Dive Club 🤿
Episode: Emmet Connolly - Transitioning into the Next Era of Design
Host: Ridd
Release Date: June 27, 2025
Introduction
In this episode of Dive Club, host Ridd engages in an insightful conversation with Emmet Connolly, the Vice President of Design at Intercom. The discussion delves into Intercom's journey towards becoming an AI-first company, the inception and evolution of their groundbreaking AI product, Fin, and the transformative impact of AI on the design landscape. Emmet shares his experiences, challenges, and visions for the future of design within a rapidly evolving technological environment.
Origin Story of Fin
Emmet begins by recounting the genesis of Fin, Intercom’s AI-powered customer service agent. He attributes the accelerated development and success of Fin to the emergence of ChatGPT, which acted as a catalyst for integrating advanced AI into their systems.
Emmet Connolly [01:29]:
"When ChatGPT arrived, it was a dawning realization over the weekend that this was going to signify a real big change for us."
Leveraging an existing partnership with OpenAI, Intercom swiftly integrated ChatGPT's API, releasing initial AI features within weeks. These early implementations, such as response rephrasing, allowed the team to experiment and understand the potential of AI in enhancing customer interactions.
Emmet Connolly [01:29]:
"Within three to four weeks of ChatGPT coming out, we released a bunch of AI features... helping us get our beaks wet and realize what was possible."
The launch of Fin coincided with the release of GPT-4, marking a significant milestone. Over the years, Fin has grown beyond Intercom’s core business, positioning itself as a formidable entity in the AI-driven customer service space.
Emmet Connolly [01:29]:
"Fin is bigger than Intercom. We've created a true AI-native startup emerging from a legacy SaaS business."
Transitioning to an AI-First Company
Emmet discusses the strategic shift Intercom undertook to prioritize AI, emphasizing the importance of adaptability and proactive disruption.
Emmet Connolly [00:29]:
"The start of an S curve is a great time to be a designer because there is so much to be figured out."
Recognizing the transformative potential of AI, Intercom centralized its design and R&D teams to focus on AI-driven projects like Fin. This reorganization allowed the company to quickly adapt and innovate without being hindered by traditional structures.
Design Challenges and Iterations
One of the primary challenges Emmet highlights is enhancing Fin’s resolution rate—the percentage of customer queries successfully handled by the AI agent. Initially, Fin achieved a 24% resolution rate out of the box, a promising start that required continuous improvement.
Emmet Connolly [08:06]:
"Our average resolution rate has gone from 24% to 60% today."
To address this, Intercom developed a comprehensive application for training Fin, involving processes like analyzing customer interactions, training the AI, testing its responses, and deploying updates. This iterative cycle was crucial in refining Fin’s capabilities and ensuring its effectiveness across diverse customer scenarios.
Emmet Connolly [08:06]:
"We had to design this analyze, train, test, deploy workflow to effectively train Fin."
Evolving the Role of Designers
Transitioning to an AI-first approach necessitated a redefinition of the design team’s roles and responsibilities. Emmet introduced the concept of specialist AI designers—designers deeply embedded within the AI team, collaborating closely with machine learning scientists and researchers.
Emmet Connolly [18:59]:
"We created a specialist AI designer role, working directly with our ML scientists on model prompting and output analysis."
Additionally, Intercom encouraged all product designers to engage with coding, fostering a culture of experimentation and technical proficiency. This initiative aimed to bridge the gap between design and engineering, enabling designers to have a more hands-on role in product development.
Emmet Connolly [18:59]:
"Everyone on the team has taken a goal to ship code to production... It's all about experimenting and giving it a go."
Design Team Reorganization
To effectively lead the AI transformation, Intercom centralized its design and R&D efforts, organizing teams around specific workstreams. This structure allowed for greater focus and agility, enabling teams to rapidly respond to evolving project requirements.
Emmet Connolly [19:22]:
"We reorganized design to be centralized, organizing around workstreams that can spin up or down dynamically."
This reorganization introduced a cross-functional team dynamic, incorporating members from sales and marketing into design projects. While this led to some initial instability and ambiguity, it ultimately facilitated more cohesive and integrated project execution.
Skills for Future Designers
Emmet emphasizes the importance of being a self-led learner and having a willingness to experiment as crucial traits for designers in the AI-driven era.
Emmet Connolly [42:16]:
"Being a self-led learner is probably the most important characteristic... curiosity, drive, determination, and agency."
He advocates for designers to embrace both design and coding skills, arguing that technical proficiency enhances creative capabilities and enables designers to more effectively engage with their craft.
Emmet Connolly [42:16]:
"The weapon designers of the future are the ones that can marry both design and coding skills."
