Key Result
Proposal prep time cut by ~50%
The Challenge
Consultants at the agency were spending hours searching through Google Drive, Notion, and Slack to find previous proposals, client research, and campaign data. This created inefficiencies, duplicated work, and led to inconsistent proposal quality. New team members struggled even more without institutional knowledge.
Our Solution
We developed a secure, private AI copilot that:
Indexes all company documents (Google Drive, Notion, Confluence)nUses semantic search to find relevant proposals, research, and case studiesnAnswers questions in natural language ("Show me all healthcare campaigns from 2023")nSuggests relevant content when drafting new proposalsnWorks entirely within their security perimeter - no data leaves their system
Built with Python, LangChain, OpenAI embeddings, and Pinecone for vector storage. Simple Slack integration for easy access.
Technologies Used
The Results
After 2 months of use:
~50% reduction in time spent on proposal preparatio
Consultants can find any previous work in seconds vs hours90% team adoption rate within first monthnNew consultants onboard 3x faster with instant access to institutional knowledge
The system has answered over 3,500 queries with high accuracy.
"This AI copilot has become our team's most valuable tool. It's like having instant access to every smart person who's ever worked here. Proposal quality has improved and our team is happier."
Sarah Chen
Head of Operations, Creative Minds Agency