Company:
Role:
Founding Product Designer
Duration:
9 months (Jan 2024 — Sept 2024)
Pivoted from Moda AI to Momentum, rebranded and rebuilt as an automation platform.
Led design across brand, UX/UI, marketing, and product strategy, from 0 to 1.
Worked cross-functionally with engineers and stakeholders to ship the core B2B SaaS product.
Enable non-technical experts to automate their own complex workflows using AI that works like they do: one step at a time.
Rachel was in charge of compliance for a real estate insurance firm.
Her job was to make sure every transaction followed strict legal requirements.
To manage the volume, her team relied on third-party BPO(business process outsourcing) firms...
Third-party BPO firms came with high labor costs.
Offshore teams worked in different time zones, creating delays and async workflows.
Language barriers and manual work led to frequent errors in the compliance process.
In early conversations with Rachel, we saw a clear opportunity: the work was repetitive, rules-based, and ripe for automation.
The compliance workflow isn’t a simple checklist, it’s a complex, multi-step process full of branching logic. If a document was missing a signature, it needed to be handled differently than one with a completed signature.
We began with a straightforward solution: chronological execution, one step after another.
But we quickly realized that linear logic wasn’t enough — real workflows don’t move in straight lines.
To be useful for Rachel (and future users), the system needed to handle flexible decision-making, just like a person would.
I analyzed how other robotic process automation (RPA) tools approached branching logic and realized the infinite canvas model enabled what we needed in a visually friendly way.
We introduced visual “blocks” to represent each step, and allowed users to connect them with branching logic: “If this, then that.”
To make it accessible to non-technical users, we layered in large language models.
Rachel could write simple instructions like “Check if the signature is on line 35” and the AI would know what to do.
This lowered the barrier even further, letting her describe each step in her own words. Rachel could now build her entire compliance process — without writing a line of code.
Rachel’s firm used legacy software that couldn’t connect with modern tools. No APIs meant no way to automate through traditional methods.
We trained Momentum’s AI to “see” the screen and interact like a human, using keyboard and mouse inputs guided by computer vision.
This allowed the system to click checkboxes, scroll through documents, and verify information — just like Rachel would.
Once Rachel built her workflow, she could run it over and over again, no matter how many documents she had.
Momentum could spin up multiple virtual instances to process them in parallel, 24/7.
With AI handling the manual work, Rachel’s day-to-day looked completely different.
She was no longer chasing tasks. She had time, clarity, and control.
With automations running 24/7, Rachel no longer had to chase tasks manually, giving her space to focus on more strategic work.
Improved accuracy meant fewer follow-ups, less double-checking, and a smoother compliance process for her whole team.
By replacing expensive third-party services, Rachel’s team reduced costs and kept ownership of their own workflows.
Momentum wasn’t just solving one problem, it pointed to a broader opportunity.
Most AI tools can handle single tasks. We built a way to chain small, smart tasks into full workflows, giving non-technical users the power to automate complex operations.
With Rachel as the expert in her workflow, and Momentum as the tool, she could now own and scale her process, without engineering help.
Alongside the product, I designed the Momentum AI website to reflect our new brand and clearly communicate the product’s capabilities.
We needed a lightweight, flexible site that could evolve quickly as the product and priorities shifted, from early demos to go-to-market positioning.
The visuals followed the same principle: lightweight, dynamic graphics that supported the content without overwhelming it.
It became our front door for demos, outreach, and recruitment, tying together product, marketing, and brand into one cohesive experience.
This project taught me how to design AI tools for real people, starting with their language and mental models, not just the technology.
My learnings went beyond design. Through working cross-functionally with engineers and stakeholders, scoping features, managing timelines, and shaping the product from 0 to 1, I ultimately learned how to own a product management project.
As the sole designer, I owned everything from product UX to marketing and brand and emerged with a deeper understanding of end-to-end product thinking.
If I had more time and the technology allowed for it, I’d explore simplifying the interface even further.
Right now, we have specialized blocks for actions like “click” or “verify,” but with better LLM performance, natural language alone could drive the logic.
The ultimate goal? A user writes: “Run compliance on property 123 Main Street,” and the system handles everything. This is the holy grail of AI: handling complex, multi-step workflows — the way humans do.
Momentum AI was a bridge toward that vision, turning expert knowledge into repeatable automation, enabling people like Rachel to turn their process into something scalable, flexible, and truly intelligent.
