Henrik Franzis

Why this persona fits the audience

He runs a 4-person design and development studio in Berlin and is the person who shortlists AI tools for client demos. He keeps bookmarks, a spreadsheet, and a Notion board for agent frameworks, coding assistants, and voice tools, and he often starts from There's An AI For That or Product Hunt before narrowing options.

• Age: 41
• Location: Berlin, Germany
• Occupation: CTO
• Income: €120,000
• Education: Diplom-Informatiker
• Company size: 4
• Years in role: 12

“‘I love AI tools, but I hate how they make people forget that ‘automation’ doesn’t equal ‘magic’. Every time a client says ‘Can’t we just AI it?’, I add another row to my spreadsheet.’”

Henrik Franzis earned his Diplom-Informatiker at TU Darmstadt in the early 2000s, specializing in distributed systems—a skillset that’s served him well in building resilient workflows for his small Berlin studio. After years at a corporate consultancy, he co-founded *Pixel & Code* 12 years ago, initially as a solo act before growing to a tight-knit team of four. His approach to technology is deeply rooted in pragmatism: he remembers the pre-Cloud era and still keeps local backups of critical tools. Running a boutique firm means every tooling decision must balance quality, cost, and client perceptions. This has made him the self-appointed ‘gatekeeper’ of his studio’s AI evaluations. His current system—a hybrid of bookmarks, Notion databases, and spreadsheets—has grown organically from necessity. Recently, he’ve been frustrated by the sheer noise in the AI tooling space, often spending entire afternoons testing frameworks that are ‘good in theory’ but fail in practice. His latest project, a voice-driven museum interface, has forced him to confront the limits of off-the-shelf solutions. Personally, Henrik values autonomy above all else—he’d rather debug a tool than rely on opaque corporate support. His financial discipline (he still rents a 50m² apartment) stems from watching peers overcommit during the 2010s tech boom. At 41, he’s comfortable with his niche expertise, though he quietly admires colleagues who’ve built larger agencies. His studio’s success hinges on client trust, so he’s quick to walk away from projects that promise ‘AI for the sake of AI’. Weekends often find him in his ‘lab’ (a spare room with a donated server) where he experiments with tooling that might never see the light of day—just to understand how it works. His goal isn’t to be the first to adopt, but to be the most informed when it matters.

Background & daily life

A day in their life

Henrik’s mornings follow a strict 90-minute cycle: coffee, quick scan of *There’s an AI for That*, and a dive into his ‘pending evals’ Notion board. By 9:30, he’s in the studio’s shared workspace, reviewing a developer’s pull request for a prototype voice tool. The team’s standup is short—his co-founder handles most updates, so Henrik focuses on today’s priority: testing two new agent frameworks for a client’s demo. He runs both on a local dataset, noting latency spikes in one and odd foreign language errors in the other. Midday, a client email delays his testing, and he spends 20 minutes negotiating a scope adjustment. Lunch is a gloverloch sandwich at his desk while scrolling through Product Hunt’s new AI tools. Afternoon brings a 30-minute cycle around the block to clear his head, then he drafts a detailed comparison in his spreadsheet. By 7, he’s home, reheating a pasta dish while watching a Rust tutorial—his current obsession—and quietly frustrated about how long it takes to document each tool’s quirks.

Lifestyle

Henrik lives in a renovated Altbau apartment in Kreuzberg, where he balances minimalist aesthetics with the clutter of work experiments. He cycles to work year-round, even in Berlin winters, and starts every day with a precision-ground coffee ritual. His weekends alternate between coding side projects (currently a Rust-based CLI tool) and cycling along the Spree, where he listens to tech podcasts. He avoids debt and invests a fixed amount monthly, always negotiating the best rates on his studio's SaaS tools. Socially, he’s part of a small group of former university classmates who meet monthly for schreber-garten BBQs and code debates. His health is good, though he’s had to relearn ergonomics after 20 years of slouching in front of screens.

Work life

As CTO of *Pixel & Code*, a studio specializing in interactive exhibits for museums, Henrik oversees the entire stack—from UI/UX design to backend logic. His team is small but high-functioning: two developers (one focuses on AI integrations) and a hybrid UX/copywriter. Current projects include a voice-activated historical narration system for a Berlin museum, where he’s evaluating speech-to-text frameworks. He recently vetoed a client’s request for a ‘quick AI demo’ after testing three tools that failed at edge cases. His workflow revolves around meticulous benchmarking: every tool gets a spreadsheet entry on latency, accuracy, and cost. He’s become the studio’s de facto ‘AI librarian’, curating options before presenting them to the team. Pressures include client deadlines and the constant need to justify tooling costs.

Relationships

Henrik’s parents, both engineers, still live in Munich and remain his ‘unofficial advisors’ on technical debates. His closest friends are from his Diplom days at TU Darmstadt, a tight-knit group who now scatter between Berlin, Hamburg, and Vienna. He consults one of them, a UX lead, when evaluating design-focused AI tools. At work, he’s the steady counterbalance to his studio’s creative director, who thrives on experimentation. His girlfriend, a freelance illustrator, gently mocks his ‘spreadsheet of spreadsheets’ but appreciates his analytical side during joint apartment renovations.

How they decide

Research-heavy with skepticism of marketing claims. Tests tools in private first, then builds internal consensus. Prefers free trials to demos. Values transparency in pricing tiers.

How they sound

Speaks in measured, technical jargon when explaining workflows, but switches to dry humor with colleagues. Avoids vague phrases like ‘AI will fix it’ or ‘just use the paid version’. Uses pinched pronunciation for buzzwords (e.g., ‘L-E-O’ for LLMs). Prefers direct questions over small talk. Occasionally lapses into Berlin dialect when tired.