Automation is evolving, and AI agents represent the next big leap. However, building a product that truly resonates with users is still an unproven and challenging process.
In a recent episode of the Startup Project podcast, I had the opportunity to dive into this topic with Jacob Bank, Founder and CEO of Relay.App. Our discussion shed light on the intricate product development journey of an AI agent-building platform. Here are some key takeaways for product managers, engineers, and entrepreneurs looking to build AI-first automation products.
The path to product clarity and the power of iteration
Relay.App’s journey highlights the power of iteration and market feedback. The early days were a period of exploration. Founded in 2021 with the vision of enhancing cross-tool coordination using AI, the team experimented with multiple approaches—building eight or nine different prototypes—before landing on a viable product direction.
Initially, the idea was to use AI to bridge gaps between various tools as the number of SaaS tools in any given organisation have been increasing over the last decade. But the real breakthrough came when Relay App shifted its focus to capturing repeated tasks that blend automation with human judgment. This pivot led to the development of a workflow tool that sits between Zapier-style automation and Asana-style task management.
The shift to AI agents
Despite some early traction, the Relay App team recognised a fundamental issue: positioning themselves as a workflow automation tool limited their reach. The “no-code workflow automation” label resonated with a niche audience but failed to capture the broader potential of AI-driven productivity.
This realisation led to a strategic shift—from an AI-powered automation tool to an AI agent-building platform. More than just a rebranding exercise, this represented a fundamental change in product philosophy. Instead of simply connecting tools, Relay.App now provides a platform where users can create intelligent agents that proactively work on their behalf.
Also Read: 4 ways to eliminate pointless tasks from your daily work
Integrations: A core competency, not an afterthought
One of the key takeaways from our discussion was the critical importance of integrations. Unlike other companies, Relay is not considering integrations as an afterthought and are not interested in open sourcing the ability to build integrations for the platform. They believe each integration has to carefully crafted and designed by top-tier engineering talent. For AI agents to function effectively, they must seamlessly interact with the tools that businesses already use.
Relay.App currently supports around 120 native integrations and is working toward expanding this number to 300-500. The goal is to cover essential business categories like email, calendar, messaging, CRM, and marketing automation. They firmly believe the usefulness of AI agents is directly tied to their ability to integrate deeply with existing workflows.
Human-in-the-loop
As AI becomes a bigger part of workflows, maintaining human oversight is crucial. Jacob stressed the importance of a human-in-the-loop mechanism that allows users to review and provide feedback on an agent’s planned actions before execution.
This approach not only builds trust but also enables continuous learning. AI agents can refine their behaviour based on user input, ensuring they align with human intent. Additionally, when AI deviates, users must have the ability to intervene and correct course in real-time. Striking the right balance between delegation and human interaction is essential to making AI augmentation truly effective.
Product-led content and community-driven growth
Relay’s go-to-market strategy heavily revolves around product-led content. Jacob actively creates content—LinkedIn posts, YouTube tutorials—demonstrating real-world use cases for AI agents. This helps users understand the product’s capabilities and lowers the barrier to adoption.
Beyond content, fostering a community is a key part of Relay.App’s growth. By enabling users to create and share templates, the platform has built a cycle of organic adoption where users contribute back to the ecosystem, making the product more valuable for everyone.
Also Read: From silicon to sustainability: Data centres in a warming world
The future of product development: AI-powered teams
Jacob envisions a future where product teams are leaner, more agile, and empowered by AI. Instead of large teams with highly specialised roles, individuals will take on “player-coach” responsibilities—combining strategic vision with hands-on execution.
This shift is possible because AI agents can automate routine tasks, allowing human employees to focus on higher-level problem-solving. The key lies in identifying the right tasks for automation and designing workflows that integrate AI seamlessly with human expertise.
Key lessons and the road ahead
Jacob Bank’s journey with Relay.App offers valuable lessons for anyone building AI-first products:
- Embrace iteration: Finding the right product-market fit requires constant experimentation.
- Listen to customers: Market feedback is essential in refining product vision.
- Prioritise integrations: AI agents must work seamlessly within existing tool ecosystems.
- Maintain human oversight: A human-in-the-loop model builds trust and ensures better AI alignment.
- Leverage product-led content: Educating users through compelling content drives organic growth.
As the agent landscape evolves, product teams need to remain flexible and adaptable. By focusing on tangible value, robust integrations, and user-centric design, companies can successfully navigate the challenges of AI adoption and build transformative products.
Relay.App’s experience serves as a reminder: the real opportunity in AI isn’t in the hype—it’s in creating practical, user-friendly solutions that empower people to work smarter and faster.
—
Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic.
Join us on Instagram, Facebook, X, and LinkedIn to stay connected.
Image courtesy of the author.
The post The product management strategy behind building AI agent platform appeared first on e27.