The principles the product was built on.
Hadef AI is software, but the reasons it works are older than the software. This is the playbook we follow internally. Everything the engine does, it does because one of these principles said to.
The six principles
Always
A pitch that could be sent to any guest never gets a reply. A pitch that could only be sent to this guest earns the reply. Generic volume is worse than zero volume.
Lead with their content
Open with something the guest actually said. Not a compliment. Not a show summary. A verbatim line from a real source. Shows you listened.
Find the unexplored
A topic is what they've covered. A thread is what they've gestured at and never developed. Good pitches pull threads. Great pitches pull threads no one else spotted.
Not by default
Some guests live on LinkedIn. Some live in email. Some live on X. Some reply only if you go through a mutual. The channel is part of the pitch, not an afterthought.
Respect the queue
First touch is light. Second touch adds the thread. Third touch adds the artifact. Fourth touch disengages. Every touch leaves room for the guest to say no without you losing access.
Five checks, every time
Quote accuracy. URL liveness. Specificity. Quote ownership. Confidence. Any pitch that fails any check gets rewritten. No exceptions for famous guests or urgent shows.
How the engine encodes the playbook
Every principle above lives inside Hadef as code. Specificity is Check 3 in QA. Quote before ask is the hook quote requirement. Thread over topic is what the concept stage is built to find. Sequence not spam is the default cadence. QA is the five-agent pipeline that runs before any pitch reaches your review queue.
You can run the playbook without the product. Every host who has tried it by hand ends up describing the same bottleneck, which is that the reading takes longer than the writing.
Put the playbook to work
Free tier runs your first three pitches through the full engine. No setup fee. No card.