×

How We Cut Proposal Time from 3 Hours to 45 Minutes with Gemini

Gemini workspace setup for sales team

Your company bought Gemini Enterprise. The sales team logged in once, got confused, and went back to manually writing proposals for 3+ hours each. Sound familiar?

This is the story of how we turned that around for a 5-person B2B sales team - achieving 80% daily usage in 2 weeks and cutting proposal time by 75%.

The Problem: $50k/Year in Unused Licenses

The client came to me frustrated. They'd invested in Gemini Enterprise for their sales team of 5, expecting immediate productivity gains. Six weeks later, usage was at 5%. The team had tried it, found it "not helpful," and reverted to their old process.

Here's what the old proposal workflow looked like:

At 20+ proposals per month across the team, that's 300+ hours of work. And Gemini was sitting unused.

The Real Problem

It wasn't that Gemini couldn't help. It was that nobody showed them how to use it for their actual workflow.

Week 1: Audit and Custom Prompt Engineering

I didn't start with training. I started by watching them work.

Day 1-2: Process Observation

I sat with each salesperson and watched them create a proposal from scratch. I documented:

Day 3-5: Building the Prompt Library

Based on the observations, I created a set of 8 custom prompts specifically for their sales process:

  1. Client Research Prompt - Analyzes company website and LinkedIn to extract pain points
  2. Proposal Outline Generator - Creates structure based on deal size and client industry
  3. Value Proposition Builder - Tailors their core offering to specific client needs
  4. Pricing Justification - Generates ROI calculations and comparison tables
  5. Objection Handler - Pre-emptively addresses common concerns
  6. Executive Summary Writer - Creates compelling one-page overviews
  7. Follow-up Email Generator - Personalized outreach after proposal sent
  8. Competitive Differentiation - Highlights advantages without naming competitors

Each prompt was tested and refined with real examples from their previous successful proposals.

Week 2: Training and Rollout

The Training Approach

Instead of a boring slide deck, I ran live workshops where we created actual proposals for their current pipeline.

Session 1 (2 hours): Each person brought a real deal they were working on. We built the proposal together using the new prompts, with them driving and me coaching.

Session 2 (1 hour): They practiced solo on another deal while I provided real-time feedback.

Daily Check-ins (15 min): For the rest of the week, quick stand-ups to troubleshoot and share wins.

The Workspace Setup

I set up a shared Gemini workspace with:

The Results: Real Numbers

After 2 weeks of implementation and 2 more weeks of monitoring:

"I was skeptical at first, but now I can't imagine going back. What used to take me half a day now takes less than an hour, and the quality is actually better." - Sales Director

Unexpected Benefits

Beyond time savings, they saw:

What Made This Work

Most AI implementations fail because companies skip the fundamentals. Here's what we did differently:

1. Workflow Integration, Not Tool Adoption

We didn't teach them "how to use Gemini." We taught them how to create proposals faster. Gemini was just the tool that enabled it.

2. Custom Prompts for Specific Use Cases

Generic prompts get generic results. We engineered prompts specifically for their industry, deal sizes, and sales process.

3. Hands-On Training with Real Work

No theoretical examples. We used their actual pipeline deals, so they immediately saw value and could apply it the next day.

4. Quick Wins First

We started with the most time-consuming, repetitive part (client research and outline). Once they saw 30 minutes saved, they were motivated to learn the rest.

5. Built-in Support System

The Slack channel and daily check-ins meant blockers got resolved in hours, not weeks. Momentum never stalled.

Key Takeaway

This implementation reduced proposal time by 75% and achieved 80% team adoption in 2 weeks. The difference? We focused on their workflow, not the tool.

How You Can Replicate This

If you're sitting on unused AI tools, here's the framework:

  1. Audit First: Watch people work. Don't assume you know their pain points.
  2. Build Custom Prompts: Generic won't cut it. Engineer prompts for specific tasks.
  3. Train with Real Work: Use actual projects, not examples.
  4. Start Small: Pick one high-impact workflow first.
  5. Support Daily: Adoption dies without ongoing support in the first 2 weeks.

The best AI implementation is one that people actually use. And people use tools that make their specific job easier, not tools they need to learn "someday."

Building something with AI?

I design and ship production systems across ML, deep learning, GenAI, LLMs, and RAG. Happy to talk through what you're working on.

Book a Free Call