← Back to Projects

User Interviews for an AI Learning Platform

User InterviewsMarket ValidationAI EducationEnterprise L&DProduct Strategy
User Interviews for an AI Learning Platform
Project Brief
Context

OOmka had a retention problem. Users would generate a course, close the window, and never come back. Enterprise L&D teams were curious about the platform but unconvinced it could fit into how they actually worked. The product needed two things: an explanation for why users dropped off, and an honest answer to whether B2B was a viable direction.

My Contribution
  • Role:UX Researcher
  • Team:Founder, Product Manager
  • Duration:2 weeks
  • Tools:Zoom, Fathom AI, Notion, Miro

Research Impact

Seven in-depth interviews with L&D specialists surfaced 6 critical insights that reshaped the product strategy revealing a clear path from a B2C course generator toward an enterprise L&D acceleration hub.

7

L&D Specialists Interviewed

In-depth interviews with enterprise learning & development professionals revealed the full complexity of their course creation workflows.

6

Critical Insights Surfaced

From AI trust issues to personalization demand, research uncovered six key problems blocking both B2C retention and B2B adoption.

2

Strategic Directions Defined

Research produced a clear prioritized roadmap for both B2C experience improvements and a B2B enterprise positioning strategy.

My Role & Constraints

Scope of Work

Responsibilities
  • Planned and executed all research end-to-end
  • Analyzed user workflows, feedback & behavior
  • Screened and recruited participants
  • Conducted 7 in-depth L&D specialist interviews
  • Synthesized insights & recommended product strategy
Operating Constraints
  • Evolving AI model during research period
  • Limited analytics access
  • Small team with limited bandwidth
  • Limited budget
  • Short timeline with tight scope

Research Methods

With a two-week timeline, I needed methods that would give both behavioral evidence and strategic depth.

1 In-Depth Interviews

Seven semi-structured interviews with L&D specialists, recruited independently through professional communities and LinkedIn outreach.

Screening criteria
  • Works in an L&D department
  • Actively involved in the end-to-end course creation process
  • Creates a minimum of 2 courses per year

This ensured every participant had direct, hands-on experience with the full course creation cycle, not a peripheral view of it. Interviews focused on workflow mapping, pain points, AI usage, and personalization needs.

2 Platform Behavior Analysis

Reviewed session recordings to observe how real users interact with the platform, where they hesitate, what they skip, and where they drop off after generating a course.

Key Insights

01

Users Overestimate AI Capabilities

Insight

Users assume the AI can deeply expand every topic and expect expert-level accuracy. When output doesn't meet expectations, trust drops and the course is abandoned.

Opportunity

Capabilities framing, transparent limitations, example prompts, and better quality signals to set the right expectations from the start.

"It works for soft skills, but for specific topics the content is just strange, better to write from scratch.", L&D Specialist
02

Lack of Trust in Source Quality

Insight

Users want citations, authoritative sources, and validation. AI's generic content makes them feel the information is unreliable, a blocker for enterprise adoption.

Opportunity

Source mode, confidence indicators, linked reading, or integrated knowledge bases to give users trust signals alongside AI-generated content.

"We used AI assistants but weren't confident in the quality, had to double-check everything anyway.", L&D Specialist
03

Retention Drops After Creation

Insight

Users forget about the course once they close the window. No reminders, no gamification, passive text-heavy structure, nothing pulls them back.

Opportunity

Notifications, streaks, quizzes, checkpoints, and micro-tasks to create re-engagement loops that bring users back to their courses.

L&D course creation spans 10+ structured phases, from brief to delivery. Material collection, scripting, video production, assessments, metrics, iterative revision. AI currently assists at just one of them.
04

L&D Specialists Face Heavy Pain Points

Insight

Unreliable AI output, confidentiality concerns, slow expert communication, multi-tool chaos, copyright issues, and the extreme cost of video production.

Opportunity

Position OOmka as an L&D workflow optimization tool, not just a generator, addressing the full spectrum of pain rather than a single step.

"For just a couple of minutes of video, we're paying tens, sometimes hundreds, of thousands of dollars. If AI could just take the script and generate the clip, that would be immediately useful.", L&D Specialist
05

Personalization Demand Is Huge but Unmet

Insight

Specialists want role-based versions, competency-level tracks, and custom quizzes per learner, but creating personalized materials manually is impossible at scale.

