Adora bets enterprise brands need AI trained on their own brand DNA

An AI marketing platform for large enterprise teams: it learns a company's 'brand DNA,' generates on-brand ad variations for channels like Meta, Google and Pinterest, routes content through human approval, and measures which visuals and messages drive conversions.

That AI trained on a brand's own assets, catalog, guidelines and first-party data can generate on-brand ads and prove what converts better than generic tools.

Seed funding from Village Global, the San Francisco early-stage firm chaired by Reid Hoffman; exact amount undisclosed.

Marco Matos spent four years at Pinterest leading its Seattle office and helping build advertising products, with earlier product roles at Facebook and Google. Kabir Shahani co-founded customer-data startup Amperity, where he was CEO until abruptly leaving in 2023, and earlier launched marketing automation startup Appature, acquired in 2013. The two founded Adora in 2023 to rethink how big brands go to market.

By its October 2025 stealth exit, Adora said early customers such as Alaska Airlines, Brooks Running and Servco had improved return on ad spend, cut acquisition costs and accelerated content creation — company-reported outcomes. The system trains on each client's creative assets, catalog, guidelines and first-party data to learn its 'brand DNA,' then generates channel-specific ad variations that teams review and approve before launch, with an analytics layer identifying what converts. Matos positioned AI as 'the next step function in how marketers go to market,' in a field that already includes Adobe and Salesforce plus startups like Typeface and Jasper.

The bet is specific about where value sits: an enterprise-tuned platform trained on brand-owned data, not a generic model wrapper.

The routing-system architecture — picking models per task — is a distinguishable technical decision in a market full of single-model tools.

The founders' backgrounds line up with the wedge: Pinterest ad products on one side, first-party customer data (Amperity) and marketing automation (Appature) on the other.

Named early customers — Alaska Airlines, Brooks Running, Servco — and a Village Global seed give the launch concrete traction markers.

In a crowded AI application market, the defensible layer is not the model but the routing: owning which model handles which task, trained on data competitors cannot reach — the customer's own brand.

As of October 8, 2025, Adora had emerged from stealth with seed funding from Village Global (amount undisclosed), around 20 employees, and early enterprise customers in Alaska Airlines, Brooks Running and Servco — with improved return on ad spend, lower acquisition costs and faster content creation reported by the company. It was competing against Adobe and Salesforce as well as AI marketing startups Typeface and Jasper, selling large enterprise marketing teams an AI platform with a proprietary multi-model routing system.