Myna AI
| 2025
Pivoting an AI product from novelty to trusted execution system
Pivoting an AI product from novelty to trusted execution system
Role
Role
Product Designer
Product Designer
Context
Context
Full-time, founding team
Full-time, founding team
Company Stage
Company Stage
Early-stage / 0 to 1
Early-stage / 0 to 1
Timeline
Timeline
12 Weeks (from audit → pilot)
12 Weeks (from audit → pilot)
Domain
Domain
B2B SaaS / AI
B2B SaaS / AI
Platform
Platform
iOS and Android
iOS and Android
Team
Team
CEO, CTO, AI Engineer, Designer
CEO, CTO, AI Engineer, Designer
Team
Team
CEO, CTO, AI Engineer, Designer
CEO, CTO, AI Engineer, Designer
TL;DR
Myna is an AI marketing assistant for independent restaurant owners. I joined as the sole designer on a lean founding team, moving from agency work into my first 0-to-1 product role. In 12 weeks, I led a full product pivot: from a swipe-based novelty interaction model to a structured task execution system.
Myna is an AI marketing assistant for independent restaurant owners. I joined as the sole designer on a lean founding team, moving from agency work into my first 0-to-1 product role. In 12 weeks, I led a full product pivot: from a swipe-based novelty interaction model to a structured task execution system.
The core problem was not usability. It was trust. Owners were not completing marketing work because the product never gave them a clear sense of what to do, why it mattered, or whether they were making progress. Fixing that required shifting prioritization from the user to the system, and that meant making design decisions that ran directly against what the founding team initially believed users wanted.
The core problem was not usability. It was trust. Owners were not completing marketing work because the product never gave them a clear sense of what to do, why it mattered, or whether they were making progress. Fixing that required shifting prioritization from the user to the system, and that meant making design decisions that ran directly against what the founding team initially believed users wanted.
My Role & Responsibilities
My Role & Responsibilities
Led product UX strategy and interface design from audit through pilot launch. Conducted and synthesised user research independently. Designed and evaluated three structural interaction models and recommended the pivot direction. Presented findings to the founding team, including decisions that required pushback.
Led product UX strategy and interface design from audit through pilot launch. Conducted and synthesised user research independently. Designed and evaluated three structural interaction models and recommended the pivot direction. Presented findings to the founding team, including decisions that required pushback.
Architected the weekly task system including progressive disclosure, tab structure, and AI constraint logic. Built the design system foundations and reusable component library. Prototyped and tested constrained AI interaction flows using Claude.
Architected the weekly task system including progressive disclosure, tab structure, and AI constraint logic. Built the design system foundations and reusable component library. Prototyped and tested constrained AI interaction flows using Claude.
Outcome
Task Completion
Task Completion
+65%
+65%
Shifted from swipe evaluation to weekly task ownership. Completion went from <20% → +65%
Shifted from swipe evaluation to weekly task ownership. Completion went from <20% → +65%
User Confusion
User Confusion
-70%
-70%
Progressive disclosure reduced cognitive load measurably
Progressive disclosure reduced cognitive load measurably
AI Complaints
-50%
Constrained tap-based flows replaced free-form prompts
Constrained tap-based flows replaced free-form prompts
Early signals also showed improved repeat usage and clearer paths toward retention and monetization. Owners began completing work instead of merely reviewing suggestions.
Early signals also showed improved repeat usage and clearer paths toward retention and monetization. Owners began completing work instead of merely reviewing suggestions.
Context
Myna helps independent restaurant owners handle marketing without agencies or complex software. Reviews, social media, campaigns. The primary user is a solo operator running daily operations with marketing as a secondary responsibility. Mobile-first. Short, interrupted sessions. High cognitive load. No tolerance for ambiguity.
Myna helps independent restaurant owners handle marketing without agencies or complex software. Reviews, social media, campaigns. The primary user is a solo operator running daily operations with marketing as a secondary responsibility. Mobile-first. Short, interrupted sessions. High cognitive load. No tolerance for ambiguity.
This was my first 0-to-1 product role after 2.7 years as an experience designer at Wongdoody and Infosys. That background gave me strong research and craft instincts, but joining a founding team mid-build was a different challenge: I was not just designing, I was helping shape what the product should be.
