Building a Product Identity for an AI Platform

CNN’s Technology & Engineering group had built a natural-language analysis engine trained on 40+ years of news content that could assess context and sentiment across text, audio, video, and galleries. I read the full engineering white paper and went deep with the engineering and product teams on performance. That’s where the competitive advantage became clear: because CNN’s training data was four decades of editorially rigorous, clean content, the model was outperforming Google Brain and IBM Watson, approaching human-level accuracy. The engineering team knew they’d built something good. They hadn’t framed why it was better than everything else on the market.

Meanwhile, the commercial side was bleeding. News spend was declining. Make-goods were increasing. Brands were freezing campaigns against keyword blocklists that couldn’t distinguish between a market crash and a plane crash. Boeing doesn’t care about economic turbulence, but they can’t be adjacent to aviation disaster coverage, so they’d pull everything rather than take the risk. Significant advertiser spend was leaving news environments entirely, and the existing tools were keyword filters with better branding.

From Engineering White Paper to Product Identity

In partnership with my EVP, I led the commercial development: product strategy, positioning, naming, visual identity, and the cross-functional alignment across Engineering, Editorial, Ad Sales, Ad Ops, Marketing, Legal, and Policy that turned a technical capability into something a sales team could sell and an advertiser could trust. We mapped CNN’s NLP engine to IAB taxonomy so media buyers could evaluate brand suitability with precision rather than fear, and packaged S.A.M. for activation across both direct and programmatic channels (DV360, Xandr).

The narrative challenge was real. We weren’t launching a feature. We were asking enterprise advertisers to move spend back into environments they’d abandoned. That required technical credibility (contextual analysis, not keyword avoidance, with performance data to back it up) and emotional confidence (your brand is safer here, and here’s the proof). Every piece of enablement and every sales conversation had to carry both.

The Design and Narrative Challenge

Every piece of the identity had to carry both. The product mark, the sales materials, the demo experience, the enablement decks, the conference presentations — all of it needed to make AI feel trustworthy and precise without being cold or intimidating. We mapped CNN’s NLP engine to IAB taxonomy so media buyers could evaluate brand suitability with precision rather than fear, and packaged S.A.M. for activation across both direct and programmatic channels (DV360, Xandr).

The visual identity decisions were inseparable from the commercial narrative. How the product looked was how it felt was how it sold. A clinical, engineering-forward identity would have reinforced the very distance advertisers already felt from AI tools. A warm, approachable identity would have undermined the technical credibility that was S.A.M.’s actual competitive advantage. The design had to hold both: precision and confidence, and every touchpoint had to be consistent enough that a sales team could walk into any meeting and the story held.

Results

It worked. Historically cautious categories returned spend to news. Campaign pauses during breaking news dropped. Make-goods declined. Revenue impact across both premium direct and programmatic. S.A.M. earned Promax Gold and Digiday’s “Best Brand Safety Strategy.”

My role: product strategy and positioning, naming, visual identity, research synthesis (translating engineering documentation into product requirements and commercial language), packaging and enablement, measurement framework, and the cross-functional process that kept seven departments aligned through multiple iterations.