/ Company
/ Company
Novo
Novo
/ Role
/ Role
Staff Product Designer
Staff Product Designer
/ Date
/ Date
2025-Present
2025-Present
Business
Business
Business
Business
Banking,
Banking,
Banking,
Banking,
Rethunk
Rethunk
Rethunk
Rethunk
[1]
[1]
/ About the project
/ About the project
Novo is a banking platform for small businesses. When I joined, the product had scaled by shipping features fast — but without a coherent identity underneath. In just over a year, I built the design infrastructure, defined the product philosophy, and initiated the vision for what Novo should become.
Novo is a banking platform for small businesses. When I joined, the product had scaled by shipping features fast — but without a coherent identity underneath. In just over a year, I built the design infrastructure, defined the product philosophy, and initiated the vision for what Novo should become.
Novo is a banking platform for small businesses. When I joined, the product had scaled by shipping features fast — but without a coherent identity underneath. In just over a year, I built the design infrastructure, defined the product philosophy, and initiated the vision for what Novo should become.




/ Getting started
/ Getting started
My time at Novo started with building the orgs first design system. The absence of one told a bigger story. There was no little infrastructure, broken parity across web and mobile, no product prioritization, and little product direction beyond "ship the next feature."
As I dug in, the deeper problems surfaced. Navigation was disorganized. Related features lived in unrelated places. Users couldn't find core functionality. Support volume was rising. The product did a lot of things but didn't feel like anything.
My time at Novo started with building the orgs first design system. The absence of one told a bigger story. There was no little infrastructure, broken parity across web and mobile, no product prioritization, and little product direction beyond "ship the next feature."
As I dug in, the deeper problems surfaced. Navigation was disorganized. Related features lived in unrelated places. Users couldn't find core functionality. Support volume was rising. The product did a lot of things but didn't feel like anything.

[Part 1]
Dicipline foundations
I decided to define the scope of the system work to go beyond the artifacts and reshape how we worked. This happened in three parallel tracks.
I decided to define the scope of the system work to go beyond the artifacts and reshape how we worked. This happened in three parallel tracks.

/ Systems
/ Systems
Beyond creating components, I worked with a recruited engineer to build out a robust token infrastructure to act as the connective tissue between design, engineering, and AI agents. We would use these layers of primitive, alias, and component tokens as the language for prompting AI coding tools to build out the system.
Beyond creating components, I worked with a recruited engineer to build out a robust token infrastructure to act as the connective tissue between design, engineering, and AI agents. We would use these layers of primitive, alias, and component tokens as the language for prompting AI coding tools to build out the system.

/ AI Adoption
/ AI Adoption
As our product systems advanced, consuming the evolving AI landscape, I wanted to ensure I was bringing everyone with us. So I worked directly with our CEO/CTO to push for broader AI adoption across the product organization. I created the job architecture framework for product design — defining levels, expectations, and growth ladders — while he built the equivalent for engineering. Together, we encouraged every discipline to integrate AI tooling into their work. Every team member now has a skill goal tied to AI adoption in their respective domain.
As our product systems advanced, consuming the evolving AI landscape, I wanted to ensure I was bringing everyone with us. So I worked directly with our CEO/CTO to push for broader AI adoption across the product organization. I created the job architecture framework for product design — defining levels, expectations, and growth ladders — while he built the equivalent for engineering. Together, we encouraged every discipline to integrate AI tooling into their work. Every team member now has a skill goal tied to AI adoption in their respective domain.

/ Product Awareness
/ Product Awareness
As our product systems advanced, consuming the evolving AI landscape, I wanted to ensure I was bringing everyone with us. So To get everyone iterating efficiently and providing thoughtful feedback, we needed everyone to start using our product. So I vibe-coded an internal testing portal — a structured tool that gave every employee a way to use and test the product systematically. It was about building a culture where everyone in the company has firsthand product awareness, and feedback comes from direct experience rather than secondhand reports.
As our product systems advanced, consuming the evolving AI landscape, I wanted to ensure I was bringing everyone with us. So To get everyone iterating efficiently and providing thoughtful feedback, we needed everyone to start using our product. So I vibe-coded an internal testing portal — a structured tool that gave every employee a way to use and test the product systematically. It was about building a culture where everyone in the company has firsthand product awareness, and feedback comes from direct experience rather than secondhand reports.