This was my first 0-to-1 product role after 2.7 years as an experience designer at Wongdoody and Infosys. That background gave me strong research and craft instincts, but joining a founding team mid-build was a different challenge: I was not just designing, I was helping shape what the product should be.
Problem
Returning users were under 10%. Task completion was under 20%. Willingness to pay was low. Owners were exiting without completing anything. The issue was not usability. It was clarity and trust.
Returning users were under 10%. Task completion was under 20%. Willingness to pay was low. Owners were exiting without completing anything. The issue was not usability. It was clarity and trust.
When I joined, Myna had already shipped a swipe-based interaction model inspired by Tinder-style gestures. The founding team had been energised by Duolingo-style gamification and investor excitement around AI speed and novelty. I ran a UX audit and reviewed usability recordings. The same breakdowns appeared in every session: owners would swipe through suggestions, feel uncertain about what they were evaluating, and exit. The experience felt like a stream of options, not a path to progress.
When I joined, Myna had already shipped a swipe-based interaction model inspired by Tinder-style gestures. The founding team had been energised by Duolingo-style gamification and investor excitement around AI speed and novelty. I ran a UX audit and reviewed usability recordings. The same breakdowns appeared in every session: owners would swipe through suggestions, feel uncertain about what they were evaluating, and exit. The experience felt like a stream of options, not a path to progress.
When asked to explain the experience, users said:
When asked to explain the experience, users said:
“What am I looking at?”
“What am I looking at?”
“It would make things faster.. but I wouldn’t pay for it.”
“It would make things faster.. but I wouldn’t pay for it.”
“I can just use ChatGPT and get all this done for free."
“I can just use ChatGPT and get all this done for free."
Where I Had to Push Back
The founding team had a strong pull toward novelty. My research showed owners were not struggling because Myna lacked novelty. They were struggling because they could not figure out what to do next. I made the case to deprioritise gamification and pivot to structured task ownership. The team trusted the evidence and we aligned.
The founding team had a strong pull toward novelty. My research showed owners were not struggling because Myna lacked novelty. They were struggling because they could not figure out what to do next. I made the case to deprioritise gamification and pivot to structured task ownership. The team trusted the evidence and we aligned.
Every interview reinforced the same pattern: owners wanted the system to prioritise work for them. They wanted fewer, better-scoped actions, not more AI-generated options. The recommendation required a significant redesign of the core interaction model. The swipe model had momentum, and changing it felt like stepping back from what made the product distinctive. This was not a comfortable conversation, but that decision is what the rest of the work was built on.
Every interview reinforced the same pattern: owners wanted the system to prioritise work for them. They wanted fewer, better-scoped actions, not more AI-generated options. The recommendation required a significant redesign of the core interaction model. The swipe model had momentum, and changing it felt like stepping back from what made the product distinctive. This was not a comfortable conversation, but that decision is what the rest of the work was built on.
What the team assumed:
What the team assumed:
Faster interaction would increase task completion
Less friction would automatically translate to trust
Owners wanted more AI-generated options, not structured guidance
Faster interaction would increase task completion
Less friction would automatically translate to trust
Owners wanted more AI-generated options, not structured guidance
Faster interaction would increase task completion
Less friction would automatically translate to trust
Owners wanted more AI-generated options, not structured guidance
Research
2 weeks. 13 interviews. One clear reframe: owners wanted the system to prioritize work for them.
2 weeks. 13 interviews. One clear reframe: owners wanted the system to prioritize work for them.
Methods:
13 restaurant owner interviews · Competitive benchmarking and SWOT · Affinity mapping and behavioral synthesis · Quantitative and qualitative patterns from usability sessions
What research discovered:
owners consistently preferred the system to prioritize work for them. They didn't want more options, they wanted fewer, better-scoped ones
What research invalidated:
speed alone builds trust
What research validated:
clear scope + defined completion increases follow-through
Sample size:
13
Competitor Landscape Review
Market Positioning & SWOT Findings
Insight Synthesis: Affinity Mapping
Research Repository: Interviews & Analysis
“If you can bring me catering orders, I’ll pay you tomorrow.”
