GEN AI PRODUCT DESIGN
HubSpot Content Assistant: AI-Powered Writing for Marketers and business owners
The goal was clear: create something that wouldn’t replace the human touch but would instead serve as a support system—enhancing creativity, not stifling it.
Timeline
16 weeks
Background
When we launched HubSpot's Content Assistant, the goal was simple: empower marketers to streamline their content creation process, overcome writer’s block, and maintain a consistent brand voice—all within HubSpot’s ecosystem. This AI-powered writing tool was designed to help marketers produce high-quality content faster, whether for blogs, emails, or social media posts.
The results? Well, they spoke for themselves. Within three months of launch, content creation time was reduced by 30%. Users not only adopted the tool rapidly but also doubled their blog output, boosting their SEO rankings and increasing engagement across the board. It felt like we had created something truly impactful—something that actually changed the way users worked.
Key Outcomes and Results
Content Output: Users doubled their blog production, thanks to a more efficient and streamlined workflow.
SEO Boost: 60% of users saw significant improvements in their search engine rankings within 30 days.
Higher Engagement: A 25% increase in shares, comments, and time on page indicated users were connecting with the content more deeply.
User Satisfaction: The Net Promoter Score (NPS) for content creation rose by 35%, showing clear satisfaction with the tool.
From a business standpoint, we saw remarkable results too:
Reduced Churn: By replacing external tools, the Content Assistant helped retain users who otherwise might have sought alternatives.
Increased Adoption: The blog tool within HubSpot saw a 40% rise in usage, reflecting the value this tool brought to marketers.
Efficiency Gains: Teams saved up to 40% of their time on content creation, leading to faster time-to-market.
Problem Statement
Before the Content Assistant, marketers faced common struggles: writer’s block, chaotic workflows, and challenges in maintaining a consistent brand voice. These pain points made content creation a slow, inefficient process. But how could we provide a solution that would keep marketers in the flow while still giving them creative freedom?
As the Lead Designer on this project, I worked alongside product managers, engineers, and researchers to craft an intuitive tool that would address these issues. The goal was clear: create something that wouldn’t replace the human touch but would instead serve as a support system—enhancing creativity, not stifling it.
TL:DR (Post-Mortem)
The design process for HubSpot's Content Assistant was filled with constraints, the most significant being time. With a tight deadline, we had limited opportunities to conduct extensive solution testing or iterate based on user feedback. As a result, we relied heavily on feedback from stakeholders within HubSpot to inform our decisions. While stakeholder insights were invaluable for aligning with business objectives, they didn’t always reflect the real-world experiences of our end users.
In hindsight, the lack of time for user testing and iterative design was a major limitation. The process was more about working within the constraints we had, rather than having the luxury to explore and refine the product based on user input. This was a major learning point for me—while we succeeded in delivering a functional tool, I recognize the importance of deep user research and longer feedback loops to create truly user-centered solutions.
Research & Planning
To kick off the project, we conducted foundational research to better understand the challenges marketers faced when creating content. We interviewed 15 participants, diving into their frustrations, workflows, and preferences when it came to writing and ideation. In parallel, we analyzed user behavior data to identify key friction points—pinpointing where users dropped off in the process and where inefficiencies slowed them down.
Additionally, we conducted a competitive analysis, benchmarking against industry-leading AI writing tools to uncover best practices for content creation and workflow integration. This research provided us with valuable insights into user needs and expectations, shaping our early design decisions and helping us define the core features of the Content Assistant.
Business Value
Why was all this important to HubSpot? The challenges of writer’s block, disorganized workflows, and inconsistent brand voice were critical for HubSpot because they directly affected marketers' productivity, content quality, and overall engagement with the platform. These issues also had a broader impact on HubSpot’s core mission—supporting the inbound methodology by enabling users to create high-quality, engaging content that attracts and retains customers.
Design strategy
With a deep understanding of the problem, user needs, and potential business impact, we kicked off a series of workshops. One key insight from our initial research was that users were highly wary of AI—they didn’t fully trust it, especially given the prevailing narrative that AI could replace marketing jobs. This made it clear that we needed to take a strategic approach in positioning the tool not as a replacement, but as a trusted assistant—a copilot designed to enhance creativity rather than undermine it.
After several discussions and iterations, we developed key strategies to ensure that Content assistance felt like an empowering tool rather than an intrusive force:
User Control: Allowing users to decide when and how AI intervenes, ensuring they remained in control of the creative process.
Adjustable AI Settings: Enabling users to fine-tune AI involvement, whether they needed simple idea generation or full content drafting.
Transparency & Editability: Making AI actions clear by showing how content is generated and allowing users to easily edit or refine outputs to maintain their unique voice.
These strategies helped us build trust with users while reinforcing AI as a collaborator, not a replacement, ensuring that marketers felt supported rather than sidelined.
Solutions
With a deep understanding of the problem, user needs, and potential business impact, we kicked off a series of workshops. One key insight from our initial research was that users were highly wary of AI—they didn’t fully trust it, especially given the prevailing narrative that AI could replace marketing jobs. This made it clear that we needed to take a strategic approach in positioning the tool not as a replacement, but as a trusted assistant—a copilot designed to enhance creativity rather than undermine it.
After several discussions and iterations, we developed key strategies to ensure that Content assistance felt like an empowering tool rather than an intrusive force:
User Control: Allowing users to decide when and how AI intervenes, ensuring they remained in control of the creative process.
Adjustable AI Settings: Enabling users to fine-tune AI involvement, whether they needed simple idea generation or full content drafting.
Transparency & Editability: Making AI actions clear by showing how content is generated and allowing users to easily edit or refine outputs to maintain their unique voice.
These strategies helped us build trust with users while reinforcing AI as a collaborator, not a replacement, ensuring that marketers felt supported rather than sidelined.
Deeper usability testing
While we conducted initial user interviews, I would have pushed for deeper, longitudinal research to better understand how marketers interact with the tool over time. This would have helped us identify additional pain points that may have surfaced after prolonged usage.
Greater Customization
One area I feel could have been improved is the level of customization within the AI’s recommendations. Marketers often have unique voices and strategies that the AI could have better supported by offering more personalized suggestions. If I had more time, I would have focused on refining these features, allowing users to tailor the content assistant even more to their brand’s tone and messaging.
More iterations
Given more time, I would have worked through more iterations of the AI’s capabilities, testing new features like multi-language support or advanced SEO insights that could have taken the tool even further. There’s always room for improvement, and I would have ensured the tool kept evolving based on user feedback.
Longer feedback loops
In the rush to launch, our feedback loops were sometimes short. I would have liked to implement longer testing phases and gather insights over a few more months to refine the product further. Testing over a longer period would have given us a clearer picture of how users were engaging with the AI and whether it continued to meet their needs as their content strategies evolved.
In the end, launching the Content Assistant was a rewarding challenge. It not only helped users improve their content creation efficiency but also showcased the power of AI when designed to enhance human creativity. However, reflecting on the project with the benefit of hindsight, there are always areas for growth. Given the constraints we faced, we accomplished a lot—but I know that with more time and resources, I could have further expanded the impact of the tool, making it even more personalized and adaptable for users.
Through this project, I gained invaluable experience in balancing user needs with business objectives, and it reinforced the importance of designing AI tools that empower users to do their best work, without compromising their creativity or voice.