Moments of impact from Nelson Joyce's career
In October 2013, HubSpot's Chief Product Officer David Cancel asked me and Andy Cook to create "startup within a startup" after HubSpot acquired Andy's company, Rentabilities. He gave us just three guardrails: build something that helps small businesses do marketing make it free or cheap use orange in the logo. The rest was up to us. Over the next two years we built Leadin, a free lead-generation and CRM Wordpress plugin that tens of thousands of people came to genuinely love. I worked closely with an engineering team based in Dublin and it taught me product and design skills I couldn't have gotten anywhere else. Leadin was announced on stage in front of thousands of people at HubSpot's Inbound conference in 2015 by Dharmesh Shah, one of the founders. Leadin grew to over 100,000 plugin installs and eventually transitioned to become "HubSpot Marketing Free", HubSpot's first freemium marketing product. Leadin was the capstone of that HubSpot chapter that gave me th confidence to leave and co-found my own company, Tettra, in 2015.




Our AI Slack bot, Kai, had become a critical bottleneck in our sales process. We weren't closing any new customers because prospects testing the bot during trials were getting too many unanswered questions, prompting them to choose competitor tools instead. With customer conversions at a standstill, improving Kai's answer rate became an urgent priority. I worked closely with our sales team and prospects to identify exactly which questions weren't being answered. Through this investigation, I discovered the root cause: we had recently migrated our search infrastructure from Algolia to Typesense, and the new platform wasn't properly tuned for our AI bot's needs. I led the effort to heavily tweak the search configuration, refine the RAG (Retrieval-Augmented Generation) prompt, and optimize the prompt context to better serve our users' queries. The results were immediate and dramatic. Kai's answer rate more than doubled from approximately 30% to over 60%, and new customer conversions resumed. This work directly unblocked our sales pipeline and restored confidence in our product during the critical trial phase when prospects evaluate whether our solution meets their needs.
At Trove, new users creating their first artifact went through a flow that generated one question at a time, form-style. It felt slow and clunky, and a lot of people never finished. Only about half of users going through onboarding actually completed creating an artifact, which meant we were losing half our new users before they got any real value out of the product. I dug into the onboarding data to understand where and why people were dropping off, then built a prototype of a new flow designed to keep people moving instead of stalling out question by question. I owned the redesign end to end as Head of Product, writing the code myself with Claude Code rather than handing it off, and moved away from the single-question, sequential form toward a faster, more fluid interaction. James reviewed the work and helped get it deployed. The new flow addressed the core drop-off problem in onboarding, replacing a slow form-style process with one built to get more new users through to a completed artifact on their first attempt.
Cofounded the company, grew the team, scaled the product to 10s of thousands of users.