We Outsourced the Legwork, Not the Judgment

What I took away from Zero's Future of GTM dinner in NYC

Last night I had a seat at Zero's Future of GTM dinner in NYC. Any time you pair good food and good people, the Introzy team is happy and grateful to be there. It was awesome to be in a private room of go-to-market leaders, from early-stage startups to teams with full marketing and sales orgs behind them. Different sizes, different motions, but the conversation kept coming back to the same challenge we're all facing.

Every GTM team is now building for at least two audiences. There are the buyers, who still make the decision and own the outcome. And there are the agents, the AI tools reading our site, summarizing our category, and deciding whether we show up at all for our buyers.

Every stage of growth brings its own problems, and ours will change as revenue grows. But one question stays the same at every stage: how do we build our brand for both audiences without breaking our process, outgrowing our team, or losing our authenticity? My answer to that challenge is sitting on a sticky note on my desk right now: "Outsource the legwork, but never the judgment."

The surprise: our team is more unusual than I thought

No two org charts at that table looked alike, and hearing how other teams were structured made me see our own setup more clearly. At Introzy, GTM isn't a department. It's the whole company. Granted, we're small, so everyone has to lean in. I own the growth strategy, and within that, our sales, marketing, and customer success flows. But the engineers who build a feature are the same people who build its landing page at launch, document it in our help center, and help shape the content we use to market it.

I knew that was different. I didn't realize how different until I heard what the other teams around the table looked like. Many had dedicated marketing teams, sales teams, or both. We're product-led, so some of our difference is by design. Still, it made me reflect on how we get enterprise-level work done without a traditional GTM org.

Being small is part of it, but flattening GTM was a choice. Our engineers are the subject matter experts on what we build, and they're genuinely passionate about solving our growth problems. So our development process isn't just about building the product. It's about building the brand and buyer journey experience that wraps around the application. Keeping that work with the people who built the product keeps its story authentic.

Having engineers living in GTM motions daily is a strength, but it takes a lot of coordination to move quickly on market opportunities and allocate our limited resources appropriately. One person has to read the signals and turn them into a strategy everyone can build against, and at Introzy that's me. But it's unrealistic to imagine a world where we ingest all the market data and manually spit out an updated strategy. This is where our GTM AI strategy kicks into gear.

Our strategy is to outsource the legwork of collecting and aggregating every market data point we have, but never pass off the judgment of what to do with them to an agent. That judgment is where our brand's creative decision-making and authenticity lives. Ultimately, we're not outsourcing the rationale behind what we build. We're outsourcing the work of organizing the inputs that inform those decisions.

How a small team matches enterprise output

A few weeks ago, our CEO, Zac, gave Evan, one of our engineers, a simple brief: figure out how to make my life easier. As Introzy's only dedicated GTM person, I had no shortage of ideas. Over a long lunch, Evan and I mapped the manual work slowing me down and imagined how the process would work if we could remove those constraints entirely. That conversation became the starting point for our GTM agent and Claude plugin.

So here's where we are today. We're three weeks into the agent build; so far, it focuses on two problems: understanding the health of what we've already shipped to the market, and predicting what we need to add to improve our place in the market in general. Those are big jobs and will likely be real roles on our team as we grow. The goal of building this agent isn't to replace functions we'll need as revenue grows. It's to build a foundation of good decision-making now, using every data point we have. At scale, our GTM team will inherit a documented history of market decisions grounded in data, not the gut calls of a small team in its early days.

Under the hood, the agent does the sourcing I used to do by hand and aggregates it into a morning report. It pulls our site analytics from Omni, social performance from Metricool, and campaign history from Notion. It uses Firecrawl to check our pages on mobile and desktop. It reads buyer behavior signals from Zero, our CRM, where our customer data lives. Then it ranks what's working, what's leaking, and what's untapped, and hands me a Notion page with the research and one to three campaign ideas. Everything it finds to fix becomes experiment tickets in Linear for engineering to action on that day.

And this is only the beginning. We have big plans for our growth agents at Introzy and we can't wait to share more.

Sitting with other GTM leaders at Zero's dinner made this work feel even more relevant. Around the table, some of the larger teams described dedicating whole resources to using AI to pull data and build custom landing pages for each client. It's the same idea Evan and I have been building toward, shaped for a different sales motion. Hearing a much bigger team describe the same problem told me we're working on the right one.

What we don't outsource

Here's the line we support: the agent gathers the data but does not make the final call.

We're very aware of the AI slop risk and I'm sure at times we'll be less on brand than I'd like. When every team has access to the same tools, the output starts to look the same. Generic copy, generic visuals, the same five hooks. Agents skim past sameness, and buyers scroll past it. In our build, we outlined early what guardrails we needed so we don't let slop slip in the door.

1. It checks every source before it analyzes anything. Each run opens with a readiness receipt for Omni, Metricool, Notion, and Firecrawl. If a source is missing, stale, or failed, it says so. It never invents data or quietly swaps one source for another.
2. Every claim carries its receipt. Each statistic, price, or product claim lists where it came from and when it was verified. No undated stats.
3. It recommends. It doesn't decide. Each run proposes one primary 10 to 14 day campaign and no more than two alternatives, then lists the decisions only I can make.
4. Everything it creates starts as a draft. Campaigns land in Notion at Planning. Asset briefs land at Draft and open with the line "Draft asset brief. Not approved to ship." Experiments stop at Drafted.
5. It can't ship anything. It never approves, schedules, publishes, deletes, launches experiments, or touches ad budgets. It never checks Brand Reviewed. That box is mine.
6. Our brand rules ride along in every brief. No em dashes, no exclamation marks, no named competitors, and current pricing only. Small rules, but they're some of the fastest tells that something was machine-written.

So what does outsourcing the legwork, but never the judgment, actually give us? I get back the hours I used to spend stitching reports together, and I spend them on taste: the creative calls that make Introzy sound like Introzy to buyers and agents alike. Our engineers get clear, data-backed direction, so every landing page and help article they ship builds on the last one. And as we grow, the people who join us will inherit a record of why we made each call, not just what we shipped. The tools will keep changing. The judgment is the part that has to stay ours.

Stop reading about referrals. Start tracking them.

Free up to 5 referrals a month. Every feature included. No card to start.