Getting Thresh Power started
Designing long term vision for a diligence tool in renewable energy project acquisition.
Summary
Thresh is an AI due diligence tool to assist buyers of renewable energy projects. I designed an interactive prototype for our product vision that demonstrates Thresh’s core features. The demo helped the founders land crucial early stage sales and pre-seed funding, along with a design partnership that informed a major pivot in our product strategy. Check out my demo designs on Thresh’s website.
An Alternative to a Hurricane of Files
Buyers of renewable energy projects rely on a hurricane of a disorganized files from a seller, confusingly named documents, and endless email chains to make sure their deals go off without a hitch. These are huge deals with millions on the line that right now are messy, tedious, and slow. Thresh’s goal is to speed up the due diligence process for buyers of renewable energy projects by using AI for file organization, risk detection, and task management.
-
The Mergers and Aqcuisition Analyst leads the due diligence process when deciding whether to buy a project. They assessing risk by skimming seller documents, and pull in domain experts to make sure the project’s environmental, legal and other requirements are satisfactory.
Getting a More Complete Picture
To get the context I needed for design the demo, I dove into the context of the problem Thresh was solving by speaking with our founders, doing my own research into renewable energy project acquisition and Thresh’s potential competitors, and reviewing Thresh’s MVP and demo. The initial design the founders had was incomplete—it only had one of three value props designed and built, so I needed to design something that could be shown to customers and investors to say ‘hey this is our vision’. The prototype needed to look like a real application, and built with something simple for engineering to code, so we used the Material UI component library.
-
No one is currently solving the problem that these M&A analysts have. A few minor competitors solved adjacent problems, but they didn't pose direct competition because they either didn't focus on renewable energy, or didn't use AI.
-
I took a look at Thresh's previous MVP and demo designs to find what was working and what wasn't. Their designs were lacking appealing UI and weren't complete in their fulfillment of Thresh's 3 value props.
-
From meetings with design partners, I learned M&A analysts in the field rely heavily upon excel sheets, email and data rooms for their due diligence process. These timelines are tedious and take months.
A core part of designing Thresh’s long term vision was deciding out how file organization would work, Thresh is a data room after all. I looked at different file organization structures like in Dropbox, Google Drive, Notion, and Mac. I settled on a plan for a similar structure to google drive because it’s used a lot by our users and wanted to make it familiar, but included a project list on the sidebar nav, and within the project page, including 3 tabs for the 3 value props: files & folders, risk report, tasks.
The Design Process
Early sketches and low fidelity wireframes
The goal: to decide on a user flow for the demo by testing different approaches file organization with different UI’s, different amounts of user freedom, and user help.
-
How much freedom do we want to give the user to opt in or out of Thresh’s AI organization? How do those decisions impact the user experience and interface?
-
How are we guiding the user through file organization? How do we make the experience as intuitive as possible to minimize Thresh having to explain how to complete actions?
-
The UI options for file organization were many. I looked at different file organization structures like in Dropbox, Google Drive, Notion, and Mac.
I settled on a similar structure to Google Drive because it is used a lot by our users and wanted to make it familiar, but included a project list on the sidebar nav, and within the project page, including 3 tabs for the 3 value props: files & folders, risk report, tasks.
Playing with different UI’s for files and folder organization—an iPad and Procreate can keep me busy for hours…
Created 4 different flows that gave the user different amounts of freedom to opt in or out of Thresh’s AI organization features
The Demo Design
The Project Home Page: Before
The demo design for the home page before I stepped in.
Dull colors lack of delight
Indistinct calls to action create too much cognitive burden
The poor visual hierarchy doesn’t lead the eye through the page
The new project home page in my demo design.
The Project Home Page: After
Harmonious color and contrast balance are more pleasing
Call to action affordances are signaled with design conventions
Improved visual hierarchy leads the eye
A New Feature: Flagging and Commenting in a Document Viewer
In order to do quick diligence on an acquisition, users can view, comment and flag text in an in-app document viewer. Comment threads that can be resolved, or if they stay unresolved, are included in a diligence checklist.
Users can assign and track tasks with colleagues with the support of Thresh’s AI autocompletion of answers and file attachments. Users have the option to create tasks out of comments on documents in the in-app document viewer.
An Improved Task Log
Risk Report: Before
Difficult to read
Unclear UI affordances like numbers and source icons
Disharmony in color, contrast and design
Risk Report: After
Improved readability with a less saturated background color
Clearer affordances by removing numbers and source icons
Harmonized colors
A New Feature: A Due Diligence Checklist
Users use an AI checklist that pulls data from the project documents that the sellers have provided in order to perform due diligence on important details of an acquisition. This checklist comes as a template that can be personalized.
Check out my full demo designs on Thresh’s website…
Next Steps
The Search Experience
Thresh plans to have AI search in their product to answer questions like ‘how does this price compare to previous project prices?’. I will bring my LLM and keyword search UI designs into higher fidelity, and make suggestions on how to implement them.
Feature Integration
One big question we have is how the features of the product integrate with one another. For example, when you flag a section of a document, does it automatically create a task in the task management section for your colleagues to review?
I will continue fleshing out these relationships.
Learnings
Don’t complete an assignment, add value
At Thresh, I learned that my work has the most impact when I focus on the core company goal that the task I am assigned to is addressing. Being a start up, Thresh’s goal was to develop the MVP product and raise money as soon as possible. After spending a while trying to make my prototype pixel perfect in Figma, I shifted my focus to creating a demo as quickly as possible that looks like a real product, and clearly demonstrates the product's value props. When I made this change, I was able to iterate on the design more quickly, and Thresh’s cofounders were able to achieve their goal of quick MVP development and raising money by September.
Don’t hide in a hole
My managers are busy cofounders. So, when I started, I was worried about bothering them with clarifying questions, and would instead try to reason my way into understanding their expectations. This got me nowhere. After pushing myself to ask clarifying questions early on, my work became better and we developed closer working relationships. And as a consequence, everything became even more fun.