“If the app could make everyday processes easy, I’d pay for it."
“I want something that’s actively chasing opportunities for me.”
Sample size:
13
Competitor Landscape Review
Market Positioning & SWOT Findings
Insight Synthesis: Affinity Mapping
Research Repository: Interviews & Analysis
“If you can bring me catering orders, I’ll pay you tomorrow.”
“If the app could make everyday processes easy, I’d pay for it."
“I want something that’s actively chasing opportunities for me.”
Sample size:
13
Competitor Landscape Review
Market Positioning & SWOT Findings
Insight Synthesis: Affinity Mapping
Research Repository: Interviews & Analysis
“If you can bring me catering orders, I’ll pay you tomorrow.”
“If the app could make everyday processes easy, I’d pay for it."
“I want something that’s actively chasing opportunities for me.”
Sample size:
13
Competitor Landscape Review
Market Positioning & SWOT Findings
Insight Synthesis: Affinity Mapping
Research Repository: Interviews & Analysis
“If you can bring me catering orders, I’ll pay you tomorrow.”
“If the app could make everyday processes easy, I’d pay for it."
“I want something that’s actively chasing opportunities for me.”
Rather than designing for an “average user,” synthesis revealed 4 distinct behavioral groups:
Rather than designing for an “average user,” synthesis revealed 4 distinct behavioral groups:
Overwhelmed Operator
Overwhelmed Operator
Daily operations · Low time · Low tolerance
Daily operations · Low time · Low tolerance
Skeptical Pragmatist
The Skeptical Pragmatist
Cautious · ROI-driven · Trust-sensitive
Cautious · ROI-driven · Trust-sensitive
Outcome-Driven Owner
Outcome-Driven Owner
Results first · Opportunist · Revenue focused
Results first · Opportunist · Revenue focused
Time-Poor Solo Manager
Time-Poor Solo Manager
Single decision-maker · High cognitive load
Single decision-maker · High cognitive load
Strategy
Stop optimising for interaction speed. Start optimising for task completion. The system takes on prioritisation so owners do not have to.
Stop optimising for interaction speed. Start optimising for task completion. The system takes on prioritisation so owners do not have to.
3 Design principles:
3 Design principles:
Clarity
Clarity
Make priorities obvious in seconds
Make priorities obvious in seconds
Execution
Execution
Make action feel lightweight and finishable
Make action feel lightweight and finishable
Trust
Trust
Make AI predictable before powerful
Make AI predictable before powerful
Evaluated Directions
3 approaches. 1 clear winner.
3 approaches. 1 clear winner.
The refined swipe model framed work as isolated reactions with no sense of progress. The AI-ranked feed shifted prioritisation back to the user, increasing decision fatigue. The task-based workflow gave owners explicit responsibilities and clear scope. They felt in control, moved faster, and dropped off less. I recommended the task-based model.
The refined swipe model framed work as isolated reactions with no sense of progress. The AI-ranked feed shifted prioritisation back to the user, increasing decision fatigue. The task-based workflow gave owners explicit responsibilities and clear scope. They felt in control, moved faster, and dropped off less. I recommended the task-based model.
Refined swipe model
Refined swipe model
AI-ranked suggestions feed
AI-ranked suggestions feed
Task-based workflow
Task-based workflow
Before committing to a direction, I evaluated three models through the lens of small-screen behavior and interruption-heavy use.
Before committing to a direction, I evaluated three models through the lens of small-screen behavior and interruption-heavy use.
The decision: Task-based workflow.
The decision: Task-based workflow.
Swipe and feed models still required owners to manage work mentally. Tasks moved that responsibility into the system.
Swipe and feed models still required owners to manage work mentally. Tasks moved that responsibility into the system.
Designing the System
Progressive disclosure solved the last mile problem, too much detail upfront caused hesitation and drop-off.
Progressive disclosure solved the last mile problem, too much detail upfront caused hesitation and drop-off.
Early versions of the task system exposed too much detail at once. Owners hesitated before they even started. I simplified the entry point to show only what mattered now, revealing detail after intent was established.