Redefining Product Development Process
The traditional linear approach to product development—starting with job definitions, wireframing, and high-fidelity prototypes—was unsuitable for AI-first products like Fin. Instead, Intercom adopted a more flexible, iterative process that prioritized material exploration and data analysis before formalizing the user interface.
Emmet Connolly [46:47]:
"For our AI-infused product, we couldn’t follow a linear approach. We had to build up the backend first before designing the UI."
This approach involved close collaboration with customers to co-create preliminary prototypes, allowing for real-time adjustments based on feedback and data insights. Once the foundational elements were validated, the team could proceed to detailed UI design and styling.
Emmet Connolly [46:47]:
"We worked with our ML scientists and customers to massaging the data... then designed the UI once we knew the shape of the solution."
The Role of Design Research
With the shift towards AI-first products, the nature of design research has evolved. Emmet notes that much of their research has focused on strategic and foundational aspects, such as industry trends and customer needs, rather than traditional user studies.
Emmet Connolly [51:24]:
"Our research has been oriented towards foundational aspects to inform our strategy rather than extensive user studies."
He anticipates that as AI-driven products mature, automated tools and AI itself may play a larger role in streamlining user research and testing, potentially transforming the methodologies designers employ.
Emmet Connolly [51:24]:
"It’ll be interesting to see if AI-driven products can manage user research aspects like automated testing and feedback analysis."
Hiring and Team Dynamics
Emmet outlines the qualities Intercom seeks when expanding their design team in an AI-centric environment. Alignment with the company’s AI-first philosophy and a demonstrated willingness to innovate are paramount.
Emmet Connolly [53:19]:
"We want new team members to be philosophically aligned with our AI-first approach and to have evidence of their proactive engagement with AI technologies."
Additional desired traits include curiosity, determination, and an openness to cross-disciplinary collaboration, ensuring that team members can thrive in a dynamic and rapidly changing landscape.
Emmet Connolly [53:19]:
"Characteristics like curiosity, drive, determination, and low fear index are crucial for navigating the complexities of AI-driven design."
Final Thoughts and Future Outlook
Emmet expresses optimism about the future of design, highlighting the potential for AI to usher in a golden age for the discipline. He envisions a landscape where designers wield both creative and technical tools, enabling them to craft more sophisticated and impactful user experiences.
Emmet Connolly [61:50]:
"Designers who embrace technical tools and AI will lead the next golden age of design, creating more impactful and sophisticated user experiences."
He underscores the importance of adaptability and continuous learning, encouraging designers to experiment and remain open to evolving methodologies and technologies.
Emmet Connolly [62:42]:
"We'll definitely get there together. Staying the same was never a viable long-term strategy for designers."
Emmet concludes by reaffirming his commitment to innovation and improvement within the design field, emphasizing that embracing change is essential for sustained relevance and success.
Emmet Connolly [62:42]:
"I'm excited to fix a bunch of stuff and move onwards, driving design into this new era with enthusiasm and determination."
Key Takeaways
-
Proactive Adaptation: Embracing AI-driven tools and methodologies is crucial for staying ahead in the evolving design landscape.
-
Iterative Development: Continuous experimentation and iterative improvements are essential for refining AI products like Fin.
-
Redefining Roles: The integration of technical skills, particularly coding, alongside traditional design capabilities will be increasingly important.
-
Centralized Teams: Organizing design and R&D around dynamic workstreams enhances agility and focus in AI-centric projects.
-
Future Skills: Curiosity, self-led learning, and adaptability are vital traits for future designers navigating AI advancements.
-
Optimistic Outlook: AI presents an opportunity for designers to achieve greater impact and innovation, heralding a potential golden age for the discipline.
Notable Quotes
-
Emmet Connolly [01:29]:
"When ChatGPT arrived, it was a dawning realization over the weekend that this was going to signify a real big change for us." -
Emmet Connolly [08:06]:
"Our average resolution rate has gone from 24% to 60% today." -
Emmet Connolly [19:22]:
"We reorganized design to be centralized, organizing around workstreams that can spin up or down dynamically." -
Emmet Connolly [42:16]:
"Being a self-led learner is probably the most important characteristic... curiosity, drive, determination, and agency." -
Emmet Connolly [53:19]:
"We want new team members to be philosophically aligned with our AI-first approach and to have evidence of their proactive engagement with AI technologies." -
Emmet Connolly [62:42]:
"I'm excited to fix a bunch of stuff and move onwards, driving design into this new era with enthusiasm and determination."
This episode offers a comprehensive exploration of how AI is reshaping the role of designers and the structural dynamics within design teams. Emmet Connolly’s experiences at Intercom provide valuable insights into navigating and thriving in an AI-driven design future, emphasizing the importance of adaptability, continuous learning, and proactive innovation.
