featured PROJECT: 2025
AI Meeting Scribe: Private Internal AI for Knowledge SUmmarization & Q&A
From discovery to deployment: an AI solution that solved a real problem, sold itself, and changed how a library collaborates.
Summary
AI Meeting Scribe is an internal AI tool I built for Boulder Public Library District to solve a growing operational pain point: meeting notes were scattered, unstructured, and hard to retrieve. Staff needed a secure way to summarize and query meeting content without exposing data to external AI platforms.
I handled everything from discovery with leadership to solution design, proof-of-concept, demo, objection handling, and launch.
The result: a lightweight, private AI assistant that turns transcripts into actionable insights.
Problem / Opportunity
The organization held dozens of meetings each month, but:
- Notes were trapped in shared drives or personal notebooks.
- Nobody had time to read full transcripts.
- Searching for “what David said about outreach last week” was nearly impossible.
- Privacy concerns prevented the use of external tools like Otter.ai or ChatGPT.
Leadership wanted an internal, secure, AI-powered way to make meeting knowledge more useful.
Solution / What I Built
AI Meeting Scribe was built as a private, in-house AI assistant that:
Summarizes meeting transcripts into structured key takeaways, action items, and speaker-specific notes.
Allows users to query meeting content conversationally, e.g. “What did Aimee say about the IT onboarding plan?”
Keeps all data secure and internal, ensuring nothing leaves the organization’s environment.
Provides clean, readable summaries that save staff hours of time and reduce missed follow-ups.
Process & Story
- Discovery: I sat down with the Director (David) and leadership team to understand their vision for AI adoption. I didn’t pitch—I listened. Took notes. Uncovered pain points and privacy concerns.
- Research: At a libraries & AI conference, I gathered examples of real-world use cases and privacy-safe architectures.
- Solution Design: I built a working prototype of AI Meeting Scribe using secure integration methods and prompt engineering optimized for meeting content.
- Demo: Instead of explaining it technically, I ran a live demo—uploaded a transcript and showed the clean, structured summary appear in seconds. Then I asked the AI: “What did David say about budget?” and let it answer.
- Objection Handling: Addressed privacy concerns by explaining how no data was shared externally. Reassured users with transparent documentation and step-by-step onboarding guides.
- Rollout: Created simple documentation and tailored presentation decks to non-technical staff. The launch was smooth, with immediate adoption from multiple departments.
Impact / Results
- 200+ staff members adopted the tool across multiple departments.
- Reduced time spent reviewing meeting notes by over 60%.
- Turned hours of manual note parsing into seconds of insight retrieval.
- Unlocked previously “hidden” knowledge, improving collaboration and follow-up execution.
- Strengthened trust in internal AI initiatives by solving a real pain point securely.
Sales Engineer Takeaway
This was pure Sales Engineering in action:
- Conducting consultative discovery to understand business problems.
- Translating complex tech into clear business outcomes.
- Creating a proof-of-concept aligned with real user needs.
- Delivering live demos that close the gap between “idea” and “yes.”
- Handling objections around security and compliance.
- Driving adoption through clean documentation and onboarding.
Tools / Tech Used
- Azure AI / Internal Chatbot Framework
- Secure Transcription Processing
- Prompt Engineering for summarization and Q&A
- Internal documentation and onboarding playbooks
- PowerPoint Launch Deck for live demos
This project showed me exactly how powerful well-positioned AI solutions can be when combined with sales engineering skills.
Let's talk if you’d like to hear how this same approach can drive adoption and ROI at your org.
© 2025 Ahmad Hakimi