Early versions of the task system exposed too much detail at once. Owners hesitated before they even started. I simplified the entry point to show only what mattered now, revealing detail after intent was established.
Before: Tasks visible upfront
Before: Tasks visible upfront
After: Shows overall progress to orient the user
After: Shows overall progress to orient the user
How this translated into the final product
Overview
Overview
What matters now
What matters now
Prioritization
Prioritization
Why it matters
Why it matters
Execution
Execution
How to complete it
How to complete it
This reduced initial scanning cost and helped owners orient quickly during short, interrupted sessions. Users moved from browsing to completing.
This reduced initial scanning cost and helped owners orient quickly during short, interrupted sessions. Users moved from browsing to completing.
Behavior shift we saw
Home oriented users
Home oriented users
Task list helped them prioritize
Task list helped them prioritize
Task Chat supported execution
Task Chat supported execution
Reframing Work from Daily Pressure to Weekly Progress
Daily tasks felt like obligation. Missing a day felt like failure. Weekly cycles changed that.
Daily tasks felt like obligation. Missing a day felt like failure. Weekly cycles changed that.
When tasks were framed daily, owners felt behind the moment they missed one, even when business impact was minimal. I grouped actions into weekly task cycles, intentionally limiting how much work appeared at once.
When tasks were framed daily, owners felt behind the moment they missed one, even when business impact was minimal. I grouped actions into weekly task cycles, intentionally limiting how much work appeared at once.
Before: Daily pressure
Before: Daily pressure
After: Weekly task cycles
After: Weekly task cycles
Weekly framing reduced anxiety, gave owners flexibility to act when they had time, and created a clearer sense of completion. Progress you could actually see, not a treadmill.
Weekly framing reduced anxiety, gave owners flexibility to act when they had time, and created a clearer sense of completion. Progress you could actually see, not a treadmill.
One necessary adjustment: Introduced a tab structure separating high-impact quick tasks from longer-running campaigns.
One necessary adjustment: Introduced a tab structure separating high-impact quick tasks from longer-running campaigns.
in real usage, campaign work took longer than a single week and couldn't fit inside a task cycle. I introduced a tab structure separating high-impact quick tasks from longer-running campaigns, keeping the weekly model clean without forcing campaign work into a frame it didn't fit.
in real usage, campaign work took longer than a single week and couldn't fit inside a task cycle. I introduced a tab structure separating high-impact quick tasks from longer-running campaigns, keeping the weekly model clean without forcing campaign work into a frame it didn't fit.
Before: Tasks and campaign steps were combined in a single list.
Before: Tasks and campaign steps were combined in a single list.
After: Quick tasks separated from ongoing campaign work.
After: Quick tasks separated from ongoing campaign work.
Testing in the Real World
Piloted with 10 restaurants across 2-week cohorts.
Piloted with 10 restaurants across 2-week cohorts.
What worked:
Owners completed weekly tasks within 1–2 days. Guided chat felt natural and supportive. The experience felt focused, not overwhelming.
What surfaced:
Campaigns still felt heavy for some users. Free-form prompts occasionally produced vague or generic AI responses, which fed directly into the decision to constrain AI interactions.
AI Design Trade-off
Free-form chat sounds powerful. During early adoption, it broke trust.
Free-form chat sounds powerful. During early adoption, it broke trust.
Inconsistent AI behavior and unclear outcomes were the fastest way to lose an owner who was already skeptical. Free-form prompts led to vague questions, generic responses, and occasional hallucinations.
Inconsistent AI behavior and unclear outcomes were the fastest way to lose an owner who was already skeptical. Free-form prompts led to vague questions, generic responses, and occasional hallucinations.
I reduced flexibility to make outcomes predictable. Using Claude, I prototyped a shift from free-form input to constrained, tap-based flows temporarily that kept AI inside a known lane.
I reduced flexibility to make outcomes predictable. Using Claude, I prototyped a shift from free-form input to constrained, tap-based flows temporarily that kept AI inside a known lane.
Before: users type anything → AI response quality varies
After: users tap through a constrained flow → AI stays in a known lane
This was a deliberate trade-off: we gave up breadth to stabilize trust during the most critical window 'Early Adoption'
This was a deliberate trade-off: we gave up breadth to stabilize trust during the most critical window 'Early Adoption'
How We Measured Success
Completed work, not interaction speed.
