The Human-AI Balance: 10X Your Output Without Losing Yourself | Bella Cowdin

Listen

About the Guest

Episode transcript

AI as a Force Multiplier for Revenue Teams

Brendon Dennewill: Bella Cowdin, welcome to the RevOps Champions podcast. Today we're going to explore a big question about humans and AI, and hopefully get into the day-to-day realities of revenue operations and how to make your CRM more valuable. For those listening, Bella is a HubSpot certified trainer and Denamico's very own Senior HubSpot Consultant. She helps teams simplify automation, strengthen trust across departments, and turn customers into loyal champions. Bella, it's great to have you.

Bella Cowdin: Thank you, Brendon.

Brendon Dennewill: I'm going to dive straight in. What is one example where you see AI having a big positive impact in revenue operations?

Bella Cowdin: That's a big question. I'll pull from HubSpot's INBOUND, which just happened this past week in San Francisco. One of the things that stood out to me was how marketing agents help your marketing operations people do their job better, sales agents and sales assistants help your salespeople sell faster, and service agents and service assistants help your customer service people build better relationships.

What stood out to me most wasn't necessarily a particular feature, but the emphasis HubSpot made on how AI can help your best people 10x, 20x their output. In terms of how AI is going to impact clients' businesses over the next two to three years, I see things moving faster and getting more personalized. I also see humans relinquishing responsibility to AI to perform repeatable tasks: making an email campaign more personalized, or personalizing a quote in a way that would take you or me 30 minutes to navigate, but an AI assistant or agent less than a second to complete.

So I see things moving faster if we give our people the right tools to train up and specify their agents to 10x or 20x their output.

Brendon Dennewill: I think you're spot on. The big annual INBOUND event this past week, and the new Loop Marketing Playbook that was launched, represent a really significant decision by HubSpot. They've essentially built their entire business around INBOUND for 19 years since their inception in 2006. Last week they came out and said that playbook is broken. SEO just doesn't work the way it did when they started. It's being replaced by an AI-driven, agentic web.

One of the paradoxes you alluded to is interesting: we've been living in a world of personalization for quite a few years now, and with AI you might think things will become less personal. But what's actually going to happen, as outlined in the new Loop Marketing Playbook, is things are going to go from being personalized to truly personal. Because with an AI agent, whether you're a marketer, salesperson, or customer service professional, you can now make every touchpoint personal rather than just personalized. HubSpot has done a really good job of reinforcing that this whole AI movement will only work if it's led by humans first, with AI making those humans better.

Bella Cowdin: Exactly.

The Risk of Over-Delegating to AI

Brendon Dennewill: So there's obviously a lot of exciting things to come. But where do you have some concerns about how people are approaching AI incorrectly in your day-to-day work?

Bella Cowdin: I want to quote one of our teammates, Tina, who made a brilliant analogy. In the late '80s and through the '90s, computers were the next big tech push. Everybody was asking themselves how they were going to use this in their day-to-day. Do I replace my Rolodex with this? What are we doing here? People didn't know how to use the computer, so they were using it improperly, just turning it off and on. That's a minor example, but it illustrates the point.

In terms of issues with AI, I think we're in a really similar boat. AI is coming out, and people are largely trying to wrap their arms around what's the best way to use it. What do I relinquish to AI versus what do I keep doing myself? What I've been seeing is too much relinquishing to AI.

Striking that balance is going to be critical, because when people relinquish too much, there's a lack of authenticity in the output. Beyond that, there was a study from an Ivy League institution that found if you give up too much of a core functionality you use your brain for, like a muscle, you use it or you lose it. You actually atrophy that part of your brain faster. So if you fully relinquish the personal touch to AI, you lose the ability to do that yourself.

Bella Cowdin: I had an instance where a client uploaded their process and what they thought were training scripts into AI and said, 'All right, we're good to go. Here is our script.' It was read over, but not edited. It was raw AI output. That client was relinquishing not only their decision-making, but also their ability to articulate what they actually wanted their people trained on, as opposed to just going with everything the AI generated. The give-and-take with AI is the biggest thing to watch as you're using it.

Brendon Dennewill: Which comes back to what you were saying earlier. People need to be thinking about how they're going to use AI to make themselves 10 times better, not about having AI do what they should be doing themselves. Because that's not going to end well.

Bella Cowdin: Right. I can actually think of another instance from a previous role where a coworker received an escalation email, put it into ChatGPT asking how to respond, and then pasted that response directly. The client actually called them out, saying, 'This doesn't sound like you. I don't feel like all of my points were addressed. Did you use an agent to answer me?'

Brendon Dennewill: Everybody's learning, and I think the one simple reframe that really helped me over the last six to nine months was: how can AI help me do this better? That's very different from asking how AI can do this for me. If AI is doing it for you, how are you getting better at your job?

Bella Cowdin: Exactly.