Opportunity

AI-powered personalization is the biggest value driver for B2B, enabling role-based tracks and automated quiz generation at a scale humans can't match.

"If you have 100 learners, you need 100 unique quizzes, that's 500 pieces of content to write. It's completely unrealistic. The problem is just left unsolved.", L&D Specialist

L&D Course Creation Flow

Understanding the full workflow revealed where AI could add value, and where it currently falls short.

Request → Brief
Define learning objectives and stakeholder needs
Deadline pressureExtreme urgencyScope creep risk
Concept & Alignment
Strategic alignment with business goals
Expert low engagementStakeholder misalignmentHigh resource cost
Material Collection
Gathering source content, expert input
Copyright issuesAI unreliabilityManual double-checking
Target Audience Research
Learner profiling, competency mapping
Generic AI outputNot domain-specificData privacy concerns
Structure Development
Modules, chapters, learning paths
Brief ambiguityMulti-tool chaosPersonalization impossible at scale
Format Choice
Text, video, interactive, blended
Format driftNo standard systemGeneric result for top management
Scripts → Video Production
High cost and time investment phase
Script writing timeMassive production costRights & asset friction
Assessments & Practice
Quizzes, exercises, certification
Systemic cheatingUnique quizzes unrealistic at scaleHard to measure outcomes
Metrics & Customer Journey
Define success criteria and journey map
Undefined success metricsLow learner engagement
Test Launch & Revision
Pilot, gather feedback, iterate
Revision loops re-open all pain pointsNo structured revision tooling
Delivery → Feedback → Refactoring
Ongoing maintenance and improvement
Update burdenVersioning chaosNo lightweight refresh workflow

How L&D Teams Use AI Today

AI already fits naturally into L&D workflows, but current tools require manual stitching together. OOmka has an opportunity to be the unified layer.

Tools in Use
  • ChatGPT
  • Gemini
  • Perplexity
  • Midjourney
  • BeautifulAI
  • Microsoft AI
Common Use Cases
  • Drafting commercial offers
  • Creating quizzes and assessments
  • Writing scripts and presentations
  • Brainstorming formats and structures
  • Research, summarization & formatting
  • Image & video asset creation
Where it breaks down
Ethical anxiety Low output confidence Not enough depth Fails on specific topics Information must be verified Data confidentiality risk

Recommendations

B2C

Improve Consumer Experience

Quick Wins
  • Clarify what OOmka can and cannot do
  • Add content source visibility and quality indicators
  • Reminders to return to in-progress courses
  • Mini quizzes after each chapter
Longer-Term
  • More interactive course elements
  • Suggested prompts for easier iteration
  • Contextual YouTube videos embedded in lessons
B2B

Support Real L&D Workflows

Quick Wins
  • Templates matching real course-creation structures
  • Secure knowledge-base ingestion
  • Confidentiality-safe enterprise mode
  • AI-assisted expert interview summaries
Longer-Term
  • Role-based & competency-level personalization
  • Automated quiz generation (unique per learner)
  • Automated versioning of old courses
Position OOmka as a course-creation acceleration hub, not just a generator, serving both individual creators and enterprise L&D teams at scale.

Opportunity Map

Impact ↑
High Impact · Low Effort
Reminders & re-engagement
Prompt templates
Onboarding improvements
Mini-quizzes
High Impact · High Effort
Secure data ingestion
Personalized learning tracks
AI-driven assessments
Automated course refresh
Expert interview assistant
Low Impact · Low Effort
UI polish passes
Copy improvements
Low Impact · High Effort
Full LMS integration
Native video production
Effort →

Research Outcomes

Clear understanding of why users drop off and the mechanics behind abandonment
Complete map of L&D course-creation workflows across 11 structured phases
Actionable feature ideas for AI-driven improvements across B2C and B2B
Strategic lens for B2B positioning as a workflow acceleration tool
Prioritized roadmap with impact/effort analysis for next development cycle

Let's work together

Reach out and I'll share more about how I can solve your problem!