I spent six years studying marketplace dynamics before joining Pulse. Cold start problems are well-documented. Uber solved it city by city, subsidizing drivers before riders showed up. Airbnb solved it by going door to door in New York, photographing apartments. Every network product in history has faced the same question: how do you deliver value when nobody else is on the platform yet?
A coordination layer for agents faces this problem at a deeper level. It's not just "users need other users." It's "agents need other agents to coordinate with." If the whole point is agent-to-agent communication, what happens when there's only one agent?
This is the question I've been working through for the past several months. And the answer — the one we're actually executing on — turns out to be counterintuitive. The cold start isn't a problem we need to overcome. It's the shape of our entire go-to-market strategy.
The Double Chicken-and-Egg
Classic marketplaces have a single chicken-and-egg problem. Riders won't download Uber if there are no drivers. Drivers won't sign up if there are no riders.
A coordination layer has two:
- Users won't join if there's no one to coordinate with.
- Even if users join, the real value — bilateral agent coordination — requires agents on both sides.
This means we can't just grow a user base and hope the network effect kicks in. We need coordination events happening from day one. Real interactions. Real value delivered. Not a promise that "it'll be useful once your contacts join too."
That framing is a death sentence for a network product. We've all seen it: "Invite your friends and this app becomes amazing!" Nobody invites their friends. The app dies.
So we refused to build that product.
The One-Sided Insight
Here's what changed our entire strategy: limited agent deployment only requires one side to have an agent.
Think about what that means. A founder creates their AI COO on Pulse, loads it with their pitch deck, product context, and calendar availability. Then they share a link. The investor who clicks that link doesn't need a Pulse account. Doesn't need to download anything. Doesn't need to understand what a coordination layer is. They just click and talk.
From the investor's perspective, they're having a conversation with an unusually responsive, well-informed AI that represents the founder. They ask questions, get instant answers, and book a meeting in sixty seconds. Done.
From our perspective, a coordination event just happened. Value was delivered to both parties. The coordination tax — the emails, the scheduling ping-pong, the 18-hour response delays — was eliminated for that interaction. And only one person needed to be on the platform.
This is the insight that Airbnb had in 2009, applied to a new domain. Airbnb didn't need both hosts and guests to sign up simultaneously. They needed hosts first. They went apartment to apartment in New York, taking professional photos, building the supply side. Guests came because the listings were already compelling.
We don't need both sides to deploy agents simultaneously. We need one side to deploy an agent that creates such an obvious "wow moment" that the other side asks: how do I get one of these?
The Conversion Funnel Nobody Designed
What happened next surprised even us.
The funnel emerged organically. A founder shares their Pulse link with fifteen investors. Those investors interact with the agent. Some percentage of them — and it's a meaningful percentage — have the same reaction: "Wait, I could use this for my own deal flow." Or: "My portfolio companies need this for their fundraising."
We didn't design a referral program. We didn't offer credits for invites. The product itself is the referral mechanism. Every time someone interacts with a deployed agent, they experience the value firsthand. They're not reading a landing page or watching a demo video. They're living the product from the recipient side.
The funnel looks like this:
- Founder deploys agent with context and calendar.
- Recipient clicks link and talks to the agent.
- Recipient experiences zero-friction coordination — instant answers, one-click scheduling.
- Recipient thinks: "I want this for my outbound."
- Recipient signs up and deploys their own agent.
- Their recipients have the same experience. The cycle repeats.
Each deployment seeds the next generation of deployments. Not through marketing spend, but through direct product experience.
The Use Cases Driving the Cold Start
Not all use cases are equal for cold-starting a network. We've learned to focus on the ones with the highest coordination tax and the most natural share-ability:
Fundraising
A founder talks to 30-80 investors per raise. Each conversation starts with the same questions: What's your business model? What's the traction? Who's on the team? A single Pulse deployment replaces 30-80 repetitive email threads. The investor experience is dramatically better — they get answers at 11pm on a Sunday instead of waiting until Monday morning. And investors talk to dozens of founders, which means high exposure to the product.
Recruiting
Candidates have questions about the role before they commit to a formal interview. Companies have the same answers to give every candidate. A Pulse link in a job posting lets candidates self-serve on information and self-schedule interviews. One deployment serves hundreds of candidates. Every candidate who interacts is a potential future deployer.
Sales Outreach
Prospects want to evaluate before they commit to a demo call. A Pulse link lets them explore pricing, features, and integration details on their own terms. The AI COO qualifies leads and books meetings with the ones who are ready. Sales teams share links with hundreds of prospects per month — massive exposure.
Each of these use cases has the same structure: one person deploys, many people interact. The ratio of deployers to recipients creates natural distribution. Every interaction is a product demo that the user didn't have to organize.
The Metrics That Actually Matter
When you're cold-starting a network, the temptation is to track MAU. Monthly active users. It's the metric every investor asks about, every board deck features. But MAU is the wrong metric for a coordination layer in cold start.
