Domino's asked an AI agent to generate ad creatives. The result? A pizza that doesn't exist on their menu. The LLM had simply imagined what a "Domino's pizza" should look like. This is exactly why most AI marketing tools have been failing — and why Hightouch just hit $100M ARR.

TL;DR
Data warehouse Brand asset connection AI agents On-brand campaigns 80% faster creative

What's actually different here?

Hightouch is a 7-year-old startup that originally built a Composable CDP (customer data platform). The idea was simple: instead of copying data into a separate system, let marketers pull directly from their existing data warehouses — Snowflake, Databricks, BigQuery — without duplication or 6-month implementation timelines.

In late 2024, they went a step further: AI agents built on top of that Composable CDP. Before, marketers used the data to manually build campaigns. Now, AI agents read the data, identify audiences, and generate ad creatives — around the clock.

Here's the thing, though: the key innovation isn't generating creatives from scratch — it's pulling from existing brand assets. Figma files, photo libraries, CMS content — Hightouch connects directly to what the brand already owns. So instead of hallucinating a Domino's pizza, the AI grabs an actual photo of their pizza and generates only the background.

$100M
ARR (April 2026)
20 months
to add $70M ARR
$2.75B
Series D valuation

The market validated this fast. Hightouch reached $100M ARR as of April 2026, adding $70M in just 20 months since launching AI products. On April 29, Goldman Sachs and Bain Capital Ventures led a $150M Series D, pushing the valuation to $2.75B. Two consecutive years of 100%+ growth.

In January 2026, Gartner included Hightouch in its Magic Quadrant for Customer Data Platforms — on the first inclusion, they scored highest in Ability to Execute. The platform now syncs 7.3 trillion+ records and processes 10 billion+ AI Decisioning events.

How is this different from regular AI tools?

"I tried using AI to make ad creatives. They all looked... off." — this is the #1 complaint from marketers. Hightouch heard the same thing from countless CMOs. Colors wrong. Fonts off-brand. Tone doesn't match. In extreme cases, products that don't exist show up in ads.

Generic AI tools (GPT, Midjourney, etc.) Hightouch Brand AI
Image generation AI imagines what the brand looks like Uses existing assets; AI generates only background
Brand consistency Colors, fonts, and tone may be off Pulled directly from Figma and CMS
Data connection Generic generation, no customer data Directly connected to existing data warehouse
Quality control Manual review required LLM judge scores outputs automatically
Speed Baseline 80% faster ad creative generation

Hightouch's stack breaks down into three layers.

  1. Context layer
    Customer data from the warehouse + design assets from Figma + external market signals — all unified. This is how AI actually learns a brand.
  2. AI agents
    Agents built on that context layer analyze audiences, propose campaign ideas, and execute across channels. Marketers can approve or let agents run autonomously.
  3. LLM judge
    A second AI automatically grades the outputs. "Does this creative pass brand guidelines?" — pass/fail handled automatically, no manual review needed.

A large publicly-traded financial services company using Hightouch saw ad creative generation speed up by 80% and reach expand by ~10%. Customers include DraftKings, Domino's, Ramp, Spotify, PetSmart, Grammarly, Warner Music Group, and Whoop.

Composable CDP vs. traditional CDP — the key difference

Traditional CDPs copy your data into their own system — 6+ month implementations, data duplication, high maintenance. Composable CDP reads directly from your existing Snowflake, Databricks, or BigQuery. Your data warehouse becomes your marketing infrastructure.

How to get started

  1. Request a demo and assess fit
    Start at hightouch.com. Hightouch is built for companies already running a data warehouse. If you're managing customer data in spreadsheets, this isn't the right tool yet.
  2. Connect your data warehouse
    Connect your existing Snowflake, Databricks, BigQuery, or Redshift. No data copying — your warehouse stays the single source of truth.
  3. Connect brand assets
    Link Figma, your photo library, and CMS. AI will learn from these assets and prioritize them when generating creatives.
  4. Configure AI Decisioning
    Define which channels (email, SMS, ads), which segments, and which goals you're targeting. AI optimizes 24/7.
  5. Scale autonomy gradually
    Start with "agent draft → marketer approval → execution." Increase autonomous execution as trust builds.

Not the right fit yet if...

You're managing customer data in spreadsheets, or your marketing team is fewer than 3 people. Hightouch is enterprise-grade infrastructure — SMB marketers should start with Klaviyo or Mailchimp to build the automation foundation first.

Go deeper

Hightouch Series D Blog Post The co-CEOs explain their agentic marketing roadmap and why they're making the shift from CDP to full AI platform. hightouch.com

Sapphire Ventures — Why We Backed Hightouch VC-level analysis of the AI Decisioning market and what makes Hightouch's approach structurally different. sapphireventures.com

AdExchanger — CDP to AI Platform Pivot Analysis A martech journalist's breakdown of Hightouch's transition from data plumbing to agentic marketing. adexchanger.com

Hightouch Composable CDP Overview The official explainer on how data-warehouse-native marketing infrastructure works. hightouch.com

BusinessWire Series D Press Release The official announcement with Goldman Sachs, Bain Capital quotes and customer case studies. businesswire.com

Gartner Magic Quadrant — Composable CDP Where Hightouch was rated highest in Ability to Execute on its first inclusion. gartner.com