33%
of squad tickets were requests for agent changes.
Case Study
AskFlow Agents: Simplifying AI activation to drive business growth.
How redesigning a complex drag-and-drop conversational editor into a simplified template library boosted monthly activations by +40%.
57%
of accounts were activated within the first 4 weeks.
1 min
average setup time per agent after redesign.
Context & Challenge
AskFlow is a drag-and-drop conversational AI automation tool for the hospitality industry. Post-launch adoption was constrained by a high cognitive load and intimidating logic blocks.
The problem
Users could build powerful automations, but the experience felt too technical for non-developers. The result was slow activation, rising support demand, and a weak first impression.
Design & business goals
- Shorten time-to-perceived ROI
- Reduce support tickets
- Lower cognitive load for non-technical users
- Impact account upgrades through simplified agent activation
Discovery & Strategy
We aligned the product triad around what mattered most: reducing friction while preserving flexibility.
Journey Mapping
Mapped the activation path to uncover where users got stuck and what assumptions were missing.
Stakeholder Workshops
Aligned design, product, and engineering around opportunities worth prioritizing first.
CSD Matrix
Used a matrix of Certainties, Suppositions, and Doubts to focus the effort on high-confidence bets.
Key insight: Even custom workflows share common baseline structures. A standard Agent Template Library could reduce friction without removing flexibility.
Design Process
From paper sketches to AI assisted prototyping, refined through group brainstorming.
Paper sketches
I first sketched the concept by hand to think through and quickly validate the idea with the product triad.
Figma Make wireframe
From there, I built a wireframe in Figma Make to test the tool and iterate through group brainstorming sessions.
Testing & Iteration
The turning point came when the experience moved from a crowded builder to a calmer, guided playground.
Before
- Highly cluttered drag-and-drop canvas
- Heavy visual blocks and noisy terminology
- Confusing logic language such as “Reference Date”
After
- Muted gray blocks and cleaner hierarchy
- Contextual sidebar for data inputs
- Simplified language such as “Reference Moment”
I ran two rounds of qualitative testing with internal staff and hotel clients. The clearest pain points were confusion around field names, hidden descriptions, and misunderstood conditional logic such as AND vs. OR.
The Solution
A library-first activation experience paired with a calmer editor for advanced users.
Agent Showcase Library
Users could browse validated AI agents and activate them in less than a minute with predefined settings.
Refined Contextual Editor
The advanced editing surface stayed powerful but significantly less visually noisy, helping users focus on what mattered next.
Product launch concept
A visual exploration of launch messaging and visual hierarchy in a modern product experience.
Outcomes
The redesign showed that reducing complexity can directly improve both user adoption and business growth.
+40%
increase in monthly activations compared to the previous builder.
100%
growth in HSM revenue share, reaching a 50/50 split with the main platform.
89%
of clients generated sales using an agent from the library within 30 days.
32%
of activations in the launch month came through the new agent library.
< 3 min
total journey time, from entering the library to activating an agent.