Building AI-Powered Sales Teams with finally's CRO Kevin "KD" Dorsey
An operators guide to implementing AI in your GTM org
Every revenue leader I speak with is trying to figure out how to implement AI in their GTM org but most are pretty lost. Recently I had the chance to sit down with Kevin “KD” Dorsey on the Revenue Leadership Podcast to discuss how leaders can move beyond the basic use cases that we often see discussed and drive real value from AI.
For those who don't know KD, he's the Chief Revenue Officer at finally and a renowned sales leader who's built his reputation through leadership roles at high-growth companies like ServiceTitan and PatientPop. Beyond his day job, he's widely known for his thought leadership on LinkedIn (where he has over 130,000 followers), his popular Sales Leadership Accelerator program, and his no-nonsense approach to sales excellence. KD is considered one of the most influential voices in modern sales leadership.
We're still in the early stages of what could be considered an AI revolution in sales, yet much of the discussion seems to remain at a surface level. As I joked with KD, AI adoption is “kind of like high school sex. Everybody's talking about it, everybody thinks everybody's doing it, but nobody's actually doing it."
Here are five thought-provoking takeaways from our conversation that might help you think about implementation in your own organization.
1. Start with your highest-value activities, not just the tedious ones
I've noticed that many revenue leaders tend to begin their AI implementation journey with low-value, tedious tasks like lead enrichment and account research, perhaps hoping for quick wins. KD offered an interesting alternative perspective:
"I've taken a little bit of a different approach than what I see a lot of people talking about on LinkedIn. Many focus on what they call the 'low value tasks' like account research or lead enrichment. I've approached it from the other way: what's the most valuable place I can apply this?"
For KD, conversations are the highest-leverage area. He's invested heavily in call scoring because it creates a cascade of improvement opportunities:
"If I know what's happening with AI on the calls, then I can do AI issue diagnosis. From there, I can do AI coaching. And from there, I can do AI role play; all because of what I know is happening on the call."
This is a really interesting first principles approach to AI implementation. While account research and lead scoring certainly have their place (and have driven a lot of value for us at Owner), what happens during customer conversations might be an even more critical determinant of success.
2. Turn what's in your head into documented processes first
You can't automate what isn't documented. The most important step before implementing AI is articulating the "WGLL" (What Good Looks Like), which is your definition of excellence in every aspect of your sales process.
KD explains: "You gotta capture the WGLL out of your head first. It'll make your AI so much better because you've captured it. Every single one of you has call scorecards; it's just in your head."
This step doesn't need to be complex. Instead of painstakingly documenting your processes, use voice-to-text tools or AI itself to quickly capture your thinking:
"Just sit down and use Superwhisper or just use the ChatGPT voice mode and say, 'Hey, I'm making a scorecard. I'm a sales leader at an XYZ company. I've got this many reps, we sell to whatever.' And then just talk through what you're looking for."
Daniel Kahneman's work on "expert intuition" in his book Thinking, Fast and Slow explains why this works. Experts often struggle to articulate their intuitive knowledge because it's processed through System 1 thinking; fast, automatic, and largely unconscious. AI tools can help bridge this gap by extracting and structuring this tacit knowledge.
3. Build vs. buy: Know when to create custom solutions
With so many AI vendors seeking attention these days, it can be challenging to decide when to build your own solution versus buying an off-the-shelf product. KD shared his thinking on this:
"The way I'm looking at it is: what do I think is completely out of my scope of understanding? For example, AI voice is completely outside my scope, so I'm looking at vendors for that. When I look at something like call scoring, it was relatively simple to build on my own."
His decision criteria include:
Complexity of the technology
How connected it needs to be to other systems
The risk if it fails
How specific your requirements are
It's worth considering that while some AI capabilities might require specialized expertise, many use cases could potentially be built using no-code tools like Zapier, Make or Clay combined with foundation models like GPT or Claude. One potential advantage of building your own solution is the ability to tailor it precisely to your needs and connect it seamlessly to your existing processes.
As the infrastructure around AI development continues to evolve, the build vs. buy decision may increasingly favor building for use cases central to your competitive advantage. I’m grappling with this myself and don’t think there’s a clear answer right now.
4. Learn by seeing what's possible, not through formal training
When it comes to learning about AI capabilities, KD suggests that comprehensive courses might not be the most effective approach as they can become outdated quickly. Instead, he recommends focusing on building awareness of what's possible:
"My first piece of advice is to see what's possible. Go on YouTube, watch videos on AI agents, watch what people are building over a weekend. The scales need to fall from your eyes first."
KD recommends resources like:
YouTube tutorials (specifically mentioning Liam Ottley and Jeff Su)
Following experts on X/Twitter/Linkedin
Free resources from companies like Microsoft's AI agent course on GitHub or OpenAI’s new academy
Once you see what's possible, you can better identify opportunities within your organization and find someone with the technical expertise to execute your vision:
"My freeing moment was when I realized I just need to know what is possible and then find someone who already knows how to do it."
This perspective has some interesting parallels with research on effective learning. Stanford's Carol Dweck has found that exposure to possibilities combined with a growth mindset can contribute significantly to skill acquisition, sometimes more rapidly than traditional training approaches alone.
5. Apply the "Could AI do this?" test to your calendar
To drive immediate impact, KD suggests a simple but powerful exercise:
"Look at your calendar next week and I want you to look at everything on it and ask those two questions: Could AI do this? Yes or no. Could AI help with this? Yes or no."
This exercise forces you to evaluate every activity through the lens of potential automation or augmentation. Even for activities AI can't fully replace—like one-on-ones with team members—it can still help with preparation, note-taking, and follow-ups.
KD performed this exercise when ChatGPT first launched and came to a sobering realization:
"I went on my phone and wrote out what I do as a leader, what I think has made me somewhat good at this thing called leadership. Then next to it I wrote what I think AI could do now, in three years, and in five years. In my personal opinion, there was basically nothing left by year five in terms of what I thought made a great leader and what I thought AI was going to be capable of doing."
By applying this test regularly, you can stay ahead of the curve rather than being disrupted by it.
Implementation Roadmap
Based on our conversation, here's a potential roadmap you might consider for implementing AI in your revenue organization:
Document your standards: Consider using AI to help extract and document "what good looks like" for each part of your sales process.
Evaluate high-value areas: Think about focusing on conversation intelligence before moving to administrative tasks.
Consider starting small: Beginning with a minimum viable implementation that delivers real value might prove effective.
Develop your learning system: Consider establishing regular habits for staying current with AI capabilities.
Create your feedback loop: Implementing continuous improvement processes could help refine your AI implementations over time.
Organizations that explore these principles may find they gain advantages over competitors who are still in the early stages of their AI journey.
As KD puts it: "There is no more important skill to learn in modern business than learning to use these tools."