Profound reached $1.8B on AI answers. A solo founder's plan for being the answer

A solo founder cannot outspend anyone. The usual assumption is that the bigger budget buys the attention. Two stories from this window suggest a second route is opening, and it rewards organization more than spend.

What happened

• September 15. Profound, which helps brands show up in the answers AI assistants give, raised $180 million in a Series D at a $1.8 billion valuation, seven months after its previous round. The company says its revenue tripled in six months, across more than 1,000 enterprise customers. Those figures are company-reported.
• September 23. Meta said more than 1,500 developers applied to build Muse connections in under a week, according to TechCrunch’s roundup of what is coming to Meta’s AI agent. That count is Meta’s.

What it means for a company of one

Two different things are moving.

The first is answers. Enterprises are paying real money to show up when a buyer asks an assistant for a recommendation. Profound’s customers are marketing teams with budgets. You are not one of them. But the question they are paying to work on (does the assistant name us, and what does it say?) is one you can start checking yourself this afternoon.

The second is agent platforms. An agent platform that accepts outside connections gives a buyer’s agent a way to reach a product without a search, a click or a demo. We do not yet know how those connections will be ranked, reviewed or paid for. What the developer count shows is that plenty of builders want in early.

Our read: you cannot win the first route by outspending incumbents, but you can be the best-organized answer. You cannot win the second by having the biggest brand, but you can be the easiest product for an agent to use.

The answer-readiness checklist

Six checks. Each fits in an afternoon for one person.

1. The ten questions. Write the ten questions a buyer would type into an assistant before buying something like yours. Use their words, not your category name.
2. One page per question. Each question gets a page that answers it in the first two sentences, then shows the evidence.
3. Plain facts in plain text. Price, who it is for, who it is not for, known limits and how to start, written as text rather than locked in images or a pricing widget.
4. One name, one description. The same product name and one-line description on your site, docs, listings and profiles. Conflicting descriptions invite conflicting answers.
5. Dated evidence. A changelog and case notes with real dates, so a reader (human or machine) can tell what is current.
6. A way in for agents. If your product’s core action is a single verb (book, quote, check, invoice), expose it: a documented API, an MCP server, or a connector on a platform your buyers already use.

Then measure it. Once a month, ask your ten questions in the assistants your buyers use. Log whether you are named, what is said, and which page is cited. Enterprises now pay companies like Profound to work on this problem at scale. At your size, a spreadsheet and an hour a month is a real start.

Nothing on this list guarantees an assistant will recommend you, and anyone who promises placement is selling something they cannot control.

Answer first, or connector first?

A simple rule for where to spend the next month:

• Answers first when buyers research before they buy and the product needs explaining.
• Connector first when the core action is one verb an agent could call on the buyer’s behalf, and your buyers already spend their day on that platform.
• Neither yet when you cannot write the ten questions. Talk to buyers until you can.

A worked example, hypothetical: say you sell a tool that turns a freelancer’s timesheet into an invoice. The ten questions include “how do I invoice from my timesheet” and “invoicing tool for freelance designers.” The product still needs explaining, so answers come first: ten short pages, one per question. But the core action is one verb, invoice, so the connector is next month’s job, on whichever agent platform your freelancers already use.

What to do this week

• Write your ten questions, in buyers’ words.
• Ask them in two or three assistants and log what comes back.
• Rewrite the page behind the worst answer so the first two sentences answer the question.
• Apply the rule above and write down which route gets next month.
• Read TechCrunch’s Muse roundup yourself before you plan any connector work. Platform terms change, and the details matter more than the headline.