A reviewed version of the final steps before creating a model. Decided to give more visibility over what’s going on under the hood with meaningful steps instead of a list of AI jargon. On the right, the final step of the new tool walkthrough is introduced.

Main Challenges

Domain immersion & taxonomy standardization

Before writing a single word, I conducted deep systemic audits of dozens of user journeys within the product.

I created a centralized jargon-to-human taxonomy that became the source of truth for both Engineering and Design teams, bridging the mental-model gap between expert data scientists and everyday low-code developers.

Product-led growth and time-to-value

When I joined, there was significant friction in the tool already. We knew users were abandoning the process during a mandatory configuration phase.

I reframed the experience from a mandatory setup burden to a progressive, value-first onboarding flow focused on expectation management and reducing cognitive friction.

Systemic governance and scalable patterns

Keeping the experience consistent within the product and across the broader company ecosystem required abstracting one-off screens into a comprehensive content governance framework.

Different versions of a mandatory initial configuration. On the left is how it was before I joined the team, on the right the reworked version.

Optimizing the first-time experience

I redesigned the initial configuration, a critical friction point, into a welcoming PLG initiative to reduce abandonment rates.

• Dropped the descriptive title for an invitation, welcoming the user.
• Focused the copy on time-to-value and benefits rather than system burdens.
• Standardized the taxonomy, using the action set up universally instead of mixed verbs like configuring or connecting to.
• Implemented progressive disclosure to manage expectations and cognitive load.

One modal to rule them all.

Scalable governance

To guarantee consistency, I created a scalable pattern system for all confirmation and cancelation modals. Titles confirmed actions, buttons were action-based, and destructive actions had strict, standardized fallbacks.

All opportunities were good ones to guide the user to better outcomes.

In-product guidance

Because machine learning is complex, I used contextual tips and empty states to provide a well-informed happy journey. All guidance was focused on first-time usage, ensuring we guided users to better outcomes without adding permanent weight to the UI.

Other projects

Main Challenges

Domain immersion & taxonomy standardization

Before writing a single word, I conducted deep systemic audits of dozens of user journeys within the product.

I created a centralized jargon-to-human taxonomy that became the source of truth for both Engineering and Design teams, bridging the mental-model gap between expert data scientists and everyday low-code developers.

Product-led growth and time-to-value

When I joined, there was significant friction in the tool already. We knew users were abandoning the process during a mandatory configuration phase.

I reframed the experience from a mandatory setup burden to a progressive, value-first onboarding flow focused on expectation management and reducing cognitive friction.

Systemic governance and scalable patterns

Keeping the experience consistent within the product and across the broader company ecosystem required abstracting one-off screens into a comprehensive content governance framework.

Other projects