DeepSolar by Raycatch

Raycatch is a start-up in the renewable energy industry which has developed a unique AI-based technology for optimising solar plants.

Raycatch offers DeepSolar as a Saas solution for plant owners and technical teams to manage their solar plants in a financial and operative manner.

Project Overview

Problem

DeepSolar is a data-intensive web application and our users feel lost within all the information and insights presented on the platform. Users struggle to process the data and properly manage their issues.

Solution

Finding the core of DeepSolar and leading our users through the system. Focusing on “Plant Insights” as our main suggested solution while providing a step-by-step action plan to fix the plant’s problems.

My Role

Lead and solo designer - discovery, user research, requirements, design

Team

Noa - Head of product
Emad - FE Developer

How DeepSolar Works

DeepSolar is an AI-based Saas platform that collects and analyses data from the monitoring systems located at the solar plants. Here is a quick overview of how it processes data:

Discovery

User Feedback

When I started working at Raycatch, DeepSolar had just gone through an intensive redesign process and was a new and improved platform. It has been used for a couple of months by multiple customers on pilot phases. During our meetings with clients we kept receiving this feedback:

“DeepSolar works and looks great, the data is accurate and there are many insights to work with, we just don’t know how to do it.”
Martin, BlackRock

We realised there is unfulfilled potential for DeepSolar, and while the technology is accurate and impressive, users felt lost when they had to use it without our guidance.

Brainstorming Session

In order to solve this challenge, we had a brainstorming session that included Noa (Head of Product), Lenna (Former Designer who created the MVP) and myself. We focused on understanding the main difficulties of our users and finding the core of what they want to know and do. We also defined the preferable user-flow we want users to take in order to accomplish their goals.

Persona

We discussed the different persona, the level of information each of them wants and can process and their goals. For example, an Owner would be interested in a high-level overview of the plant’s insights, while an Operator would be interested in drilling down to find specific devices which have problems.

User Goals

• Reviewing all the problems across the plant
• Detecting low performing devices
• Understanding the steps for solving each problem
• Targeting “quick wins” - faulty devices with high ROI

Brainstorming Conclusions

• “Plant Insights” is the most valuable section on DS
• The insights should be more practical
• The flow should allow a quick and easy detection of problematic devices

Design Process

Recreating the Sections Hierarchy

“Plant Insights” section lists all the problems, which we refer to as “insights”, that were detected across the solar plant. Each insight also offers a step-by-step personalised recommendation to solve the problem. The current design does not do justice to the Insights section, it can easily be skipped on and the recommendations are not prominent enough.

Based on the realisation that "Plant Insights" is the most valuable section for most users, there is a need for a designated page for it, so users can explore it conveniently while we emphasise the important points for each insight.

1. Current state - Plant Insights is on the right, to see all insights users have to scroll down
2. Adding a designated page for the insights
3. The page will contain all insights and will allow a convenient view

Leading Our Users Confidently - User Flow

So far we did not a set a preferred user flow and we noticed users were wandering around aimlessly. Once we focused on our strengths we defined 2 main user flows:

  • Direct access from the Plant Overview page to the Device Drilldown page to investigate a specific insight.

  • A route via Plant Insights page which allows users to explore all the problems on the plant.

Presenting a New Concept - ROI Based Action Plan

Following the feedback we received that the insights were not “practical” enough, we considered a new approach. We suggested dividing all insights to 3 levels according to the urgency of the problem:

1. Immediate
Corrective actions with high gains that can be carried out immediately.
2. Mid-level
Corrective actions to be considered on the next scheduled visit.
3. Long-term
Potential asset improvements that may require long-term processes.

After sharing a mock-up with some leading clients and collecting feedback, we knew we were on the right path. They embraced this concept and linked it to practical actions for the maintenance team.

“This 3-priorities division helps us to understand where we need to put our efforts and prioritise the tasks according to urgency-level.”
Sara, Fortum

Final Design

Designing a New Plant Insights Page

Final Design
The first block presents the general state of the plant: total recoverable energy and the breakdown according to the Action Plan priorities. Following the first block are the insights, divided to 3 groups: Action Plan, additional insights and healthy insights.

Insight Division
The first group of insights contained all the actionable insights - problems that could be fixed by clients and therefore were the most valuable.
The second group I created was for insights that were Faulty but not actionable and therefore were not a part of the Action Plan. The last group contained all Healthy insights - insights that performed well and did not need any care on the client's side.

Sketches
When I approached designing a new page from scratch, I first needed to map out all the options and levels of information we wanted to show. I wanted to maintain all the information that appeared on the cards and add the priority level we set for each insight.
Here are some of the sketches:

Color Choices
From the initial state of planning this feature, we visualized the colors as red-orange-yellow to indicate the level of severity of the problem. We needed to conduct some color testing to make sure our minds are not set on these colors because we used them to express an idea when collecting feedback from clients.

Plant Overview - Widget Redesign

The Plant Overview page contains snippets of all the features on DeepSolar and we needed to give the Plant Insights widget a redesign.
I took 3 key decisions for the Action Plan widget design:

• Showing 3 “highlight” insights
• Maintaining the cards format
• Maintaining the existing grid

Grid Exploration
The decision to maintain the existing grid was mainly due to developers resources and with that in mind, I needed to find the best solution for the layout. After some sketching and testing, the final decision was to adjust all the widget to a mirrored grid.

Widget Evolution

Final Design

Final Design - Both Screens

Conclusions

I loved working on this feature since it required an innovative approach and courage to make profound changes in DeepSolar. In order to solve this problem, we needed to take a wider view of the platform and have a deep understanding of our users. The process also involved a high level of research and iterations which helped promote this feature in the right direction.

Next Steps

To make this feature even better, we decided on the next steps for upcoming versions:

• Adding a task manager
• Apply the Action Plan approach cross-system (portfolio, histogram & heat map)
• Allowing users to personalise the settings of the priorities
This is it for now.
There is much more to see from my work at Raycatch. Please contact me via morangibson@gmail.com or Linkedin to learn more.

More projects