How Webhound added 200 users with Saava's intent data

What wasn’t working

Webhound charges for research actually done, with no seats and no subscription, so its best users are the ones with real questions to answer again and again: people already spending heavily on AI, running the same kind of research every week, with money riding on the answer. None of that shows up in a job title. An analyst, a founder and an engineer can all be the right user, or none of them can.

What they changed

Webhound described its buyer to Saava in one sentence: prosumers with high AI spend and a repeatable research workflow, where a large dollar amount rides on the research outcome. An intent monitor listens for seven ways people say that out loud, from our OpenAI bill and LLM spend to automating our research with AI. Tracked profiles watch the audiences of AlphaSense and Perplexity leadership, where people already interested in research tools gather. Saava scores everyone it finds against the description.

How it played out

1. Research that stops when the budget does

   Most AI agents stop searching as soon as they have enough to sound confident. Webhound keeps following leads and checking sources until the budget it was given is spent, and returns the answer with the work behind it: cited reports, datasets where every cell has a source, and claims an agent can filter by confidence. Agents call it over MCP or an API, and people can drive the same engine from its own interface.

2. Buyers who already pay for AI

   Because Webhound bills by the research done, the users who matter are the ones with a steady stream of questions worth paying to answer properly. The team wrote that down as the buyer description instead of a list of titles, so Saava looks for the habit and the stakes rather than a role.

3. Listening for the bill

   People describe the problem Webhound solves when they talk about cost and workflow: what they spend on AI tools, their API costs, their OpenAI bill, the research pipeline they are building, the research they want to automate. The monitor watches for exactly those phrases.

4. The audience next door

   AlphaSense and Perplexity sit close to Webhound in what their users care about. Tracking AlphaSense's page and founder, and Perplexity's cofounder, turns the people reacting to their posts into names Webhound can reach while research tools are on their mind.

5. What came back

   Since July, Webhound has added 200 users that it credits to Saava's intent data and lead generation.

The steps, if you want to try this

1. 01Describe your buyer by what the work is worth to them, not by title. Webhound wants people whose research decides a large amount of money.
2. 02Listen for people talking about what they already spend: AI tool bills, API costs, LLM spend, research they are trying to automate.
3. 03Track the leaders of the tools your buyers already use. The people engaging with AlphaSense and Perplexity are already paying attention to research tools.
4. 04Let every person be scored against the description before they reach your list, and start with the strongest fits.