Prompting Frameworks: Getting Better Results from AI

Brendon Dennewill: Have you come across any tools or frameworks that help people think through this more effectively? When they're giving ChatGPT or Gemini or whatever AI they're using more insight about what their prompt should include so that the result is more likely to be useful? A good prompt isn't the solution to everything, but additional context will definitely give you more to work with. That said, you still have to edit and make sure you're good with whatever the AI produces.

Bella Cowdin: Off the top of my head, I don't recall what CRIT stands for, but that's the four-step process you mentioned. And it lines up with what was shared at INBOUND, which introduces the concept of 'express and curate.'

Brendon Dennewill: CRIT comes from Geoff Woods, who wrote The AI-Driven Leader. The C is Context, the R is Role, so you ask the AI to take on a role that has the best perspective to give you the result you need. The I stands for Interview, where you tell the AI to interview you. An additional step I use is asking it to pose one question at a time, because otherwise you get a hundred questions at once, which is overwhelming. That helps the AI gain deeper context. Then the T, the last piece, is Task: define the specific objective you want AI to help you with.

If you start using this framework, within a few days you very quickly start to realize that AI is there to help you, not to do things for you. It's a very different mindset.

Bella Cowdin: And I was going to connect this to what HubSpot just laid out with the Loop. The 'express and curate' concepts in my mind almost lend themselves to the same idea for the user: tailoring how you want the AI to think, expressing who you are and the context in which you want the output. Regardless of which framework you use, the idea of putting yourself first in order to use AI is still what 'context' means, or what 'express' means, when you think about that first step of any AI framework.

And that circles back to the earlier point about people relinquishing too much to AI rather than using it in a strategic, ethical way that actually produces good output.

Brendon Dennewill: Absolutely. Whatever framework it is, because there are increasingly more of them, the good news is they're all essentially saying the same thing. So search for a framework to get better results from your AI tool of choice. Just understanding what the framework is trying to do gives you insight as a human into how you can use AI to help yourself do things better.

HubSpot's AI Memory and the Power of Persistent Context

Bella Cowdin: Another thing worth pointing out while we're on this topic: HubSpot's AI now hosts memory. It will remember what you prompted it months later, and you can reference that context going forward.

Something I've done personally that I want to try professionally now that HubSpot's AI has memory: I use a prompt when I engage with an agent that says, 'Hi, I'm Bella. I live in central Colorado with my partner and my dog. I enjoy these things, I don't enjoy these things. That's me. Now give me recommendations based on that.' With HubSpot, I very much see myself using one of their service agents or assistants the same way: 'Hi, I'm Bella. I'm a Senior Consultant. I do some solutions architecture work. I'm primarily involved in the services and delivery portion of consulting. Here are some of the things I'm doing. Here are some of the challenges my clients face. Use that context moving forward.' I haven't done this yet professionally, but I know the memory is there, and I think it could be really helpful.

Brendon Dennewill: Absolutely. And I think that's a really good point. It confirms what AI leaders across the planet are talking about: as we get better at working with AI, AI will get better at working with us. But there are fundamental rules we're all learning quickly. The faster we learn them, the better off we'll all be for ourselves, our families, our companies, and our communities. If we lead this AI era with a human-first approach, that'll help us avoid a lot of the potential pitfalls from a negative perspective.

Bella Cowdin: And I'd just like to circle back on relinquishing too much to AI. A really simple solution is just: don't. Here's a funny example. I heard about someone who uploaded a picture of all their clothes to AI and said, 'Dress me for the week.' Do you really want to relinquish that small, unique part of presenting yourself every day to AI? I'm not saying our clients are doing that, but it's an extreme example that illustrates the point: you're almost relinquishing your identity when you do that.

Brendon Dennewill: If you have a paid license and you're keeping this in a custom project, then it's not going to be used anywhere else. But I'd absolutely agree that if you're not using a paid license and you do that, essentially anyone could know what your closet looks like. It's about knowing the consequences of not doing this the right way, without being so scared that you don't do anything with it at all.

Bella Cowdin: I was more coming from the lens of not wanting to relinquish that unique part of my decision-making, let alone the compliance side of it. But I see what you're saying.

Brendon Dennewill: Some of the smartest people who've ever lived, Steve Jobs among them, never had a wardrobe. Or if they did, everything was the same, so they didn't have to think about it. Every day: black shirt, jeans. That's where every person has to decide for themselves what makes sense in terms of simplifying their life.

Bella Cowdin: I totally see what you're saying. Though I think that gets into the difference between male and female. I was at lunch with a group of friends one day and I mentioned, 'When was the last time you wore a yellow shirt?' And all the men at the table said they'd literally never thought about it.

Brendon Dennewill: Yes, I hear you.

HubSpot Data Studio: Breaking Down Data Silos in RevOps

Brendon Dennewill: Looking ahead, are there any concrete examples or opportunities where you see AI delivering the next big competitive advantage in RevOps over the next six to twelve months? Specifically around moving beyond basic automation?