[Part 2]
Dicipline foundations
/ Listening
/ Listening
While the org was adopting these new working patterns, Novo hired its entire brand and marketing function. They immediately started on a brand redesign, and part of that work was reanalyzing the small business market and our ideal customer profile. I worked directly with the brand team, sharing prior research insights, and performed market research about our industry and product positioning.
I used that research as the inspiration for the three product value types Novo could be for our users:
While the org was adopting these new working patterns, Novo hired its entire brand and marketing function. They immediately started on a brand redesign, and part of that work was reanalyzing the small business market and our ideal customer profile. I worked directly with the brand team, sharing prior research insights, and performed market research about our industry and product positioning.
I used that research as the inspiration for the three product value types Novo could be for our users:

Banking
tool
“Where my money lives”
Storage - Balances - Transfers - Payments

Financial
insights tool
“Where my money lives”
Cashflow - patterns - forecasting - recommendations

Operations
tool
“Where my money lives”
Invoicing - bill pay - payroll - lending
/ Forming
/ Forming
Sharing with our executive team, the banking-first identity of the tool won out for addressing the most primary user needs and leaning into our core revenue drives (float capital, interchange, and credit).
This decision shaped my subsequent product vision: what belongs in the navigation hierarchy, page information structures, and scalable product systems. I honed these ideas into a comprehensive information architecture proposal that covered the entire product ecosystem and shared this vision across product, engineering, design, and leadership to workshop and cultivate shared direction.
Sharing with our executive team, the banking-first identity of the tool won out for addressing the most primary user needs and leaning into our core revenue drives (float capital, interchange, and credit).
This decision shaped my subsequent product vision: what belongs in the navigation hierarchy, page information structures, and scalable product systems. I honed these ideas into a comprehensive information architecture proposal that covered the entire product ecosystem and shared this vision across product, engineering, design, and leadership to workshop and cultivate shared direction.
There are 5 core building blocks to creating effective information architecture:
Organizational systems - hierarchical, sequential, and matrix structures
Labeling systems - categorization, semantics, keywords, and taxonomy
Navigation system - the global, local, and contextual paths one takes to get from one location to another
Findability systems - known item searching and assisted browsing
Information systems - the visual display of information

[Part 3]
Visioning & Delivery
With the identity defined and the architectural framework in place, I then recommended a series of enhancements in packaged workstreams. Each was achievable in the near term, high impact, addressed systemic issues beyond its own scope, and built momentum to continue improving the product. I prototyped each concept and worked across teams to build alignment.
01

Increase speed of task completion
Clear entry points and efficient navigation will result in quicker task completion.
02

Increase product awareness
Prioritized information and obvious labeling results in feature discovery and awareness.
03

Increase customer satisfaction
Confident navigating improves product utility resulting in higher product satisfaction.
04

Decrease customer service tickets
Obvious wayfinding decreases confusion resulting in less need for customer support.
05

Grow core revenue drivers
Prioritized navigation steers users to intended, banking-centric, usage.
06

Better scalable infrastructure
Broader IA philosophy can better ingest new features and products as the business grows.
07

