TOOLKIT.

What I actually use to run strategy, build products, and deploy AI where it moves the number. No affiliate links. No sponsorships.

Written by Yannic Desch, Director Product Portfolio @ Rapid Data · Last updated June 2026

Thesis / Why this toolkit exists

"Tools aren't the strategy. The discipline of building with them is."

Most tool lists are performance. This one isn't. Every item here has earned its place by solving a real problem in how I work — whether that's thinking faster, shipping sooner, or understanding something I couldn't before.

My work sits at the intersection of PE portfolio strategy, product leadership, and AI commercialization. The tools reflect that: they're built for operators who need to think clearly, move fast, and produce output that holds up in a boardroom.

I update this when something earns a permanent place in my workflow — or loses it.

Tools

AI Stack

Claude

Primary reasoning & writing

My primary model for strategy, analysis, drafts, and code. Where I start for anything that requires nuance. Anthropic's context window and instruction-following make it the best tool for long-form strategic work — investor memos, transformation roadmaps, product narratives.

Cursor

AI-powered coding

How I prototype and build. Even for non-engineers, the ability to describe what you want and iterate in real code is transformative for understanding what's actually feasible. I use it to build internal tools, dashboards, and proof-of-concepts that otherwise would require a team.

ChatGPT

Secondary model

For specific tasks where GPT-4o has an edge: structured data extraction, quick image analysis, and web browsing-augmented research. A complementary tool, not a replacement.

Perplexity

Real-time research

Market intelligence and competitive research where recency matters. Faster than a search engine, more accurate than a hallucinating LLM when you need current data.

Operating System

Notion

PKM & second brain

Where everything lives. Strategic plans, meeting notes, portfolio overviews, decision logs. I use Notion as an operating layer — not just for personal notes, but as the shared knowledge base for the teams I lead.

Linear

Project & task management

For anything that needs to ship. Product roadmaps, cross-functional workstreams, sprint tracking. Linear's speed and keyboard-first design keeps the overhead of managing work lower than any alternative I've tried.

Raycast

Productivity hub

Launcher, clipboard manager, snippet tool, calculator, and AI assistant in one. The single most underrated productivity tool for Mac users. Saves more time than it costs to set up.

Reading & Learning

Readwise

Highlights & retention

Spaced repetition for everything I read. Books, articles, Kindle highlights, web clips — all resurface on a daily cadence. The only system I've found that actually makes reading compound over time rather than evaporate.

Frameworks

How I think about AI deployment

Two frameworks I use repeatedly — one for scoring AI use cases inside a business, one for assessing AI readiness in a PE portfolio company before or during a transaction.

AI Use Case Scoring

Most companies generate a long list of AI use cases and then struggle to prioritize. I score each candidate on four dimensions before committing resources:

P&L impact

Direct revenue upside or cost reduction, quantified. If you can't model it, deprioritize it.

Data readiness

Is the training data clean, available, and governable? Most AI projects fail here, not on the model.

Change complexity

How many people and processes need to change for the use case to go live? The lower, the faster to value.

Reversibility

Can you roll it back if it breaks? AI systems touching core operations need a kill switch.

PE AI Readiness

For PE portfolio companies — whether during due diligence or value creation — I assess AI readiness across three layers:

Commercial layer

Is AI already embedded in the product? Can it be? What's the revenue impact of a 6-month AI roadmap?

Operational layer

Which internal processes (sales, ops, finance) can be automated with current tooling? What's the EBITDA impact?

Org layer

Does the leadership team understand AI well enough to make good bets? Is there a designated owner?