When Beth Dunn looks back on her career, there's a single thread that ties it all together. Whether she was writing onboarding guides, building HubSpot’s first content design team, or shaping the future of product experience at Agent.ai, her work has always been about making technology feel more human.
Beth is quick to admit that she never set out to work in artificial intelligence. Her path into agents was not the result of a strategic five year plan, but rather a willingness to experiment, to follow her curiosity, and to automate herself out of repetitive work. That willingness eventually gave rise to what she calls the Sweet Spot Framework, a way for all of us to think about how AI can amplify what we love doing most.
Beth’s career began at HubSpot in 2010. At the time, “UX writer” was not even a common job title. She was officially in customer onboarding, but quickly noticed a problem. Every time she explained the product to a new customer, she was forced to say the same thing: “If the words on the screen were clearer, I wouldn’t have to explain this.”
So she started pestering the engineers. If they could change the words on the interface, maybe customers would need fewer explanations. Soon, she became the de facto writer for the product, and eventually the company’s first UX writer.
As HubSpot scaled, the demands on her time grew. She couldn't edit every screen or review every string of text that went into the product. That's when she built an internal tool that colleagues nicknamed BethBot. It wasn't AI. It was a simple rule-based program that flagged typos and enforced HubSpot’s brand voice.
But BethBot had a surprising effect. By automating the repetitive edits that drained her time, Beth freed herself to focus on the more strategic and creative parts of her role. She jokes that she “automated herself out of a job,” but what really happened is that she created space for a much more interesting one.
This was the first glimpse of a principle that now guides her work at Agent.ai: AI doesn't need to replace people. It can remove the drudgery so we can lean more fully into what we do best.
Fast forward more than a decade. AI is everywhere. Beth had long since moved into leadership roles and was watching closely as new technologies emerged. She subscribed to Agent.ai founder Dharmesh Shah’s newsletter, curious but not yet sure how AI would fit into her own work.
When she saw an early note about building an agent that could edit web copy, it struck a chord. Could she make an AI-powered version of BethBot? That thought pulled her into experimenting with agents, and she quickly realized the broader potential.
At INBOUND 2025 she shared the culmination of that realization: the Sweet Spot Framework.
The Sweet Spot is designed for anyone who has heard the message “you need to use AI” but does not know where to start. Rather than beginning with a list of tools or features, it starts with a simple question: what do you love doing most?
Beth explains that every creative process can be broken into four stages: Investigate, Dream, Explore, Act. Some people love the upfront research, others thrive on brainstorming wild ideas, some are at their best during testing and iteration, and others shine when shipping the final product.
The Sweet Spot quiz asks just nine questions. Based on your responses, it generates an archetype that reflects your natural strengths. Maybe you are a Professor, drawn to early research and strategy. Maybe you are a Producer, who loves shipping and execution. Or maybe you are a Scientist, combining exploration and investigation.
The point is not to box you in. The point is to show you where you can lean in, and where an agent could help you cover the gaps. If you love dreaming and brainstorming but hate proofreading, an agent can take the polish off your plate. If you thrive on research but get bogged down in formatting, an agent can step in.
By mapping your creative flow, you find your Sweet Spot. And when you let agents handle the rest, you make more room for the work that energizes you.
Beth and her colleagues did not stop with the quiz. To demonstrate what this looks like in practice, they built a set of sidekick agents aligned to each stage of the creative process.
Each one has a specific purpose. They're not meant to replace the entire creative process. Instead, they show how agents can take on focused roles that make the human parts of the process more enjoyable and effective.
Beth points out that this is often where new builders find traction. They don't always need to invent something brand new. They can take inspiration from existing agents, adapt them to their needs, and make them more personal.
For users, the lesson is simple. You don't need to be a coder to benefit from agents. Start with something that solves a problem you recognize in your own work, and then customize from there.
Another theme Beth emphasizes is the changing role of generalists.
In the old model, many professionals advanced by becoming increasingly specialized. They sharpened one skill, became an expert, and built a career on that expertise. Today, some specialists fear that AI will replace their niche.
Beth sees a different opportunity. Generalists who can connect across domains are thriving. Agents can handle deep, narrow tasks like cleaning up text, analyzing data, or generating code snippets. The person who knows how to orchestrate those outputs and bring them together in a meaningful way becomes incredibly valuable.
She describes it as a shift in perspective. Instead of asking “Will AI take my job?” the better question is “Which tasks can AI take so I can focus on higher value work?” In her words, “Agents are everybody’s opportunity to take a piece of that cloud that is hovering over all of us that is AI, and pull it down to earth and make it work for them.”
This is the heart of Beth’s message. AI is not just for builders or engineers. It's for anyone who wants to reclaim their time, lean into their strengths, and do more of the work they love.
Beth also looks ahead to what might come next. She imagines agents that evolve over time, learning from your preferences, and suggesting changes to their own prompts. If you consistently ask an agent to make outputs more playful, it might eventually propose, “Do you want me to adjust my default style?”
She envisions teams of agents with a “team lead” agent that notices gaps in your workflow and offers to fill them. While this is not reality yet, the seeds of it are already here. For Beth, this future feels like a natural extension of the BethBot experiment that started it all.
Beth ends her conversations with a piece of advice that is both practical and philosophical: lean into the weird.
This moment in technology is unusual. It's uncertain and sometimes unsettling. But it's also full of possibility. Rather than resisting it, Beth encourages us to meet it with curiosity and optimism.
For users of agents, the invitation is simple. Don't overthink your first step. You don't need to build the perfect agent on day one. Instead, try something small. Take the Sweet Spot quiz, see your archetype, and explore the sidekick agents that fit your style.
You might just find, like Beth did, that automating away the work you like least creates space for a far more interesting job.
Beth Dunn’s journey from content design to AI agents is a reminder that the best innovations often begin with a simple problem and a willingness to play. By identifying what you love most, automating the rest, and staying curious, you can make agents a partner in your growth.
Ready to discover your AI Sweet Spot? Take the Sweet Spot Archetype Quiz on Agent.ai and see which archetype fits you best.