Asset condition insights: AI powered dashboard
Context
This research included interviews and concept testing to evaluate how users interacted with the proposed condition insight AI feature within an asset management dashboard and assess the added value for users.
This research focused on understanding what asset information users find valuable and actionable and to explore how users access and use asset data, identify key decision-making needs and challenges and their viewpoints about AI.
Problem: Customers struggle to find asset information within Maximo
Within Maximo-an application for asset management, customers on an older version of the application have to navigate between different pages and applications to find asset information. This is time-consuming and difficult to do. They’re overwhelmed by a flood of KPIs, scores, and charts that demand complex configuration and deep analytical effort. Extracting meaningful insights requires cognitively intensive evaluation. This slows down decisions-making and increases the risk of missed issues.
With the proposed new asset dashboard, customers can find all the important asset information in one place including information about asset health and performance as well as important metrics. But first, we needed to test how customers responded to this updated application.
Research goals
I wanted to understand customer needs, experiences and expectations related to finding and managing asset data as well their experience with using AI to gain insights about their assets.
I focused on finding answers to the following questions:
What key asset data or metrics are critical for your decision-making, and how do current tools meet or fall short of these needs?
What types of insights or recommendations about asset health would enable you to take immediate action to improve performance or reliability?
Are you currently using AI in your asset management role? If so, how, and if not, what barriers prevent adoption?
Research process: this involved daily collaboration-through slack, email, Mural & MsTeams
To understand the subject area and background research on this topic:
I led the kickoff call with the team to understand the problem space.
I used the internal research repository to conduct secondary research.
I focused on creating a research plan and the discussion guide.
I conducted participant recruitment using Respondent. I needed to recruit customers & business partners as well as non-Maximo users as the aim was to also focus on attracting new customers. Non-Maximo users were recruited through respondent based on a screener.
I scoped the feasibility and viability of the project within the set timeline.
Interview Procedure:
1-1 interviews with 12 customers & business partners over MS Teams-this was 1 hr long.
1-1 interviews with 5 non-Maximo users recruited through Respondent.
The personas interviewed were Reliability engineers, Asset managers and Data analysts.
The interviews were split into 30 minute interviews to understand how participants managed data within their systems and 30 mins of concept testing to get their feedback on the proposed designs.
Cross-functional Teams involved:
2 UX Designers
1 Product Manager
1 Content Designer
1 UX Researcher (me)
Research timeline
Overall: 12 weeks. As a team we worked in 2 week sprints.
Discovery & Research planning: 2 weeks
participant recruitment & interviews: 8 weeks
Research synthesis & final presentation: 2 weeks
Summary of research interviews
“It’s going to take a lot of trial and error to trust the output of AI...I would need to see accurate reports generated over a long period of time before I feel confident about its output.”
I identified 4 pain-points we needed to address through data analysis and identifying themes from the interviews
13/17 participants expressed the need to validate AI outputs, highlighting concerns about relying on AI-generated information to make critical decisions. This remains a key area of concern.
Balancing technical and financial feasibility is a big area of concern for customers if they want to adopt and integrate AI especially as many customers are using assets and software that may be very old.
Attitudes toward AI are impacting adoption especially among ; we need to support users by demonstrating how it can be gradually integrated into APM practices.
Customers are looking for demos showing how AI integration works with their data and helps reduce costs-whether by minimizing downtime or enabling Technicians to focus on higher-priority tasks
Recommendation
Improve data accessibility and usability to help customers extract more meaningful insights and drive proactive asset decisions; this is a known barrier within the asset performance management space.
Introduce role-based training and onboarding materials to support employees in adopting AI-powered tools, addressing common fears and promoting trust in automation.
Reflections on this project
Regular communication drives alignment: With multiple related products and teams involved, weekly updates on research findings ensured everyone stayed informed and could adjust design decisions based on new insights.
Smooth research handoffs are essential: Since the project spanned multiple phases, creating thorough documentation and context was key to helping other researchers quickly onboard, especially in unfamiliar domains.
Managing scope creep and different opinions on solutioning: conducting workshops with the entire team to provide preliminary research updates and findings and work through different ideas was key to moving the project along. It also helped people feel heard and by fleshing out ideas as a team, we could work out what was relevant and high priority or what could be moved to another phase of research.
