DigitasAI
App Feature
Redesigning an AI agent creation tool, boosting task success by 40%
DigitasAI's core capability, creating custom AI agents, had low adoption. I led the research, design, and delivery of an improved workflow for agent creation. Leading to increased feature adoption, task success, and reduced support tickets.
App
Growth
End-to-End
Shipped
Jessica Tosti, Product Manager / Jair Reina, Architect / Alex Fumes, Quality Assurance
Project Overview
Problem
Professionals found it difficult to create agents
DigitasAI is an internal platform with a suite of AI tools for professionals. One of its most-used capabilities is creating custom AI agents through the “Add Agent” feature. But the original single-screen interface was overwhelming with unclear steps, heavy technical jargon, and no sense of progress.
Business impact: Low adoption meant the platform's core value was underutilized, and high support ticket volume was draining resources.
Existing experience
Solution
Optimized the Add Agent feature for clarity and comprehension
The redesigned multi-step guided flow replaced the intimidating single screen. It used progressive disclosure to present only the necessary information at each stage, layered in contextual help via tooltips and instructional text to bridge the technical knowledge gap.
Key improvements:
Give users a sense of momentum with visible progress indicators
Clarify each step without overwhelming the UI with tooltips and inline education
Accelerate development and maintain consistency with component reuse
Redesigned agent creation flow
Tooltips and instructional text increased comprehension around the agent creation process.
Tooltips and instructional text
Impact
+40%
Increase in task success
Process Highlights
Research
Professionals found the experience to be daunting, unclear, and difficult to understand
Existing experience and interview snippets
Define
Creating a problem and goal statement drove consensus with stakeholders and the focus of the project
Journey
Mapping the end-to-end journey for creating an AI agent helped prioritize the information we could show at each step
User journeys mapped with and without adding an index
User testing
10 professionals tested and validated that the redesigned interface increased comprehension and understanding of task requirements
Agent creation designs
Delivery
I delivered high-fidelity design specs, partnered with engineering for feasibility checks, and collaborated with QA to validate final functionality before release
Preview of final design specs