Improved brand reputation
Polished design systems and branding improve user delight and association with the brand.
/ Project 1 – Inbox
/ Project 1 – Inbox
Cleanup came first in this process. The Novo product had a growing attention problem: pop-ups, banners, tags, and alerts were all competing to grab user attention with no hierarchy or coordination. Messages, notifications, and action items were scattered across the experience.
An in-app inbox provides a single surface that unifies all accessory information within the product. Without a dedicated PM or engineering IC, I wrote the PRD and RFC, outlining product journeys to determine the inbox contents and volume (messages, notifications, tasks, system statuses, and alerts), and partnered with marketing to build coordinated lifecycle messaging across email, push, and SMS. I stood up the content management workflow in Customer.io and defined the logic for product and system-related messages, empowering the marketing team to take on future messages.
Cleanup came first in this process. The Novo product had a growing attention problem: pop-ups, banners, tags, and alerts were all competing to grab user attention with no hierarchy or coordination. Messages, notifications, and action items were scattered across the experience.
An in-app inbox provides a single surface that unifies all accessory information within the product. Without a dedicated PM or engineering IC, I wrote the PRD and RFC, outlining product journeys to determine the inbox contents and volume (messages, notifications, tasks, system statuses, and alerts), and partnered with marketing to build coordinated lifecycle messaging across email, push, and SMS. I stood up the content management workflow in Customer.io and defined the logic for product and system-related messages, empowering the marketing team to take on future messages.

/ Project 2 – Money movement
/ Project 2 – Money movement
Next, we needed to deliver on a small business owner's most crucial job for their banking tool, moving money.
This workstream started with a specific observation. We had added a bill pay feature to the send money flow — letting users upload a document and extract information to populate payment fields. But the send money flow itself was already long and cumbersome.
The problems went deeper. Invoicing, adding money, and sending money had distantly separated entry points, despite the mental models in our users' heads. Even within these disjointed flows, there was unnecessary complexity and decisions users were forced to make, including manually adding recipients to each of the 5 payment methods.
I consolidated everything into a single "move money" button in the primary navigation that opens a sub-menu with the various send and add methods, then triggers the appropriate flow. The fragmented entry points were unified into a single interaction, and the flows themselves were consolidated into simple forms. Again, I worked without a dedicated PM or engineering IC — I went to those teams for feedback and alignment, but drove the design, logic, and documentation myself. (maybe add this note to the reflection)
Next, we needed to deliver on a small business owner's most crucial job for their banking tool, moving money.
This workstream started with a specific observation. We had added a bill pay feature to the send money flow — letting users upload a document and extract information to populate payment fields. But the send money flow itself was already long and cumbersome.
The problems went deeper. Invoicing, adding money, and sending money had distantly separated entry points, despite the mental models in our users' heads. Even within these disjointed flows, there was unnecessary complexity and decisions users were forced to make, including manually adding recipients to each of the 5 payment methods.
I consolidated everything into a single "move money" button in the primary navigation that opens a sub-menu with the various send and add methods, then triggers the appropriate flow. The fragmented entry points were unified into a single interaction, and the flows themselves were consolidated into simple forms. Again, I worked without a dedicated PM or engineering IC — I went to those teams for feedback and alignment, but drove the design, logic, and documentation myself. (maybe add this note to the reflection)

/ Project 3 – AI Chat
/ Project 3 – AI Chat
The AI chat was the natural extension of this thinking — a conversational layer that could pull from the same data and support content we'd been restructuring, and surface it wherever users needed it instead of making them hunt through static pages. I designed the prompt architecture, safety constraints, and rollout plan, partnering with engineering to ship a system that could analyze spending data and answer product questions without crossing into financial advice.
The AI chat was the natural extension of this thinking — a conversational layer that could pull from the same data and support content we'd been restructuring, and surface it wherever users needed it instead of making them hunt through static pages. I designed the prompt architecture, safety constraints, and rollout plan, partnering with engineering to ship a system that could analyze spending data and answer product questions without crossing into financial advice.
/ prompt engineering

We're shipping in four phases — internal dogfooding, closed beta (current phase), open beta, then GA — each gated on guardrail violation rates before advancing.
We're shipping in four phases — internal dogfooding, closed beta (current phase), open beta, then GA — each gated on guardrail violation rates before advancing.

