensuring continuity across teams.
JTBD-Centered Platform
Used the JTBD framework to uncover not only functional but also emotional and social needs, shaping the solution around confidence and credibility.
PROCESS EVIDENCE
Research & Insight
We grounded the project in operators’ daily realities. Through interviews, on-site observations, and surveys, we identified that inefficiencies weren’t just technical—they were human. Operators struggled with stress, lack of clarity during shift transitions, and skepticism toward AI. JTBD analysis revealed their goals went beyond “running the plant” to seeking reassurance, credibility with peers, and confidence in decision-making.
Design Approach
Personas and journey maps captured these insights and guided ideation. We sketched wireframes, then prototyped dashboards, explainable AI modules, and handover features. Cross-functional workshops with engineers and compliance managers kept technical feasibility in sync with user priorities. Iteration was continuous, with operators validating assumptions and refining flows.
Testing & Iteration
Pilot programs at live plants became proving grounds. Usability testing revealed where dashboards felt cluttered, where AI needed clearer explanations, and how handover tools could better fit workflows. We adjusted layouts for readability, enhanced transparency in AI recommendations, and fine-tuned predictive alerts. The system steadily became both more powerful and more approachable.
Challenges & Resolutions
Algorithms provided transparent reasoning, not just black-box outputs, building trust with cautious operators.
Building Trust in AI
Operators were wary of automated recommendations. We addressed this by pairing every suggestion with plain-language explanations and maintaining manual override options.
Integrating with Legacy Systems
Compatibility with SCADA infrastructure was critical. Close collaboration with IT and engineering teams ensured smooth adoption without disrupting ongoing operations.
PROCESS EVIDENCE
Outcomes at a glance
%
Lower resource consumption
Reduced energy and chemical usage, aligning with cost and sustainability goals.
%
Improved compliance
Fewer non-compliant events, decreasing the need for manual audits.
%
Less Downtime
Predictive alerts shortened equipment outages and boosted productivity.
Qualitative Impact
Operators reported greater confidence in daily decisions.
Teams found shift handovers smoother, reducing stress and improving continuity.
Management highlighted improved trust in data transparency across operations.
Lessons Learned
Centering on Jobs-to-Be-Done reframed the problem. Operators didn’t only want efficiency—they wanted confidence, credibility, and reassurance. Designing for these human motivations ensured adoption and satisfaction.
The project also underscored that explainability is key in industrial AI. Transparent systems are far more likely to be trusted and embraced by users.
Finally, we saw that efficiency alone doesn’t drive adoption. Empowerment, trust, and clarity were just as critical as energy savings or compliance gains.