Briefbot

BriefBot is an AI-powered assistant that transforms chaotic client calls into structured briefs in real time. I was brought in as the sole product designer to turn this concept into a working prototype used in investor demos and early agency pilots.

May 15, 2025

April 10,2025

Problem Inspiration

Rooted in real agency struggles

Problem Inspiration

Rooted in real agency struggles

Problem Inspiration

Rooted in real agency struggles

Role

Sole Product Designer (0 → 1)

Role

Sole Product Designer (0 → 1)

Role

Sole Product Designer (0 → 1)

key Strategic Decisions

Trust-First AI UX, Human in the loop control, Scope to Impact

key Strategic Decisions

Trust-First AI UX, Human in the loop control, Scope to Impact

key Strategic Decisions

Trust-First AI UX, Human in the loop control, Scope to Impact

Context

How Might We

How might we help agency PMs run smarter client calls and capture actionable briefs without adding more tools or interrupting their flow?
How might we help agency PMs run smarter client calls and capture actionable briefs without adding more tools or interrupting their flow?

Key Challenges

Key Challenges

Key Challenges

Pain, Insight & Opportunity

Pain, Insight &
Opportunity

Scattered inputs, vague requirements, and lost context.
Agency PMs often struggle to extract actionable briefs from conversations, leading to project misalignment and wasted hours.
Most PMs “wing it” and rely on memory or scattered notes
Designing for Speed and Simplicity
Indiability to focus on instruction and student engagement. Need for
streamlining these tasks and support professional growth.
Creating simple worksheets takes approximately 15 minutes each day, indicating a need to streamline and speed up the process
Indiability to focus on instruction and student engagement. Need for streamlining these tasks and support professional growth.


Lack of Visibility into Individual Performance
While each student's performance is mapped across various topics,
there is a lack of clear visibility into individual strengths and
weaknesses making it challenging to identify areas for improvement
High Administrative Burden Reducing Instructional Focus
Indiability to focus on instruction and student engagement. Need for streamlining these tasks and support professional growth.
Provide real-time structure without disrupting flow
While each student's performance is mapped across various topics,
there is a lack of clear visibility into individual strengths and
weaknesses making it challenging to identify areas for improvement
While each student's performance is mapped across various topics, there is a lack of clear visibility into individual strengths and weaknesses making it challenging to identify areas for improvement

My Approach

AI Driven Design

The biggest challenge? Making AI feel helpful, not disruptive. I focused on trust, transparency, and user control while balancing automation with human agency.
The biggest challenge? Making AI feel helpful, not disruptive. I focused on trust, transparency, and user control while balancing automation with human agency.

0-1 Product Launch Strategy

Collaborated with PMs, engineers and internal stakeholders to create seamless AI exerience
Discovery Interviews
5 agency PMs personas across branding and dev agencies
Key insight: “I never know what’s missing until the project is halfway done.”
Need: A tool that listens and nudges

01

Brief Schema Design
Define what a “great brief” includes with user journey mapping. Allign user goals with business success criteria.

02

AI Interaction Design + Design System
Designed trust-building patterns with a fallback system for vague responses or hallucinations
Card-based brief editor with real-time AI annotation
Clean, neutral system aesthetic for adaptability

03

Testing + Iteration
Ran 3 pilot simulations with sample calls and prompts
Feedback: “Love the nudges. Just don’t overdo it.”
Iterated to prioritize user agency over AI automation

04

Deliverables

Ensuring Product-Market Fit

  1. Scrollable brief editor with live AI-generated cards
  2. Visual nudge system with missing-field reminder
  3. Summary review page with confidence marker

Next Steps on Future Directions

Refine the AI interaction model to enhance natural language understanding and increase the accuracy of personalized content generation. Begin user pilot testing to gather real-world feedback on AI-driven workflow improvements.

Lesson learned

Human-in-the-Loop by Design: Balancing Automation with Control

Designing with AI ≠ full automation: It’s about co-creation, letting humans lead
Trust patterns are essential: Labels, toggles, and control points build user comfort
Designing with AI ≠ full automation: It’s about co-creation, letting humans lead
Trust patterns are essential: Labels, toggles, and control points build user comfort

Send a Message