Bella Cowdin: The biggest thing for me, hands down, from INBOUND was HubSpot acknowledging head-on that not all customer data lives in HubSpot. Up until INBOUND this year, we kept hearing the sentiment: 'If it's not in HubSpot, it didn't happen.'

This was the first year that HubSpot stood up in front of everybody and said, 'We know not all your data is in HubSpot. It could be in a Google Sheet a CSM made five months ago, or ad hoc anecdotal data in an email that's not captured anywhere other than an activity record, or in another system entirely.'

So this year HubSpot announced Data Studio: an AI-driven, codeless integration with outside, non-HubSpot sources to bring your data into HubSpot and enrich your records. The example used was Snowflake, where they accessed Snowflake Datashare outside of HubSpot and brought that data in under a tab in the contact or company view.

I still want to spend more time with it and see what comes native with Data Studio, and whether there's an option to connect a third-party platform without code. But the biggest thing worth noting is the ability to de-scatter your data with Data Studio. And then, with marketing, sales, and customer service 10x-ing their output with AI, they can bring in that contextual data to have truly enriching conversations that they probably wouldn't have had access to before.

That is the biggest one for me. Step one, version A: a codeless, AI-powered integration to an outside third-party system that brings data into HubSpot. Amazing.

Brendon Dennewill: That is huge. And I think that, in conjunction with HubSpot announcing their integration with Claude about six weeks ago, then ChatGPT three or four weeks ago, and then Gemini last week at INBOUND, which is arguably the biggest of the three because there are so many businesses already operating as Google shops, that becomes really, really interesting.

What you're describing is essentially the rebrand of what used to be called Operations Hub, now called Data Hub, which is where Data Studio lives. And we're very excited about the power that'll bring. We've been looking at this for the last twelve to fifteen months around how to help companies get ready to use AI, and the very first step is clear: you can't really do much with AI if you don't have organized data. HubSpot has spent a lot of hours and energy over the past year or two making sure it can now do a lot of that for you.

Allowing HubSpot customers, whether directly or with the help of a partner like ourselves, to get access to all their data and make decisions in real time is one of the most powerful things you can do as a leader.

Bella Cowdin: Another one I want to highlight while we're on this topic is AEO: moving from SEO to AI Engine Optimization. For those who may not know, it's the shift from search engine optimization to AI engine optimization.

As most people have probably experienced by now: if you Google 'Top 10 Indian restaurants in San Francisco,' AI will serve up its own recommendations drawn from multiple sources. So instead of SEO, where you're trying to rank first, second, or third on a page by keywords, AEO is about enriching your content for helpfulness and your ability to show up in an AI recommendation.

Some people probably aren't even using Google anymore. They're just talking to their assistant. So in order to get recognized, it's not about ranking on a search page. It's about being cited in an AI-generated answer. That's another massive shift we're going to see, not just with HubSpot customers but across every industry.

Brendon Dennewill: That's a really important call out, and I'd agree it was a huge announcement. Kip Bodnar, the CMO of HubSpot, came on stage and noted that HubSpot has mountains of data, so they could see what was happening. Traffic was just tanking. As they started testing AEO, they found that the conversion rate from AI-generated results was three times higher than anything from traditional search engines.

Now they've made that available to all of us. Anyone listening can go to their favorite AI or search and look up the AI engine optimization grader in HubSpot, put in your domain name, and it'll show you how you're doing and give you clues on what to improve so your business starts being found by ChatGPT, Claude, Perplexity, Gemini, and others. It's a huge shift.

Practical AI Tips: HubSpot Breeze Enrichment in Action

Brendon Dennewill: Bella, thank you so much for joining. One last question: any final words of advice you'd leave listeners with as they think about optimizing their CRM or anything else you've been sharing with clients this week?

Bella Cowdin: HubSpot has made a lot of AI enrichment natively available. Just this week, I uploaded a new segment for a client and noticed that in the import process there was a checkbox that never existed before: 'Would you like to enrich all of these records with AI?' I said yes, uploaded the records, and clicked on one of the example companies.

This client didn't have Breeze credits for that enrichment, but it worked natively regardless. It said Breeze was thinking, and then it generated a quick summary letting me know the last time that organization received funding, how many employees they have, and even suggested the CEO as a recommended associated contact.

It will give you data to fill in the missing puzzle piece on your accounts. A lot of clients will have, say, the Director of Operations and the Director of Marketing at a company, but they don't have the Director of Procurement or the actual decision-maker. With this enrichment tool, it's very possible to pull in that missing piece and get your full picture of the company.

Brendon Dennewill: Another very powerful example. I think we've talked about some really exciting things today, and I realize we're probably just scratching the surface. But I think that's enough to whet everybody's appetite for what's to come. I'm personally very excited for the next six to twelve months to unpack all the value, because it still has to be customized for every single business. Bella, thanks so much for coming on the show and sharing your insights.

Bella Cowdin: Absolutely. Thank you so much for having me, Brendon. I had fun.

Brendon Dennewill: Thanks again.

Bella Cowdin: Have a good one. 

Subscribe