Genie Assistant

An AI-powered contextual assistant embedded across Bigtincan's platform

Overview As large language models (LLMs) like ChatGPT gained momentum, our team at Bigtincan saw an opportunity to deliver similar capabilities natively within our sales enablement platform.

But instead of creating just another generic chatbot, we focused on building a truly context-aware assistant that added value for our diverse user base—sales reps, marketers, and enablement teams across industries.

My RoleAs Lead Product Designer, I drove the end-to-end design strategy—from research and early concepts to prototyping and production—working closely with the Product Manager and the Engineering team.

I also coordinated internal feedback loops with Sales Engineering and Customer Success to ensure alignment with real user needs.

Key Design Goals
Lower the barrier to AI adoption
Help users who are unfamiliar with prompting LLMs by offering contextual suggestions and guided interactions.

Respect user context
Genie adapts based on where you are in the platform (e.g., viewing a document, searching, reviewing content) and offers intelligent actions accordingly.

Balance AI power with UX trust
Communicate how answers were generated, where data came from, and what actions are possible—without overwhelming the user.

Design Process

Discovery & Use Case Mapping We began by mapping tasks that users already performed within our platform: searching content, preparing for meetings, summarizing documents, learning about products. Then, we explored how LLMs could augment those tasks—rather than replace them.

Prompt Framing & Input ModelsOne challenge was the blank prompt dilemma—what do users ask? To solve this, I introduced:

Prompt pills: prefilled contextual chips (e.g., “Summarize this content”) inside the input field

Onboarding samples: suggested prompts based on user persona and page context
Conversational memory: ability to follow up naturally based on earlier messages

Context-Aware System DesignGenie is not a standalone chat—it pulls context dynamically:

In the Content Hub, Genie knows which document or folder you’re in.
While searching, it references your query and current filters.
During meetings, it accesses speaker notes, agenda, and audience.

This required close collaboration with engineering to define how data is passed to the LLM and how to structure prompts invisibly under the hood.

Visual & Interaction DesignThe UI needed to feel lightweight but powerful. I created:

A responsive assistant panel that slides in from the side
Highlighted sources and confidence tags
A guided “Ask me anything” input with placeholder examples

I also contributed design tokens and interaction specs to our design system to support future assistant use cases.

What I Learned

Designing for AI is not just UX—it's about orchestrating data, backend logic, and human trust

Contextual guidance is more important than raw LLM power

Successful AI features require thoughtful onboarding and ethical transparency