Designing an AI tool to help investors get deals done faster.
Project Brief
Playbook is an early-stage AI startup revolutionizing company evaluation and deal execution processes for investors. While there was early user interest in Playbook’s AI-backed tool, they had yet to fully validate their hypothesis with paying users.
My Contributions
✔️ Conducted user research to validate their problem statement
✔️ Helped shaped their product vision and value proposition to set them apart
from competitors
✔️ Redesigned their website app for a more modern look in line with their brand
and for a more intuitive experience
✔️ Crafted compelling copy for email marketing campaigns; designed marketing
website
Collaboration with the team
I worked closely with the product manager and engineer with an agile design approach. We met several times a week to analyze user feedback and ensure design decisions were balanced with business objectives and technical constraints, driving iterative improvement.
Before
After
CONTEXT
Closing deals is hard. M&A professionals spend 80+ hours per week on manual tasks.
The founder of Playbook, with 10 years of prior experience in investment banking and private equity, was frustrated by how tedious and manual a lot of the work was. He envisioned a game-changing solution:
What if we create an AI-powered, all-in-one resource to streamline company research and deal evaluations for investors?
Project goals
✦ Identify target users and validate the current problem statement and solution
✦ Revamp visual design to reflect a sleek and modern brand image
✦ Enhance usability and scannability to deliver value and foster user
engagement and retention
Despite Playbook's large potential customer base, each user group had unique needs and frustrations. I sourced users from various groups with differing seniority levels to interview and understand the following:
✦ Which user group should we target, and what are their specific needs?
✦ Does our current solution solve a real problem?
✦ Is our value proposition compelling enough to drive user engagement?
INSIGHT #1
75% of users, primarily PE investors, expressed a strong preference towards using existing tools.
Early interviews revealed a major obstacle: current tools are deeply engrained into investors' workflows. Despite their inefficiencies, investors are reluctant to switch over to new tools.
"This might have a bit more than Pitchbook or Crunchbase. Not sure how much I'd pay given I have information on a lot of these already"
One promising insight was that VC investors tended to be more open than others to trying new tools. However, Playbook still needed a unique and compelling value proposition to set them apart from competitors.
INSIGHT #2
Users were intrigued by AI, but 62% had concerns about privacy.
While investors expressed strong interest in the AI document chat, testing showed that they were hesitant to use it due to privacy concerns. As privacy was very important in the deal world, working with AI was a significant challenge. AI was also prone to inconsistencies and inaccuracies in its outputs.
"I'm not sure if I'd feel comfortable uploading confidential docs. How do we ensure our information stays confidential?"
INSIGHT #3
67% of users thought the website looked clunky and hard to scan.
Investors also thought the website resembled that of Pitchbook, which was our biggest competitor. Playbook needed to stand out from its competitors somehow.
"This is a lot of information to take in. It would be nice if it was displayed in blocks, and not a huge list."
Identifying problem spaces and opportunity areas to create value
From research findings, it seemed that Playbook's current value proposition wasn't strong enough. I created a customer journey map to help the team zoom out and identify other areas of opportunity to explore within the deal process.
Once I identified problems that needed to be solved in each step of the deal process, I created a prioritization matrix to identify and advise the client on which opportunities to explore first.
EXPLORATION #1
How might we use design to address privacy concerns with AI tools?
Hypothesis: Users need a workflow that would allow them to upload private docs and work in a private instance, where they could utilize AI comfortably.
Solution: I decided to use the existing design pattern of Incognito, or dark mode, to indicate to the user that they are working in a private workspace.
Result: The introduction of the private workspace mode increased the usage of Document AI by 8%.
However, we found that users were still hesitant about using an AI chat that stored data in the cloud, and the increase in usage wasn't enough to justify spending a lot more time and money on the feature. At the same time, other apps like ChatGPT and CoPilot had developed powerful document reader tools.
We decided to drop the AI document feature and are experimenting with using AI in other areas. Nonetheless, I think the light mode/dark mode design idea to indicate a private workspace has great potential for future usage.
EXPLORATION #2
How might we create a strong value proposition and distinguish ourselves from competitors?
"The most tedious part of my job is creating competitor matrices - I need to know all the different competitors, what series they are, their growth, founders, revenues, HQs, etc. It involves a lot of Googling and manually filling out, usually taking about 2-3 hours."
Hearing the same insight across three user interviews, I advocated to my team to explore the problem of creating competitor maps. Through a productive brainstorming session with the product team, we connected the dots.
