We just wrapped up two incredible days at the reIGNITE 2026 conference in Muskoka. I attended with Kevin Butler, our Chief Strategy Officer, and Chris O’Neil, our VP of Revenue and Customer Growth and we came back with a wealth of new insights, and we’re still busy unpacking everything we learned. Between a great kickoff from the District Municipality of Muskoka and the stunning views, the real value for me was in the conversations. At the end of the day, growth comes from relationships, not slide decks.
Later on, I had the opportunity to take the stage and share one of the tougher calls we’ve made at Goose Digital this year: walking away from Salesforce after eight years to build our own CRM from scratch.
The Problem We Were Solving
At any given time, we’re managing around 125 open sales opportunities; translating to 350 to 400 a year across new logos and existing clients, with roughly 150 closed. That entire pipeline lived in Salesforce, our marketing automation ran through HubSpot, project management sat in Notion, and time-tracking lived in yet another system. That meant four disconnected platforms, endless duplicate data entry, and zero intelligence connecting them. Not to mention, limitations around connecting a knowledge base, HR tools, and using AI to help streamline our workflows. And, we are trying to double the company in 3 years…
For a marketing agency that builds revenue operations and strategy for clients, that disconnect was impossible to justify internally. My challenge as CEO was twofold: how do we actually put AI to work in the business, and how do we get our own house in order at the same time.
What We Built
The answer was two connected pieces. AgencyDimensions is the CRM and order-management system we built in-house; the record of every account, contact, and opportunity. GD Command is the AI orchestration layer sitting on top of it, replacing a string of small, individually vibe-coded tools we’d built over the past two years for things like on-brand image generation, content planning, and landing page execution.
Going from first commit to live production took just 16 weeks; 701 commits, about 80 schema changes, all reviewed by the team. We migrated roughly 3,100 accounts, 4,000 contacts, and 3,700 sales opportunities, deliberately leaving behind closed-lost history that wasn’t tied to a real prospect or customer relationship.
Behind the scenes, the workflow was a true AI partnership: Claude Cowork served as our planning and spec-writing layer, driving deep architectural discussions before a single line of code was written, while Claude Code handled implementation and review.
Why It Mattered Enough To Do
Anyone who’s run a sales team knows the real problem isn’t the CRM software, it’s getting reps to keep it current. So we built the friction out. The system reads meeting data straight from Google Calendar and the Google Meet summaries already being generated, attaches it to the right account, brings in any new contacts who were on the call, and recommends the next pipeline stage based on what actually happened in the meeting. If it’s a client project meeting, the same summary generates tasks in our delivery system automatically. Our reps don’t have to spend hours doing manual cleanup after every call anymore; and when our CSO, Kevin Butler, leads pipeline reviews, we’re working off data we can trust.
We also rebuilt our reporting environment exactly the way our business needs to see it; new versus existing sales, MRR versus one-time project revenue, rep and account performance; instead of stitching it together from exports every week. And we added account enrichment, so a new account created in the system gets automatically researched rather than sitting incomplete.
The cost comparison is worth sharing too: we’re running the whole platform for about $500 a month in cloud infrastructure, against roughly $70,000 a year we’d been spending across our previous tool stack.
Where It’s Going
We call the assistant inside the platform ADAI; it’s Gemini under the hood, scoped to specific tools and functions inside AgencyDimensions. The longer vision is a single conversational interface into everything the business runs on, with the database as the source of truth underneath it. UIs get better every year, but I think the interface most of us actually work through keeps shrinking down to a chat window. We’re not there yet, but that’s the intended direction.
The Bigger Lessons
Day two of reIGNITE stuck with me more than the technical build did. Mark MacLeod’s framing that a CEO’s job is to remove friction to create revenue growth is exactly right, and it applies as much to your own internal systems as it does to your go-to-market. Get out and meet people. Enjoy the process. And make space to reflect on your health and your family, not just the business; because none of the platform work matters if that side falls apart.
Thanks to everyone who came out to the session and asked sharp questions, including the honest conversation around meeting transcription and compliance-sensitive industries, which deserves its own deep dive at some point.
Kevin Butler, also spoke at RevOps Camp this year on how businesses should map their sales process before trying to accelerate growth through it. He’s writing that up now, so keep an eye out for it.
If you want to see more of what we’re building at the intersection of AI and marketing operations, it’s all at goosedigital.ai.
— Michael Turcsanyi, CEO, Goose Digital