Breaking the Failure Loop: Crafting an Adaptive AI Habit Engine for Modern Professionals
Impakt started as a personal build — no developers, no team, just a problem I couldn’t ignore. Frustrated by the gap between ambition and execution, and tired of productivity tools that punished inconsistency instead of supporting it, I took ownership of the entire journey — from research to strategy to prototyping — to explore how AI could create adaptive, emotionally intelligent systems for real life.
ROLE
Sole Product Designer
TEAM
Personal Project
DURATION
5 months
SKILLS
Qualitative Synthesis
Survey Design
UX Architecture
Interaction Design
Feature Prioritization
Feedback Loops
Impact
600+ professionals surveyed
3 Core systems designed
Reduced cognitive overwhelm
64% users feel overwhelmed when translating big goals into daily routines.
7/10 reported dropping new habits within a month due to lack of guidance or motivation.
85% seek more flexibility and emotional support than existing productivity tools provide.
Big goals overwhelm me. I am usually clueless where to start.
— Jin, 26, Health Coach
Missing days makes me quit. Apps don’t support real life.
— Maria, 23, Freelancer
Building habits alone is hard. I want connection and support.
— Aisha, 26, Educator

Pain Points
With all this research in mind, we scoped out the key 4 objectives we wanted to achieve for this project:
Design adaptive micro-steps that reduce cognitive load
Replace streak pressure with flexible recovery systems
Build transparent, collaborative AI support
Scope
I defined the boundaries early to prevent feature sprawl disguised as innovation. The goal was not to build “a productivity app,” but a focused planning system where AI and manual control coexist without conflict. By narrowing the scope, I protected the core experience from unnecessary complexity.
User Journey
I mapped the journey from first-time onboarding to daily plan adjustments to understand behavioral friction points. The emphasis was on reducing cognitive load at moments of decision — especially when editing or negotiating with AI. Every interaction was designed to feel controlled, reversible, and intentional.
Sketches
Sketching allowed me to test structure before committing to polish. I explored layout hierarchies, task-state visibility, and AI chat placement through rapid iteration. Those early explorations clarified how structure and branding could merge into a system that feels cohesive rather than assembled.
Example of practitioner feedback for the major app pages












