Case Study: AI RFP project - Tim Walter, CIO Edward Don & Company

How AI Turned RFP Bottlenecks into Revenue Growth

A Leadership Case Study with Tim Walter, CIO of Edward Don & Company

Executive Summary

AI discussions often focus on experimentation, pilots, and future potential. In this Technology Executives Club® leadership case study, Tim Walter, CIO and VP at Edward Don & Company, shares a practical example of how AI was applied to a very specific operational constraint and produced measurable revenue impact.

Edward Don faced a growing backlog of complex customer RFPs, many containing hundreds or thousands of line items with inconsistent descriptions, customer-specific rules, and vendor constraints. The problem was not demand. It was throughput.

Rather than approaching the initiative as an “AI project,” Tim and his team framed it as a business problem with clear economic consequences. Their objective was straightforward: respond to more RFPs, faster, without compromising accuracy or customer trust.

The Real Business Problem Was Missed Revenue

Edward Don serves restaurants, hospitality groups, and large organizations across the country. Their RFPs frequently include massive product lists with nuanced requirements, preferred manufacturers, and distribution limitations.

Historically, a large RFP could take five to six weeks to complete. The work was manual, knowledge-intensive, and dependent on a small number of experienced employees who understood customer preferences and product substitutions.

As a result, the company could only respond to a small fraction of incoming RFPs, leaving millions of dollars of potential revenue untouched.

The challenge was clear: increase RFP capacity without simply adding headcount.

Why This Was Not an “AI-First” Initiative

Tim emphasized that the team did not start by looking for ways to use AI. They started by asking a more basic question:

Where is friction preventing us from serving customers and growing revenue?

Only after defining the problem did they explore whether AI could help reduce manual effort, accelerate comparisons, and surface institutional knowledge embedded across sales and merchandising teams.

This framing shaped every decision that followed, from technology selection to rollout strategy.

Embedding AI Inside the RFP Workflow

The solution combined an AI-powered product matching engine with existing sales and merchandising processes.

The system ingests unstructured inputs such as RFP documents, competitor quotes, product images, and inconsistent descriptions. It then maps those inputs against Edward Don’s catalog, customer rules, vendor constraints, and internal business logic.

Crucially, humans remain in the loop.

Sales and merchandising teams review AI-generated matches, validate recommendations, and make corrections where needed. Those interactions feed reinforced learning, allowing the system to improve continuously based on real-world decisions rather than static rules.

This approach positioned AI as a team member, not a replacement.

Speed Improved Immediately, Accuracy Improved Over Time

During pilot testing, the results were striking.

A 1,600-line RFP that previously required weeks of effort was processed in approximately 20 minutes. Initial accuracy exceeded 85 percent, meaning the majority of manual comparison work was eliminated on day one.

The system was not perfect at launch, and Tim was explicit about that. Expectations had to be managed. Accuracy improved as teams interacted with the system and trained it through real usage.

The tradeoff was intentional: faster responses now, smarter responses over time.

Change Management Was as Important as the Technology

One of the most important lessons Tim shared was the role of change management.

Many users initially expected AI to be flawless. When it wasn’t, skepticism surfaced. The leadership team anticipated this and reinforced a clear message: the system would improve through use, not magic.

By positioning AI as an assistive tool and emphasizing human oversight, adoption increased and resistance declined. Over time, teams began asking not whether the system worked, but how it could do more.

Revenue Impact Made the Case

By increasing RFP throughput, Edward Don unlocked access to previously unreachable revenue.

The organization set targets to increase RFP response volume by more than 400 percent. Early results showed progress toward capturing $20–30 million in new revenue annually, with expectations to grow further as adoption expands across the sales organization.

The project moved from concept to production in roughly three months, required a fraction of the investment typical of large AI initiatives, and delivered tangible business outcomes quickly.

Key Takeaways for Technology Leaders

• Start with a real business constraint, not a technology trend
  
  
• Use AI to reduce friction inside existing workflows
  
  
• Keep humans in the loop to preserve judgment and trust
  
  
• Expect learning curves, not perfection
  
  
• Measure success in revenue, capacity, and speed, not novelty
  
  

Closing Thoughts

This case study demonstrates a pragmatic model for enterprise AI adoption. When applied to a clearly defined problem, embedded into real workflows, and supported by strong change management, AI can move beyond experimentation and deliver meaningful operational and financial results.

▶ Watch the full case study to hear Tim Walter walk through the architecture, results, and lessons learned in his own words.

Hosted by Alex Jarett, Founder, Technology Executives Club®