/ Project 4 [planned] - Account Centricity
/ Project 4 [planned] - Account Centricity
With the three core enhancements, I've been getting ahead of the next phase: rethinking how we display accounts. Currently, internal checking and credit card accounts are separate from external accounts, which are separate from integrations, which are separate from lending tools like MCA and installment loans. They're all financial accounts, but they live in completely different places.
With the three core enhancements, I've been getting ahead of the next phase: rethinking how we display accounts. Currently, internal checking and credit card accounts are separate from external accounts, which are separate from integrations, which are separate from lending tools like MCA and installment loans. They're all financial accounts, but they live in completely different places.
/ Sign-up before

/ Sign-up After

I've began scoping a project where all accounts — regardless of type — roll up into a unified dashboard where users see the aggregate of their business financial health. Right now, the dashboard tries to show a little of everything and therefore communicates nothing. A banking-centric dashboard surfaces the complete financial picture and lets complementary features, such as cash flow insights and reserves, emerge in context.
I've began scoping a project where all accounts — regardless of type — roll up into a unified dashboard where users see the aggregate of their business financial health. Right now, the dashboard tries to show a little of everything and therefore communicates nothing. A banking-centric dashboard surfaces the complete financial picture and lets complementary features, such as cash flow insights and reserves, emerge in context.

Along with this infrastructural redesign of accounts, page layouts will also see revised philosophy.
Along with this infrastructural redesign of accounts, page layouts will also see revised philosophy.


/ Status
/ Status
From a product standpoint, the design system, testing portal, inbox, and money movement restructuring have shipped. The AI business partner and dashboard restructure are in active development. The brand redesign is underway, and we're leveraging the design system to efficiently transition the product to the new visual language.
From a process standpoint, every team member now has a skill goal to adopt AI tooling in their domain. I'm coaching designers to work directly in VS Code and prompt component updates through code. We're continuing to evaluate all productivity developments as a potential process shift. We're also continuing to coach other disciplines to prototype their ideas so the whole team can evaluate together.
From a product standpoint, the design system, testing portal, inbox, and money movement restructuring have shipped. The AI business partner and dashboard restructure are in active development. The brand redesign is underway, and we're leveraging the design system to efficiently transition the product to the new visual language.
From a process standpoint, every team member now has a skill goal to adopt AI tooling in their domain. I'm coaching designers to work directly in VS Code and prompt component updates through code. We're continuing to evaluate all productivity developments as a potential process shift. We're also continuing to coach other disciplines to prototype their ideas so the whole team can evaluate together.
/ Reflection
/ Reflection
The most important thing I've done at Novo was resist the temptation to redesign screens. The problems were structural — in the architecture, in the process, in the team's relationship to its own product. Building the design system first, then the testing culture, then the IA vision, then the individual workstreams — each layer made the next one possible.
Working without dedicated PMs or engineering ICs on several of these workstreams taught me that the most effective way to drive change is to make the work so tangible that people can't ignore it. Prototypes create momentum in a way that decks don't. When I walked leadership through a working concept of the unified money movement flow, the conversation shifted from "should we do this?" to "when can we ship this?"
The AI work reinforced something I keep coming back to: a conversational interface is only as trustworthy as the architecture behind it. The chatbot concept only works because there's now a structured information architecture, a consistent design system, and a clear product identity underneath. Without that foundation, you're just adding to the chaos.
The most important thing I've done at Novo was resist the temptation to redesign screens. The problems were structural — in the architecture, in the process, in the team's relationship to its own product. Building the design system first, then the testing culture, then the IA vision, then the individual workstreams — each layer made the next one possible.
Working without dedicated PMs or engineering ICs on several of these workstreams taught me that the most effective way to drive change is to make the work so tangible that people can't ignore it. Prototypes create momentum in a way that decks don't. When I walked leadership through a working concept of the unified money movement flow, the conversation shifted from "should we do this?" to "when can we ship this?"
The AI work reinforced something I keep coming back to: a conversational interface is only as trustworthy as the architecture behind it. The chatbot concept only works because there's now a structured information architecture, a consistent design system, and a clear product identity underneath. Without that foundation, you're just adding to the chaos.




























