How Inflow became HumanDeploy.

A year ago I started a company called Inflow. The pitch was simple. An AI-managed creative team that lives in your Slack. You ask for a landing page, a deck, a launch video. The AI scopes it, prices it, routes it to a senior expert, and ships it back to the channel you asked from. No portal. No project manager. No briefs.

We built it. Twenty-seven kinds of work. Real experts. A brand-intelligence layer that read your site on day one. An agent in Slack that could scope a sprint in twelve seconds and hand it off without anyone writing a Notion doc. It worked. People paid for it. We were proud of it.

The whole time, something kept nagging at me. The product we were shipping was a creative team. The company we were building was something else.

What Inflow taught us

The first thing you learn when you put an AI agent in a customer's Slack is that the agent is the company. Not the website. Not the dashboard. Not the contract. The conversation in the channel is the entire product. If the agent is sharp, present, and useful, the customer stays. If it isn't, nothing else you build will save you.

The second thing you learn is that creative work is the visible tip of a much deeper iceberg. Behind every can you make this deck is a positioning question. Behind every landing page is a question about who the buyer really is. Behind every launch is a campaign that needed to be planned three weeks earlier. Customers kept asking us for the deck and then, almost apologetically, asking for the strategy underneath it. We started saying yes to both.

The third thing you learn is that the most valuable thing the system owns isn't the work it ships. It's the context it accumulates. By month three our agent knew our customers' positioning, their voice, their last six campaigns, what their CEO hated, what their sales team kept asking for. That context was the product. Every new request was cheaper, faster, and better than the one before it.

What we kept hearing

By the start of 2026 almost every customer call sounded the same. They had AI everywhere. Agents writing copy. Agents drafting outbound. Agents summarizing Gong calls. Tools layered on tools layered on tools. And they were exhausted.

Not because the AI was bad. Because the AI was unsupervised. There was nobody in the loop who knew the brand, the buyer, the strategy, last quarter's experiments, and could just make the call. The agents were generating output. The humans were drowning in review. The work that actually went to market kept feeling slightly off.

The thing the market was missing wasn't more AI. It was the thing that goes after AI. The senior human in the loop. The judgment call. The taste. The "no, that's not us," followed by a fix that took ten minutes instead of two days because the system already had every piece of context that mattered.

Pulling the lens back

So we pulled the lens back. We took the architecture we'd built for Inflow. The Slack-native agent. The context graph. The expert network. The AI-scoped handoffs. The MCP integrations into the customer's stack. And we asked a different question. What if the wedge isn't creative? What if it's the entire go-to-market?

Brand. Web. Story. Sales. Campaigns. Five areas of work. One agent in your Slack. A expert network behind it. A context graph that gets richer every month. AI that handles the volume. Humans deployed against the work that needs judgment.

That's a different company. It needed a different name.

Why HumanDeploy

The name is on purpose. Everyone in the AI stack talks about deploying agents. The more interesting move is to deploy humans the same way. Programmatically. In context. Only where they're needed. With everything they need already loaded in their hands when they show up.

The shape of work is changing fast. AI is going to do more and more of the volume. That's already true. The honest question is what happens to the humans. The lazy answer is they go away. The honest one is they get pulled toward the work that actually requires them. The strategy. The taste. The judgment. The conversation that moves a deal. The call that changes a positioning. And away from the work that never should have required them in the first place.

HumanDeploy is a bet on the second answer. AI raises the floor. Humans raise the ceiling. The system in the middle is what makes it possible to do both at once.

The point of AI in work was never to remove the human. It was to deploy them where they matter.

What we're shipping

An AI agent that lives in your Slack and runs your GTM. It connects to your CRM, your analytics, your sales intelligence, your content stack. Through MCP, not screen-scraping. It maintains a compounding context graph of your business. Positioning, ICP, voice, competitors, pipeline, what worked, what didn't. When the work is in its lane, it ships. When it isn't, it deploys a senior expert from a curated network with the brief already written, the context already loaded, and a clear bar for what good looks like.

You don't manage anyone. You don't log into anything. You don't write a brief. You message your agent in the channel you already work in, and the work shows up.

$2,500 a month. No contracts. Cancel any time. We have to earn the month, every month. That's a feature, not a concession.

What's next

We're opening to a small group of founding customers this month. Most of them came from Inflow. A few came from this blog. If you're a B2B founder or revenue leader between roughly $1M and $50M ARR, and you've felt the same thing (too much AI, not enough judgment, too many tools, not enough team) we'd like to meet you.

We'll write more here. About the architecture. About the experts. About the things we're getting wrong as we go. The blog is the place we'll think out loud, so the people who care about this can think along with us.

Inflow became HumanDeploy because the thing we were really building deserved its real name. The point was never the design service. The point was never even the AI. The point was the team you actually want, in the place you already work, getting smarter every month, and shipping the work that matters.

— Steffan