I've Got the Power: Using AI to Even the Playing Field

Procurement professionals often feel under-resourced compared to their sales counterparts, who have entire teams dedicated to responding to RFPs. To “even the playing field,” procurement needs tools that can optimize data and automate the heavy lifting.

In this special “jam session” episode of Proc & Roll, Conrad Smith is joined by David Johnson for a hands-on demonstration of how to use AI tools like Google’s Notebook LM and Gemini Gems right now. David shares his unconventional path into the industry and provides a live walkthrough of how he used AI to manage five simultaneous RFPs without a large team.

Topics Discussed in This Episode:

• Falling Through the Ceiling: David shares his unique origin story of “falling through the ceiling” into a procurement career, where he independently invented a PO system in Excel to solve invoicing confusion.
• Why Procurement is Made for AI: David argues procurement should be at the “forefront of AI” because it relies on structured data like SKUs and contracts, allowing teams to “even the playing field” against sales departments.
• Automating RFPs with Notebook LM: The episode features a practical demo of Google Notebook LM, showing how to synthesize vendor responses, generate audio overviews, and query specific details like business continuity plans.
• Building an Automated Scorecard: David explains how to use AI to build a full scorecard based on requirement documents, streamlining the process by asking the tool to rate vendor responses against specific criteria.
• Notebook LM vs. Gemini Gems: The guest clarifies that Notebook LM is a learning platform for large data, while a Gem is a “customized creative partner” that can be trained on specific tones and emails.
• The Future: Predicting Outcomes: David predicts the next wave of AI will focus on “predicting outcomes,” with agentic AI modeling negotiation scenarios and analyzing financial filings to build strategy.

Watch now or read the transcript below.

Transcript: Proc-N-Roll | I’ve Got the Power: Using AI to Even the Playing Field

Conrad: Hello everybody and welcome back to Proc and Roll. Today I am really excited for another jam session with David Johnson. We’re going to do hands-on interaction with a couple of different AI tools that you can use and implement in your work right away . But first, I have to ask about this incredibly unique story about you falling through a ceiling .

David: I didn’t go to school for procurement; I was a theater kid who liked computers. My first office job was at a brokerage firm as a runner. The head of IT knew I was in school for computer science, so he had me running cables through the ceiling . I was halfway in the ceiling running cable over a trader’s head when the drop ceiling crumbled. The trader was the owner’s son . To keep me out of his face, my boss put me in an inventory stock room. While there, I saw the Accounts Payable lady struggling with invoices, so I built a system to create POs in Excel and do a three-way match. The controller found out, gave me a raise, and said, “This is what you do from now on” .

Conrad: That’s amazing. We are primed to take advantage of AI because data is the food AI lives on, and we have so much of it . I’d love for you to walk us through the use case of Notebook LM as an automation framework to process RFPs .

David: The key difference is that generative AI like ChatGPT is for ideation, while Notebook LM is an education tool meant to take large swaths of data and make it manageable . For an RFP, I took vendor responses and my requirements document, converted them to PDF, and threw them into Notebook LM . It instantly generates an “audio overview,” which is essentially a podcast of two people going through your documents .

Conrad: I’m imagining I can ask questions like, “Which of these four vendors has the best disaster recovery plan?” .

David: Exactly. I can ask, “Which of these responses has the most comprehensive business continuity response?” and it gives an overview of the availability guarantees, support coverage, and data flow for each vendor . I even used it to build out a full scorecard based on the requirements document, asking it to give me a score for each subsection based on the vendor responses .

Conrad: You’ve been doing RFPs for 30 years. How has this changed the efficiency of the process? .

David: When I did this, I was managing five RFPs at the same time with a total of 28 vendor responses . This tool made it a thousand times more efficient. I got double the response rate from stakeholders because I shared the notebook with them, and they could listen to the audio overview or create mind maps instead of just reading raw documents .

Conrad: I want to pivot to another technology: Gemini Gems. How do you think about when to use a Gem versus Notebook LM? .

David: Think of a Gem as a customized creative partner or AI assistant, whereas Notebook LM is a learning platform. A Gem is limited to 10 files, but you can train it on your emails and tone . I built a Gem for a Salesforce contract where users could ask, “Tell me about my Salesforce licensing,” and it would answer based on our specific enterprise agreement .

Conrad: I actually created a “Proc and Roll Assistant” Gem a few minutes ago and told it to create all responses as a poem. It just output: “A chorus of wisdom shared in rhyme, on modern procurement for this very time.” That is fantastic .

David: Looking ahead, the supply chain is ripe for AI to predict outcomes . Imagine an agent looking at contract terms, modeling scenarios like “what happens if I cut usage by 15%,” and reviewing a vendor’s 10K to build a negotiation strategy .

Conrad: This is the renaissance for tech nerds right now. As we wrap up, what is your advice to your younger self?.

David: AI is not going to take your job; it’s going to take the parts of your job you don’t want to do, like auditing POs . My advice is: Always be curious. Never stop learning. The more you know about how your business and processes connect, the better you’ll be

This transcript has been edited for clarity while maintaining all substantive content