What I Learned Advising 5 Companies on AI Adoption

Kings League is one of the fastest-growing sports properties in the world — built from scratch by Piqué and Ibai, scaling to tens of millions of fans across digital platforms in just two years. When I came in, the challenge wasn't enthusiasm for AI. The challenge was abundance. The team had identified compelling AI opportunities across fan engagement, broadcast, mobile, content creation, and competitive data. All of them were real. None of them had priority.

The advisory work was almost entirely about constraint. We built a forcing function: rank AI investments by two axes — revenue proximity (how directly does this connect to a business outcome within two quarters?) and infrastructure leverage (does this compound, or is it a standalone tool?). The answer cut the AI roadmap from eleven initiatives to three. Those three shipped. The others are still "being evaluated" twelve months later at comparable organizations that didn't make the same cuts.

Cyanite is building AI infrastructure for the music industry — intelligent tagging, mood analysis, and search for music libraries at scale. The product thesis is technically compelling and the team is strong. The challenge when I came in: the AI product roadmap and the fundraising narrative were evolving independently. What the team was building, what investors were being told, and what customers were actually paying for had begun to diverge.

The advisory work here was largely about convergence. We spent significant time aligning the internal product thesis — what AI capabilities they were actually building and why — with the investor-facing narrative. The goal wasn't spin. It was finding the version of the story that was both true and compelling. When those two things align before the fundraising process, conversations are faster and terms are better. When they don't, you're negotiating against your own ambiguity.

Blanket Homes is building AI-powered tools for property management — automating workflows across leasing, maintenance coordination, and tenant communication. The product ambition is high. The constraint is real: property management is a trust business. Landlords and tenants have relationships that extend years. AI that reduces friction for the platform but introduces friction or opacity into those relationships destroys the core product, not just a feature.

The advisory work here was about sequencing. Before any engineering resources were committed, the leadership team needed to answer a specific question: where does AI serve the relationship, and where does it threaten it? We mapped the workflow end-to-end and categorized every AI opportunity by that lens. Automation behind the scenes (maintenance routing, document processing, renewal triggers) went straight to the roadmap. AI in direct tenant communication required a different approach — a more careful, human-in-the-loop architecture. Getting leadership aligned on that line before engineering started saved months of rebuild.

UniperCare operates in a space where the stakes of AI errors are not measured in conversion rates — they're measured in patient outcomes. The leadership team understood this. What they hadn't resolved was how that understanding should translate into product decisions. There was an implicit tension between the commercial pressure to ship AI-powered features and the clinical and regulatory caution that the product domain requires.

The advisory work was about making that tension explicit and giving it structure. We built a three-tier AI decision framework: AI that operates fully in the background (scheduling, documentation, routing) with no clinical adjacency; AI that supports clinical staff with clear human-in-the-loop requirements; and AI that directly influences patient experience, which required a separate review process and explicit regulatory mapping. Making those tiers visible gave the product team a clear vocabulary for every new AI proposal — and gave leadership confidence that moving fast on the first tier didn't create exposure in the third.

Shuffle-X is building in a media space where AI is simultaneously the competitive opportunity and the identity risk. The question every media organization faces: how much AI is the product before the product stops being yours? For Shuffle-X, this wasn't a philosophical question — it had direct implications for advertiser relationships and audience trust, both of which are the business.

The advisory engagement focused on establishing clear editorial boundaries before the AI roadmap was finalized. We separated AI as infrastructure (recommendation, optimization, workflow) from AI as content (generation, curation presented as editorial). The first category had essentially no internal resistance once the distinction was drawn. The second required explicit commercial and brand guardrails that didn't previously exist. Establishing those guardrails before they were needed — rather than after the first audience backlash — changed the speed and confidence of the product development process.