Playbook to the moon! 🚀
We already had comprehensive product and financial details on every company scraped by AI, as well as accurate competitor lists. Why don’t we take advantage of that to:
1. Give users a comprehensive competitor view
2. Use AI to find key differentiation between the products?
First iteration
Final iteration
Labeled how the data was generated
Added customization tools like bucketing, searching, sorting, and adding columns
Removed unnecessary table lines for less visual noise
Added zebra stripes for easy distinction between rows
Much more scannable and powerful!
"Having all of this information on one screen saves me so much time. It's helpful to quickly know how my product differentiates from competitors... But how do I know how accurate the AI information is?
So far, users have expressed positive feedback in both improved scannability and quality of content, but we are continuing to adjust AI prompts to get the most accurate and useful information. As a next step, I am exploring other options to display the similarity score in a more meaningful and user-friendly way.
EXPLORATION #3
Users don't want to leave Excel.. how might we integrate our tool more seamlessly into Excel?
Insight: Users spend a lot of time modeling in Excel.
Hypothesis: Users need a tool that could help automate their waterfall models in seconds, saving hours of time.
This flow was unclear – users weren't sure where to start and which inputs were required.
Modeling tool: first iteration
Some tables required many rows, which led to excessive
horizontal scrolling.
We thought this button would allow for easy integration into excel, but users weren't clicking this button, as they dropping off before this point in the flow.
I decided to simplify the complex user flow by transitioning to a step-flow approach, segmenting it into with smaller, more manageable steps. The use of progressive disclosure should help reduce cognitive load.
Second iteration
However, this multi-step flow introduced more challenges. We found that:
✦ Users began abandoning the long process halfway through
✦ Users still preferred to use Excel
Rather than moving away from Excel completely, users want to seamlessly integrate new tools with Excel to make their current processes more efficient.
Following this observation, I proposed the idea of creating an Excel plug-in that would further streamline investors' workflows.
This turned out to be a significant breakthrough in overcoming the barrier to tool adoption.
Final iteration
_gif.gif)
Despite some UI constraints with the plug-in format, we ensured the UX mirrored the same web app flow, facilitating easy data transfer to Excel.
The plug-in has received overwhelmingly positive feedback, adopting 8 paid users so far! Users appreciate that the tool saves them time on current processes without needing to learn a new tool.
"This is super helpful for sense checking my work and for when I need to know distributions really quickly. I do think the future is about having more automated spreadsheets - it doesn't make sense how manual a lot of this is!"
We are actively continuing to test and refine the tool, making further flow improvements to enhance usability. Next steps involve further user acquisition and expanding the scope of our Excel plug-in to incorporate AI technology.
🎨 UI Design
UI EXPLORATION #1
How might we establish a sleek and modern brand that sets us apart from competitors?
I initially used gradient backgrounds to achieve a glassmorphism effect for a modern look that complemented its AI features.
"Cool UI, but slightly distracting."
I decided to opt for a simpler UI leveraging mostly grayscale to highlight content, keeping the gradient and pops of colors in only certain areas like the data visualizations, logos, AI features, and buttons.
First iteration
Final iteration
A grayscale color palette with gradient accents (right) is much more practical for a data and content-heavy platform.
UI EXPLORATION #2
How might we elevate our brand with typography?
I started with a neutral (and slightly boring!) Inter font but switched to IBM Plex Sans for a techy body font and Helvetica Neue for modern titles. Additionally, I reduced the number of font sizes for simplicity and easier implementation.
Before
After
Design System
I initially chose the AWS Amplify UI Builder for its rapid prototyping and development capabilities. However, I found that its components were rather limited. Next time, I would use a more extensive UI kit with more dashboard design components for even faster design.
UI EXPLORATION #3
How might we redesign our logo to further establish our brand identity?
The original logo was created by mashing two font styles with two different cap heights together. For the redesign, I utilized the pen tool in Figma to create a vector icon with a “P” inside of a book to represent Playbook (clever, right?).
I also used a gradient as its unique brand color to represent modern technology and differentiate it from competitors with an older UI.
📈 Results
Impact and next steps
So far, the redesign efforts have driven a 25% increase in user engagement, and have helped with acquiring 30 new users! 🎉 In addition, we've received positive qualitative feedback, with users expressing strong interest in the Excel plug-in and redesigned web app, particularly with the competitor mapping tool.
This project is ongoing, and Playbook is continuing to do rapid iteration and testing with users to find their product market fit.
Lessons learned
✦ Stay flexible: maintaining flexibility to rapidly discard ideas and pivot when the product doesn't resonate with users is critical in a product's early stages.
✦ Prioritization is key: an a fast-paced startup environment, a designer should work closely with the PM to determine what is worth solving, and what to solve first.
✦ Engage with early users: finding product market fit is hard. Focusing on solving a specific user groups' needs first will help with gaining early adopters and identifying the product's value proposition before scaling.