250M+ people · 200,000+ signals · 15T precomputed relationships

Generate audiences no one else can reach.

Describe the people you want to reach in plain English. Watt composes a sized audience from 200,000+ behavioral signals, refreshed daily, and pushes it to your channel of choice. Driven by a Claude plugin.

Request access Signal engineering at work.

Compatible with:

Connectable to:

Running audiences in production at

Watt Audiences. Compose audiences nobody else can.

Inside the Claude org your team already uses, Watt builds the audience the way your best operator would, if your best operator had 200,000+ behavioral signals at their fingertips.

It:

Analyzes how your best customers actually behave
Surfaces relevant signals: purchase behavior, life events, category affinities
Lets your operator rank what matters
Composes a precise audience definition
Pushes it to Meta, Google, TikTok, or your ESP
Refreshes it daily against new data

No UI to learn, no seat licenses. You own the logic.

See Watt Audiences →CAC down up to 80% · Conversion up to 10×

Audience summary

In production today.

Generate$100M beauty brand

CAC down 80%

$100M beauty brand customer cut acquisition cost by 80% through autonomous ICP analysis and lookalike agents built on Watt.

ActivatePerformance marketing agency

24× ToFu conversions

National men's health customer increased top-of-funnel conversion rate 24× through autonomous audience refinement agents running on the Signal Graph.

The graph

The Signal Graph.

The same caliber of infrastructure that powered hedge funds and government labs, now sitting inside the Claude org your marketing team already uses.

250M+

Verified people

60M+

Businesses covered

145,000+

Signals on people

55,000+

Signals on businesses

15T

Pre-computed relationships

The audiences nobody else can reach.

Three patterns from agencies running this today, across very different verticals.

Niche products · thin signal

Bioplastics, on Meta's dropdown? Nothing useful.

Meta's prebuilt interests don't know what a bioplastics buyer looks like. Watt surfaces:

Eco-packaging purchase history
Sustainable-living subscription patterns
Composting household signals
25+ other behavioral patterns that correlate with the actual buyer

Regulated industries · sparse menus

Healthcare. Financial services. Meta's prebuilt categories don't go where your product does.

When the platform's interest taxonomy has nothing useful and demographic categories are off-limits, Watt surfaces behavioral signal that does the same job, legally:

Chronic-condition management indicators
Complex-financial-product evaluation signals
In-market behavior the interest taxonomy doesn't carry
Composed without the protected attributes the platform won't let you use

Hyper-local DTC

The last mile is physical. The targeting has to be too.

Stores, dispensaries, service businesses. Anything where the customer has to physically get there. Watt builds audiences as compositions of:

25–50 mile radii drawn around each of your locations
Behavioral signal intersected with each radius
Boolean expressions across multiple sites
Not just a static demographic ring around a zip code

If your audience is hard to find on Meta's dropdown menus, that's exactly the audience Watt is built for.

Getting started

Become a Signal Engineer in minutes.

01Install the plugin

One command. No account to create, no config to wire. The Watt plugin bundles the MCP server, skills, and sub-agents. Install it once in Claude Code or Cowork, log in, and you're live.

02Describe your audience

03Activate anywhere

Claude Code

Trusted by world-class teams.

RevenueBase

We build and sell data products for a living, so we aren't easily impressed. Most 'AI-ready' data is a warehouse with an MCP bolted on. Watt is the first thing I've seen where an agent can actually reason across the signals and compose things nobody pre-built. That's the hard part, and they built it.

Mark

Senior Director

Lahzo

No matter what you want to do, a marketer will always say 'Yeah, we do that.' But when you say, 'Do you know who's researching your category and products? Do you know their financial situation? Do you know the specific signals that matter for your category?', they're like, 'No, we don't do that.' Watt does.

Mitchell

Director of Strategy

A handful of companies hoard the world's signals.

Now one Signal Engineer can access them.

Latest thinking.

Product

The Watt waitlist is open.

The Watt public waitlist is open. Connect the plugin in Claude, brief it like an analyst, and generate your first custom audience from 160,000+ behavioral signals. No setup, no data team.

June 2026Read more →

Engineering

What is the Signal Graph?

An agent can't reason over petabytes of raw vendor data. Watt Chief Architect John Zila explains the substrate that turns it into 162,000 composable signals, with the answers already on the shelf.

June 2026Read more →

Research

Signal Drop 01

The first Signal Drop. Education signals on 118M+ people (alumni queries in one line), geographic filters that finally compose like every other signal, plus company founding year, LinkedIn contacts, work-vs-personal flags, mobile ad IDs, and 1,700+ new intent topics. See you in two weeks.

June 2026Read more →

View all research →

Build your first audience.

30 minutes from signup to your first audience in production.

Request access Signal engineering at work.

Common questions.

Generate audiences no one else can reach.

Watt Audiences. Compose audiences nobody else can.

Inside the Claude org your team already uses, Watt builds the audience the way your best operator would, if your best operator had 200,000+ behavioral signals at their fingertips.

It:

Analyzes how your best customers actually behave
Surfaces relevant signals: purchase behavior, life events, category affinities
Lets your operator rank what matters
Composes a precise audience definition
Pushes it to Meta, Google, TikTok, or your ESP
Refreshes it daily against new data

No UI to learn, no seat licenses. You own the logic.

See Watt Audiences →CAC down up to 80% · Conversion up to 10×

Audience summary

In production today.

Generate$100M beauty brand

CAC down 80%

$100M beauty brand customer cut acquisition cost by 80% through autonomous ICP analysis and lookalike agents built on Watt.

ActivatePerformance marketing agency

24× ToFu conversions

National men's health customer increased top-of-funnel conversion rate 24× through autonomous audience refinement agents running on the Signal Graph.

The audiences nobody else can reach.

Three patterns from agencies running this today, across very different verticals.

Niche products · thin signal

Bioplastics, on Meta's dropdown? Nothing useful.

Meta's prebuilt interests don't know what a bioplastics buyer looks like. Watt surfaces:

Eco-packaging purchase history
Sustainable-living subscription patterns
Composting household signals
25+ other behavioral patterns that correlate with the actual buyer

Regulated industries · sparse menus

Healthcare. Financial services. Meta's prebuilt categories don't go where your product does.

When the platform's interest taxonomy has nothing useful and demographic categories are off-limits, Watt surfaces behavioral signal that does the same job, legally:

Chronic-condition management indicators
Complex-financial-product evaluation signals
In-market behavior the interest taxonomy doesn't carry
Composed without the protected attributes the platform won't let you use

Hyper-local DTC

The last mile is physical. The targeting has to be too.

Stores, dispensaries, service businesses. Anything where the customer has to physically get there. Watt builds audiences as compositions of:

25–50 mile radii drawn around each of your locations
Behavioral signal intersected with each radius
Boolean expressions across multiple sites
Not just a static demographic ring around a zip code

If your audience is hard to find on Meta's dropdown menus, that's exactly the audience Watt is built for.

Common questions.