The research tells us that almost half of marketing leaders have invested in some form of AI tool. Meanwhile, sales reps are catching up. In working with client sales teams at Salted Stone, we’ve seen adoption rise from 24% in 2023 to 43% last year — and we expect that to grow even more. But plenty of others haven’t got to work yet.
Knowing some businesses struggle to keep pace, Salted Stone explored this dilemma on the Under The Hood podcast with HubSpot guest expert Nicholas Holland. We unpacked the evolving relationship between B2B brands and AI integration, HubSpot's efforts to empower team adoption, and where AI agents and avatars can help.
Read or watch the episode with Nicholas below. Alternatively, listen in on Spotify, SoundCloud, or iHeart.
For those businesses that haven’t taken up the opportunity to fully harness AI, Nicholas says analysis paralysis is to blame. 54% of marketers reportedly feel overwhelmed at the prospect of implementing AI tools, according to new HubSpot research. Many leaders don’t know where to start or what to do first — and that's understandable. “You do have to have a degree of courage to jump off that cliff,” says Nicholas, in order to start using AI meaningfully.
The second problem points to AI hype. Often teams buy shiny new tools that don’t deliver because strategy isn’t guiding the selection and usage of those tools. “Across one department you might have bought four different tools that all do somewhat of the same task, so you see an explosion of cost that can't be really turned into ROI,” says Nicholas. Just buying a ChatGPT plan isn’t enough — you need a strategy for deployment and transformation.
To this he adds a third problem: Once people get serious about AI, they soon realize importing data into AI systems can be complex and messy. (We explored this dilemma in our recent article on HubSpot's Deep Research Connector, where we dive into data hygiene and AI.) These three common challenges breed skepticism towards AI, which can hamper adoption efforts in an organization.
However, keeping up with advancements in AI is critical. A move must be made, no matter how difficult getting started may be. Either leaders sit on the sidelines and watch everybody else figure it out, or they can step up to the plate and learn at least one tool. Nicholas advocates for the latter. Competition is fierce, he says, citing an oft-quoted saying in the start-up realm. If you don't keep up, it's going to be at a detriment to your business:
“Out there in some garage, an entrepreneur is forging a bullet with your company’s name on it. Once that bullet leaves the barrel, you won’t be able to dodge it. You’ve got one option: you have to shoot first.”
— Gary Hamel (HBR, 1999), American management consultant
Which brings us to how HubSpot empowers marketing, sales and service to take that crucial first shot.
HubSpot has always been a champion for small-and-medium businesses (SMBs), for the teams who don’t have budgets to waste on moonshots that don’t work out. In order to bring them value fast, HubSpot has taken AI and applied it to the real problems customers have. Though, the journey has taken years.
When Nicholas’s team started working with AI about three years ago, it was in exploratory mode. They tested models like OpenAI’s GPT-3 and integrated them into the tools. At that stage, it was “neat,” but far from being robust enough to bring immediate value.
Two years ago, the models started to improve and HubSpot got serious about AI, adding more capabilities to its arsenal. But the quality still wasn’t there. The images looked odd, and the text was robotic and impersonal. Everything changed in 2024.
Nicholas explains that if AI was going to help SMBs, it had to be necessary. That meant formulating strong opinions on what an AI-first SMB looks like, then building tools able to deliver meaningful outcomes.
Here’s the result:
The idea is that all three sets of features are powered by a unified data layer, both structured (like CRM fields) and unstructured (emails, calls, chats, and so on). When you harness that, Nicholas believes any SMB can become an AI-first company in a remarkably short amount of time.
AI’s capabilities are vast, yet people are still required — no matter what other AI pundits say. Many AI vendors today target large enterprises, focusing on ROI, cost-cutting, and reducing headcount. But that’s not the HubSpot approach. Instead, it's about empowerment.
“Technology unto itself is not inspiring. It's what technology does for humans that really makes it inspiring in my opinion,” explains Nicholas. “Everything we do [at HubSpot] puts the human at the center.
“We want employees to have AI superpowers, that’s why we give them AI assistants. We want teams to gain the leverage they need, not because we’re trying to get rid of people, but because there’s so much work to do!
“It's very much about empowering the marketing, sales, and service people we work with,” says Nicholas. “That’s who we think are the real winners here.”
We can now delegate mundane tasks that used to take up our time, without forgoing complete control. Humans must be involved the whole way through. That begins with setting guardrails around responsible AI use. While crucial, empowering teams also means providing the flexibility to automate repetitive tasks and deepen contributors’ strategic thinking, creativity, and understanding of who they serve.
So where should one start? Nicholas recommends not worrying about finding the perfect do-it-all tool. Pick one, knowing that you may not be tied to it for years, and focus on learning by doing.
“You learn how to interact with AI, you learn how to bring data or context to AI, you learn that the hype of it ‘doing all the work’ is kind of smoke and mirrors,” says Nicholas. “You have to put a different type of work into making the AI outputs really good, and those are all durable learnings.
“So when you switch from that tool to another tool, you take those [learnings with you].”
Emerging toolsets that excite us at Salted Stone are AI avatars and AI agents, which offer smarter ways of getting infinite work done. Essentially, AI agents automate tasks and are best used for problem solving, whereas AI avatars deliver instructions where a human touch helps.
| AI Agents | AI Avatars |
|
|
| Example use case: Automatically applying a brand’s voice to a body of text in HubSpot at the click of a button. | Example use case: Delivering a standardized sales training to a new remote hire as part of an onboarding program. |
Of course, HubSpot offers a host of integrated agents suitable for all teams: Breeze Customer Agent, Breeze Prospecting Agent, Breeze Content Agent, and Breeze Social Media Agent. For a high-level overview, read this article on Breeze AI.
A widely quoted statement on AI adoption, attributed to content creator Lucas Rizzotto, is: "AI won't replace humans, but humans using AI will replace humans who don't." We have to agree. AI integration is rapidly becoming essential, rather than a competitive advantage, and the sooner teams embed AI-augmented ways of working, the better.
To begin, take Nicholas’s advice and experiment with whatever you have at your disposal, carrying the learnings forward into the next tool or iteration of your choosing. And if you have already invested into HubSpot's AI-powered customer platform, make it your first sandbox. You might be surprised at just how far it takes you.
Beyond experimentation and ready to integrate? Talk to us about incorporating AI into your HubSpot strategy.