I came in not knowing the domain. I left with expertise that opened the next door.
2024–2025 | Freelance, via TheTalkingStick.co + direct with EasySend
1. What I walked into
2. How I got oriented
3. What I contributed
4. What I learned about AI in complex systems
1. Learning the domain: workflow automation and multi-touchpoint systems
I came into this engagement not knowing what workflow automation was. Not the patterns. Not the product logic. Not what it actually means to design for a system where the user is not filling out a form but constructing a process.
EasySend is a no-code platform for enterprise companies that need to build digital customer journeys: insurance claims, loan applications, onboarding flows. Processes where data moves between multiple people, systems, and states before the job is done.
A single journey can involve a customer filling a form, an agent reviewing it, a third party signing a document, and an automated follow-up if the signature does not arrive in time.
That is a different kind of product than anything I had worked in before. I knew it immediately.
The Challenge
Coming into a complex domain without prior context is a real challenge, especially when the work is strategic rather than executional. I could not rely on intuition. I had to learn the domain before I could contribute to it.
The EasySend platform covers a wide surface: form building, document generation, e-signatures, workflow logic, integrations with CRM and ERP systems, and AI features layered across all of it. Understanding how these pieces relate to each other, and how a user moves through them, took deliberate effort.
There was also a framing challenge. I was coming in through a strategic consulting engagement, working with Shahaf Ben-David from TheTalkingStick.co, who led the initiative. My role was to support and execute, not lead. That meant being useful without overstepping, and contributing without pretending to have authority I did not have.
The Process
I spent the first weeks immersed in the product. I signed up for trials of competing platforms. I went through the full first-time user experience on each one, step by step, the way a real new user would. I did not just click around. I built things, deleted them, changed them, and documented every screen.
This is how I learned what a workflow builder actually feels like from the inside: where it creates confidence, where it creates confusion, and what the critical decision points are.
I also paid close attention to how AI features were being introduced across these platforms. Not what AI could theoretically do, but how companies were actually surfacing it to users: as a side panel, as a step in an automation, as a conversational assistant, as a suggestion engine. Each pattern has different implications for UX.
One research session ran nine hours. By the end I was exhausted, but I had a clear map of the competitive landscape and a strong foundation for the strategic work that followed.
What I Understood
Workflow automation is a different kind of product. The user is not filling in a form. They are constructing a process. Every design decision carries different weight because the output is not a completed task but a system that will run without them.
That understanding changed how I approached every subsequent part of this engagement.
The Result
By the end of this phase I could hold a real conversation about how these platforms are architected and where the real UX problems live. That knowledge carried directly into the research work that followed.
EasySend
• Learning a new domain from scratch: workflow automation and multi-touchpoint system design
• Deep competitor research: FTUE analysis, pattern documentation, AI feature mapping
• Contributing to a strategic product experience transformation initiative
• Designing within real constraints: technical, product, and organizational
2. Research: competitor analysis and AI integration patterns
Shahaf ran the company-level strategic analysis. He knew the landscape.
What he needed from me was more specific: go inside the actual platforms, document how they handle the UX of AI integration step by step, and structure it in a way he could use directly for his stakeholder presentation.
The Challenge
Shahaf defined the platforms. Together we shaped the research questions. My job was to go answer them consistently across all of them.
The questions were specific. Not "how does AI work in this product" but things like: when you create a new folder or business group, what does the platform ask you? What does it show, what does it hide, and what does it assume you already know? That level of detail, applied to every platform, across every key area of the product flow.
You cannot answer questions like that from a review site. You have to be inside the product.
The Process
I mapped patterns across products. Where does AI appear in the flow? What
triggers it? What does it offer? What happens when it is wrong?
I also looked beyond the obvious. what does the page highlight, what are my observations, and what is missing. Not just the happy path. Empty states, error states, the moments before a user gives up.
One pattern worth noting was Airtable's empty state approach:
before you type a single word, the system offers examples, templates, and
suggestions. You are never alone with a blank screen. That is not a feature. It is a
position on how much cognitive load the system should carry for the user.
What I Understood
The underlying principle across everything I have learned: AI integration is a UX problem
before it is a technical one. The question is not whether the AI can do something.
It is when the user needs it, where it should appear, and whether showing it there
builds trust or creates noise.
Platforms that got this right made AI feel like a natural step in the flow. Platforms
that got it wrong made it feel like an interruption. The difference was almost never
the AI itself. It was the information architecture around it.
The Result
The output was a structured research file covering multiple platforms in depth. Shahaf used it for his presentation to EasySend's stakeholders.
Because I had done this work myself, I understood the material the way Shahaf did. When he handed me wireframes for the prototype, there was almost no ramp-up time. I could move fast, ask the right questions, and execute with precision. In strategic UX work under time pressure, that kind of shared understanding is not a nice-to-have. It is what makes the collaboration work.
Together, Shahaf and I were able to give EasySend something concrete: a clear picture of how AI could be integrated into their platform in a way that actually serves users, not just checks a feature box. That kind of clarity is hard to reach from inside a product. It takes someone with strategic distance and someone willing to go deep on the ground level. That is what we brought.
3. Prototyping the strategic vision
Shahaf Ben-David (TheTalkingStick.co) led the product experience transformation initiative and produced the wireframes. I built the interactive prototype that brought the vision to life, and contributed research and UX input throughout.
The work explored what EasySend's AI-integrated product experience could look like: AI as a conversational assistant, AI as a step in an automation workflow, AI as a guide that reduces friction for new and experienced users alike.
I was not the author of this vision. I was the person who made it tangible. Shahaf needed something he could use to storytell the transformation direction, something stakeholders could see and interact with, not just read about. That was my job.
Building a prototype from someone else's strategic framework requires a different kind of precision than building from your own brief. You have to fully understand the intent before you can express it in interaction.
Final Takeaways
I came into the EasySend engagement as someone who understood design systems, UX process, and how to work inside complex products. I did not understand workflow automation. I did not understand what it meant to design AI as a participant in a multi-step, multi-person process.
I left understanding it well enough that my next client hired me specifically for that expertise.
That is the part worth noting. Not that I learned something new, which is always happening, but that a deliberate decision to immerse deeply in an unfamiliar domain produced knowledge that transferred directly into new work. The nine-hour research sessions, the competitor walkthroughs, the process of building a prototype from someone else's strategic direction: all of it accumulated into a capability that did not exist before.
When I think about AI integration now, I think about it differently than I did before EasySend. I think about the difference between AI as a tool and AI as a workflow participant. I think about cognitive load: not just what the user has to do, but what the system should carry for them. I think about the moment of first contact, and whether the experience gives users something to hold onto before they have to figure it out themselves.
That thinking is now part of how I work. EasySend is where it started.
Credits: Shahaf Ben-David, TheTalkingStick.co. He led the strategic initiative and the product experience transformation direction. Watching how he runs that kind of engagement, professionally and methodically, was a lesson in itself. Thanks also to the EasySend product team for the trust and collaboration throughout both phases of the work.