Overview
First stage development of an AI product.
Pre-seed to seed stage start-up aiming to create a B2B AI powered system designed to help accountants and businesses manage their financials seamlessly under one roof.
Key Contributions
(1) Oversight and curation of a scalable brand to support fundraising efforts now and market sales later.
(2) Research to uncover our highest risk assumptions and validate them.
(3) Working with stakeholders to create rapid lean designs for MVP.
(4) Scout, hire & manage external design partners
Role
Head of Design
Tools
Figma, Framer, Slack
Company and Year
Alexion.ai | 2025
The Problem
We quickly needed a Brand that was versatile and could appeal to distinct and the widely different audiences of investors and our early adopter end users.
Our leadership team needed help to identify some blind spots and help to mitigate their high risk assumptions. Specifically, we had some misalignment on who our target MVP user was and some key assumptions about what their biggest problem was to solve.
The Dev team needed user experience deign thinking input as they sprinted to create a rapid workable prototype.
Opportunities
Create a Brand that had foresight for scalability and could adapt as we grew.
Gather insight into our assumptions by talking to our beta users to help steer us towards success.
To ensure our rapid product building didn't sacrifice usability for speed of delivery
The Process
Discovery & Stakeholder Alignment


Led a cross-functional assumptions matrix workshop with product, engineering, and leadership teams to identify areas of highest risk.
We prioritized highest risk assumptions related to:
How much trust users were willing to place in an AI-driven system.
Expectations around automation vs. control.
Delivery of MVP as a Slack integration.
Consideration about our focus on solving the problem of onboarding new clients for accountants and firms is the largest and most fruitful area of opportunity in the given problem space.

Early Design Exploration
Designing an AI interface with a few guiding principals in mind:
Mimicking familiar tools provides a comfortable and intuitive experience.
Minimizing the learning curve with pre built examples and low friction onboarding.
Minimal design with strong hierarchy, clear visual feedback and total user control.
Using low/ medium fidelity UX wireframes to explore with leadership the possibility of allowing users to choose from three different levels of AI interface; as assistant, as invisible software, and as friendly personality.

Working with external vendors, I guided the development of a working set of brand guidelines and subsequent application of that brand for our investor pitch deck.
Next the dev team built a working minimal product to release to our early beta testers.
I planned and conduct research, one-on-one interviews and competitive analysis to verify our assumptions and guide the direction of our post MVP improvements.


Lydia Deerheart 2026