Lessons Learned from a Successful AI Project – Part One

A Leadership Briefing with Tim Walter, CIO & VP, Edward Don & Company

Executive Summary

What separates successful AI projects from the many that fail to deliver ROI?

In this Leadership Briefing, Tim Walter shares practical lessons learned from Edward Don’s highly successful AI-driven RFP initiative. Rather than focusing on tools or hype, Tim outlines the foundational principles that made the project succeed: starting with a real business problem, setting realistic expectations, and leading disciplined change management.

In conversation with Alex Jarett, Founder of Technology Executives Club®, Tim explains why AI is not a magic solution, but a capability that must be integrated into existing business processes with clarity, alignment, and accountability.

This session represents Part One of a two-part Lessons Learned series.

Start with the Business Problem — Not the Technology

The first lesson is straightforward but frequently overlooked.

Before selecting AI, Edward Don identified a specific constraint limiting growth. Their RFP response process was manual, complex, and slow. Only about 5% of RFPs could be addressed due to staffing and process limitations. A 1,600-line RFP could take five to six weeks to complete.

Instead of asking, “How do we use AI?” the team asked, “What problem are we trying to solve?”

By defining the business issue first and working backward from the desired outcome, AI became a targeted solution within a broader business process — not an experiment in innovation.

Set Realistic Expectations in an AI Hype Cycle

AI carries enormous visibility and pressure. Boards, executives, and employees often expect immediate transformation.

Tim emphasizes that one of the most important lessons was clearly defining what AI would and would not do. The system would handle repetitive, research-heavy comparisons. Humans would still validate outputs, apply customer-specific knowledge, and make final decisions.

Setting expectations prevented disappointment and ensured alignment. Executive sponsorship was essential, but so was clarity. AI was positioned as part of the process — not the entire solution.

Without realistic framing, even a technically successful AI initiative can be perceived as a failure.

Change Management Determines Adoption

Technology does not drive transformation. People do.

Tim explains that the greatest risk was not technical — it was emotional. When employees hear “AI,” many fear job replacement. Resistance emerges when people feel they are losing something.

The leadership team reframed the narrative. AI would remove repetitive tasks, allowing team members to focus on higher-value work: customer engagement, validation, and strategic decision-making.

Instead of replacing roles, the initiative realigned responsibilities.

Communication, engagement, and executive alignment ensured that the project was championed internally. AI became a team member, not a threat.

Measurable Business Impact

The results were tangible.

Pre-project, Edward Don responded to roughly 5% of incoming RFPs. A large RFP required weeks of manual processing.

With AI integrated into the workflow:

• A 1,600-line RFP could be processed in approximately 20 minutes
  
  

• Accuracy reached approximately 86%
  
  

• Throughput doubled, tripled, and in some cases quadrupled
  
  

• The organization set a revenue growth target of $20–30 million in the first year
  
  

More importantly, AI shifted the focus of human effort toward exceptions and value-added judgment instead of repetitive research.

Key Takeaways for Technology Leaders

Begin with the business constraint, not the technology.

Define clear outcomes and realistic expectations before launching AI initiatives.

Align executives, teams, and stakeholders around what AI will and will not deliver.

Treat AI as part of a broader business process, not a standalone solution.

Prioritize change management as highly as technical implementation.

Closing Thoughts

AI success is rarely about algorithms alone. It is about disciplined leadership.

By starting with a real business problem, managing expectations, and guiding change thoughtfully, Edward Don transformed a manual bottleneck into a scalable growth lever.

This is Part One of Tim Walter’s Lessons Learned series. In Part Two, Tim explores reinforcement learning from humans, why data does not need to be perfect to begin, and additional insights from scaling AI inside the enterprise.

Be sure to look for Part Two to be published soon.

Guest Speaker: Tim Walter, CIO & VP, Edward Don & Company
Host: Alex Jarett, Founder, Technology Executives Club®

Learn more about Tim Walter and watch all of his Leadership Briefings on his Board of Advisor page here.