AI powered learning ecosystem

Evolving learning from compliance to a true learning experience for 4M+ learners. Transforming traditional, destination driven learning portal into learning ecosystem that supports learners in the flow of work.

Customers + Impact
Learning management system serving 4M+ Amazon employees

Role:
Leadership interfacing, Concepts and Visual design, Interaction design, Prototyping

Timeframe
6 months

Preview of solution

AI assistant teaching mode provides on demand, custom learning with conversational sessions, editable plans and interactive check points

Learning progress tracking, nudge reminders and badges within employee’s own profile

View in depth training history, grades, learning with AI and popular courses on My learning dashboard

Before

Cumbersome discovery

Learners had a hard time finding learning as it was buried in the menu on their work portal. Some learners could not find their way into the tool.

No motivation to learn independently

Learners visit their learning portal through notifications about mandatory and compliance trainings and do not expect much in terms of active learning.

Outdated page with minimal experience

As the main priority of the product previously was moving compliance trainings over from a third party platform to in house LMS, minimal investment was made on the learner experience.

🕵🏻‍♀️ Understanding the problem

Traditional training approaches, scattered across multiple platforms with static content, fail to provide the personalized, just-in-time instruction that modern professionals require to remain competitive and confident in their role. With 39% of core skills expected to change by 2031 and AI accelerating industry disruption, employees need seamless, personalized learning integrated into their workflow.

  • 57% of employees reported difficulty finding due to fragmented systems and poor discoverability

  • Only 2% of training was self-initiated by employees in 2025

  • 70% of employees lack time to balance learning with daily responsibilities

  • 72% of employees want more career-aligned development

Vision

Transform employee learning through AI and deliver personal experience that allows seamless and continuous learning to support individual growth. This customization will be based on data such as trainings completed, searches performed, role, goals, and insights from the skills graphing engine, which maps the user’s skills to those required for various roles across the company.

🛠️ Solving the problem

The initial exploration and design was aimed at introducing micro-learning and learning ingresses into users’ news feed and redesigning outdated page. Additionally, we designed an initial AI built custom learning paths.

  • Serve micro-learning content designed to keep development engaging and accessible.

  • Deliver smart recommendations and tailored skill development

  • Provide AI assisted custom course creation based on Ai skills assessment

  • Introduce social proof and goals tracking, to increase motivation

AI Powered custom learning path creation

  • Request a topic from AI assistant

  • Skills chart powered customized path

  • Insight into training set selection

Phase 1 design

Learn home redesign into a full feature hub

  • Relevant recommendations

  • Goals section

  • Curated learning paths

  • Assigned training, Managing and filtering mandatory training

🧭 Leadership feedback

We showcased our new design vision at a yearly Amazon Leadership summit. The designs were overwhelmingly well received and earned a Most Innovative award from shareholders. We also had the opportunity to gather reactions to the tools and direction of the vision from stakeholders and customers. We received some actionable feedback for iteration.

  • Learning hub is still too destination driven and does not promote organic discovery enough

  • AI assistant needs to be more advanced and not only create collection of existing trainings but rather offer a full interactive learning experience

  • We need to push the designs further to solve customer painpoints

Next steps

After rollout of our designs at a leadership summit and gathering feedback, we advanced the solution to integrate learning across employee experiences vs maintaining one destination. Additionally, we advanced the AI assistant from just building courses to offering direct lessons.

  • Surface learning via multiple touchpoints vs via one destination

  • Deliver hyper-personalized learning in the flow of work

  • Offer interactive formats like AI Teaching, podcasts, microlearning

  • Transform development into an integrated, seamless part of daily work, supporting continuous skill growth without disrupting productivity

Phase 2 design

The next design phase was aimed at analyzing how we could introduce learning across most frequented employee touch points and offer fully interactive AI assistant learning experience. To get more holistic perspective, we collaborated across multiple teams and held a design workshop to get inputs on ideal employee touch points.

Meeting learners across multiple touchpoints. Whether it’s an employee hub news feed, AI assistant, or their personal Profile, they”ll find easy to access learning opportunities.

Closer look: AI Assistant teaching mode

  • On demand, custom learning

  • Conversational session and editable plans

  • Quizzes and progress badges

  • Different learning formats like podcasts

Closer look: Employee hub news feed

  • Digest reminders for due soon and overdue trainings

  • Microlearning modules and videos in feed informed by team interests, personal interests, and smart engine recommendations

  • Social proof and public recognition of learning achievements

  • Ready made AI assistant lessons

  • Knowledge tests that allow user to test out from some trainings

Learning from within employee’s own profile

  • Progress tracking, nudge reminders and badges

  • Shortcuts to required and recommended trainings

  • and recommendations based on their role, goals and interests directly into employee’s profile via Learn widget.

In depth learning dashboard

  • View all learning in depth including training history, grades and certificates

  • Information chips highlight due dates, learning with AI and popular courses

  • Easy ingress to learning with AI

  • All relevant recommendations

🕵🏻‍♀️ User research and feedback

We tested the new experience for personalized learning and AI assistant teaching mode with our customers.

Biggest takeaway

To make personalized and AI assisted learning truly impactful, we must build learning that adapts, not intrudes. Give users tools to shape their path, see their progress, and connect learning to outcomes, while making the experience feel lightweight, contextual, and safe.

Learning recommendations need clarity and restraint

Users liked seeing learning in their news feed if it was visually distinct, well-placed, and not overwhelming. Too many or too generic recommendations reduce perceived relevance. Social proof helps but only when it’s personal (e.g. “trending in your job family”).

AI assistant teaching mode

Users saw clear value in the AI teaching mode. The biggest wins were:

  • Bite-sized, easy to accomplish learning

  • Quizzes with real-time feedback were viewed as the most valuable part of the experience not just for validation, but for actual learning.

  • Visible progress felt focused and motivating, especially when compared to longer static courses.

Suggested learning works, when recommendations are transparent

Users preferred proactive course recommendations, but trust hinges on transparency: users want to know why something is recommended and how it benefits them. Opt-in models and light customization were favored over full automation.

Users want personalization, but on their terms

Participants favored customization (choosing duration, depth, style of AI learning session) over aggressive AI-driven personalization and are cautious about deeper personalization using sensitive data (like their performance or manager feedback).