Here's what we actually track:
Coordination events per deployment. How many meaningful interactions does each deployed agent generate? If a founder deploys one link and it generates 40 conversations and 12 booked meetings, that's a high-yield deployment. This metric tells us whether each node in the network is actually creating value.
Recipient-to-deployer conversion rate. What percentage of people who interact with someone else's agent go on to deploy their own? This is the viral coefficient of the network. Above a certain threshold, growth becomes self-sustaining.
Time-to-first-value. How quickly does a new user go from signup to their first coordination event? If it takes three days and fifteen configuration steps, we lose them. If it takes ten minutes and one shared link, we keep them. We obsess over this number.
Recipient satisfaction without signup. This is subtle but critical. If recipients have a bad experience — if the agent gives wrong answers or feels like a gimmick — it poisons the well. Every bad interaction is anti-marketing. So we measure how recipients rate their experience even though they never signed up for anything.
These metrics tell a different story than MAU would. They tell us whether the network is actually working as a coordination layer, not just whether people created accounts.
The Inflection Point We're Building Toward
Right now, Pulse operates primarily through single-sided deployments. One person has an agent. The other person just clicks a link. This is Phase 1, and it delivers real value today.
But there's an inflection point coming, and we can feel it approaching.
It happens when enough recipients have deployed their own agents that bilateral coordination becomes possible. When a founder's agent can talk directly to an investor's agent, and the scheduling, context-sharing, and follow-up happen entirely between agents — that's when the network of agents stops being a vision and starts being reality.
The math is simple. If each deployment generates an average of N recipients, and some percentage of those recipients become deployers, then the network grows geometrically once the conversion rate crosses a threshold. Below the threshold, growth is linear and paid-marketing-dependent. Above it, growth is organic and self-sustaining.
We're not there yet. We're honest about that. But every single-sided deployment moves us closer, because every interaction is a chance to convert a recipient into a deployer.
Phase 2: When Both Sides Have Agents
Here's what Phase 2 looks like, and why it's worth building toward:
Today, a founder shares a Pulse link with an investor. The investor talks to the founder's agent. Great — the founder saves hours.
In Phase 2, both the founder and the investor have agents. The founder's agent proactively identifies investors whose criteria match the company's profile. The investor's agent screens inbound opportunities based on thesis fit, stage, and check size. When there's a match, the agents coordinate: they share relevant materials (with access-aware security), find mutual calendar availability, and schedule a meeting — all before either human lifts a finger.
That's not an incremental improvement. It's a category change. Pulse stops being a tool for sharing context and becomes infrastructure for professional coordination. The coordination tax doesn't just decrease — it effectively disappears for routine interactions.
But you don't get to Phase 2 by announcing Phase 2. You get there by making Phase 1 so good that people pull Phase 2 into existence by demanding it.
Why the Cold Start Is Our Advantage
Here's the part that took me longest to see.
Most people view cold start as a weakness. "You don't have network effects yet" is the polite version of "you might never get them." Every investor meeting, every advisor conversation — someone raises it.
But the cold start period is when we do something that an established network can't: we hand-select the first nodes. We work directly with early deployers. We learn exactly which use cases generate the highest coordination-event-per-deployment ratio. We tune the recipient experience based on real conversations, not aggregate analytics.
Uber couldn't hand-pick which riders to onboard in San Francisco. We can choose which founders, which sales teams, which recruiting operations to enable first. And those choices compound — because a well-chosen early deployer creates more recipients, who convert at higher rates, who create more deployers.
The cold start isn't the phase before the product works. It's the phase where we build the foundation that determines whether the network works at scale.
What We're Not Doing
It's worth naming the strategies we've deliberately rejected:
- We're not running a "invite 5 friends" campaign. Forced virality creates spam, not networks.
- We're not trying to build MAU before revenue. We'd rather have 500 deployers generating 10,000 coordination events than 50,000 signups generating nothing.
- We're not pretending the network already exists. The value proposition today is single-sided deployment. It's honest, it's real, and it works.
- We're not subsidizing both sides. The deployer gets clear value (hours saved). The recipient gets clear value (instant answers and scheduling). Neither needs a discount to participate.
The First 1,000 Coordination Events
We're not trying to build a billion-user network on day one. We're building toward the first 1,000 coordination events that prove the network works — that prove agents can replace the coordination tax for real professional interactions.
Every one of those events teaches us something. Which prompts do recipients ask most? Where does the agent fall short? What makes someone convert from recipient to deployer? What makes them share the link with their own network?
A thousand coordination events, deeply understood, are worth more than a million page views. They're the proof that agent-based coordination isn't theoretical — it's operational.
If you're a founder who sends the same pitch to dozens of investors, a recruiter who answers the same candidate questions every week, or a sales lead who knows the demo-scheduling dance too well — you're exactly who we built this for.
Deploy your first agent. Share one link. Let the network start with you.
Join the waitlist · Launch Pulse · Read about limited agent deployment