Completed work, not interaction speed.
Success was defined by whether restaurant owners could confidently complete real marketing work, not how quickly they could interact with AI. Every metric was chosen as a signal of trust, clarity, and execution.
Success was defined by whether restaurant owners could confidently complete real marketing work, not how quickly they could interact with AI. Every metric was chosen as a signal of trust, clarity, and execution.
+65%
Task Completion
Task Completion
The clearest indicator of trust and value. Rose from under 20% after shifting to weekly task ownership.
The clearest indicator of trust and value. Rose from under 20% after shifting to weekly task ownership.
-70%
User Confusion
User Confusion
Tracked through usability sessions and support feedback. Progressive disclosure and a simplified entry point drove the reduction.
Tracked through usability sessions and support feedback. Progressive disclosure and a simplified entry point drove the reduction.
-50%
AI-Related Complaints
AI-Related Complaints
Used as an early warning for AI reliability during onboarding. Constrained interactions resolved the pattern.
Used as an early warning for AI reliability during onboarding. Constrained interactions resolved the pattern.
See impact breakdown here
Retention & Monetization Signals
Retention & Monetization Signals
As owners started completing weekly tasks instead of reviewing suggestions, repeat usage improved and paths toward monetization became clearer.
As owners started completing weekly tasks instead of reviewing suggestions, repeat usage improved and paths toward monetization became clearer.
Design System
As the product expanded, consistency needed to scale with it.
As the product expanded, consistency needed to scale with it.
I built a lightweight design system to support the growing product: foundations (color, typography, spacing), components using atomic design principles, and reusable interaction patterns. These patterns enabled scale without reintroducing cognitive load.
I built a lightweight design system to support the growing product: foundations (color, typography, spacing), components using atomic design principles, and reusable interaction patterns. These patterns enabled scale without reintroducing cognitive load.
Trade-offs Made
Every call prioritized trust and follow-through over speed, flexibility, and novelty.
Every call prioritized trust and follow-through over speed, flexibility, and novelty.
Traded away
AI breadth
Chat flexibility
Swipe novelty
In favor of
Task completion on a focused set of workflows
AI reliability and predictability during adoption
Outcome-focused tasks that owners could finish
Key Learnings
What a 12-week pivot from novelty to execution taught me about AI product design.
What a 12-week pivot from novelty to execution taught me about AI product design.
1
Trust must come before intelligence
Trust must come before intelligence
People return for usefulness, not novelty. An AI product that moves fast but feels unpredictable loses users faster than a slower, more constrained one.
People return for usefulness, not novelty. An AI product that moves fast but feels unpredictable loses users faster than a slower, more constrained one.
2
Validate early, observed behavior is more reliable than assumptions
Validate early, observed behavior is more reliable than assumptions
Every root cause in this project traced back to a decision made without user input.
Every root cause in this project traced back to a decision made without user input.
3
Constraint can be a feature when users are overloaded
Constraint can be a feature when users are overloaded
Limiting AI flexibility during early adoption wasn't a compromise, it was the design
Limiting AI flexibility during early adoption wasn't a compromise, it was the design
What's Next
1
Strengthen trust and accuracy before expanding flexibility
Strengthen trust and accuracy before expanding flexibility
2
Identify the single workflow users would pay for
Identify the single workflow users would pay for
3
Build retention loops tied to real restaurant activity.
Build retention loops tied to real restaurant activity.
Open to product designer roles at companies building AI products, complex B2B systems, or anything where earning user trust is part of the design problem - not an afterthought.
If the problem is complex
and the users are skeptical,
I've been waiting for your email.
21:05:28
|
V.2026
Created by Pavan Suresh
I'm glad you made it here;
I'm currently open for new and exciting opportunities. Let's connect and create something nice.
Open to product designer roles at companies building AI products, complex B2B systems, or anything where earning user trust is part of the design problem - not an afterthought.
If the problem is complex
and the users are skeptical,
I've been waiting for your email.
21:05:29
|
V.2026
Created by Pavan Suresh