about dan

I'm Dan Schoenemann. I write Getting By On AI — plain notes from someone who ships real product work with AI tools, not from someone selling a course about them.

For twelve-plus years I've worked in digital media and adtech: account work, solutions engineering, product, and platform leadership. The through-line is the messy middle between data, activation, identity, integrations, and the teams trying to make money from all of it without breaking trust or governance. I've facilitated both batch and realtime / API integrations — file pipelines and live connectivity — so audiences and data actually move between systems that were never designed to agree.

I currently lead product for a data-activation / syndication platform inside one of the world's largest advertising holding companies — the connectivity layer audiences move through to reach destinations. That includes owning how partners and destinations integrate: batch delivery where it still wins, APIs where latency and feedback matter. I've also built and shipped on tiger-team agentic campaign systems: integration layers agents can call, supply-side activation paths, and the operating rules that keep AI-assisted builds from drifting.

Earlier: product on a global consumer-tech company's onsite search advertising business (preserving value when identifier loss hit targeting); product and solutions work at a major credit-bureau / identity company on retail-media and clean-room integrations, including a realtime API a top-tier retail media network beta-launched for audience syndication into the buy side; and years at a leading demand-side platform across partnerships, product integrations, and major accounts.

I'm not a career software engineer. I am a product person who builds when building is the fastest way to know the truth — APIs, agent tooling, local model setups, the boring scaffolding that makes "AI in the workflow" survivable.

what i'm good at

• Strategy that survives contact with the system. Not decks that dissolve at implementation. I turn ambiguous bets into an operating model: naming the system, drawing domain boundaries, sequencing what to build vs. integrate, and making the next decision obvious for other people.
• Integrations — batch and realtime. Partner and platform connectivity end-to-end: specs, data contracts, API design, file/SFTP-era pipelines, activation protocols, failure modes, and the product choices that keep integrations operable at scale. If two systems have to exchange audience or campaign truth, I've lived in that gap.
• Data activation and identity-shaped product. Clean rooms, first-party governance, syndication/connectivity, retail media, measurement that has to survive scrutiny. If audience data has to move and still mean something, I've probably fought that fight.
• AI that ships inside real surfaces. Not "add a chatbot." Agent-ready APIs and integration surfaces, eval and review gates, human-in-the-loop checkpoints, and the documentation habit that stops silent assumptions from becoming production debt. The newsletter is me practicing that in public.
• Cross-functional translation. Product ↔ eng ↔ data ↔ privacy ↔ partnerships ↔ leadership. I write the thing people can execute against — then hand ownership off so I'm not the bottleneck.

why the newsletter exists

Most AI writing is either hype or homework. I wanted a third lane: first-person experiments from actual work — what worked, what flopped, what I'd skip next time. The posts are the lab notes. This page is so you know those notes come from platform, integration, and product seats, not from a weekend prompt hobby.

If the writing is useful, subscribe. If you want help applying the same standards inside your company, that's the advisory work.

advisory / consulting

I advise product, platform, and leadership teams who need strategy and integrations that hold up in production — activation, identity, partner connectivity, and AI operating surfaces — not slideware.

• Integration strategy & design — batch and realtime/API paths between platforms, partners, and internal systems; contracts, sequencing, and operability
• Activation / identity / connectivity strategy — when audiences and data have to move across an ecosystem without fantasy architecture
• Agentic operating-model design — how product and eng actually work with AI (specs, gates, review, ownership), including the integration surfaces agents depend on
• Build-with-you advisory — short, sharp engagements where I help you structure the problem, ship a wedge, and leave you with a system other people can run

I am selective. Best fit is complex, cross-functional, strategically fuzzy work with real integration load. Weak fit is "make us a chatbot" or fixed-scope feature backlog babysitting.