How Forgefy uses AI, and what it refuses to do
AI content tools happily invent a customer who said a thing that never happened. Post that once and you have burned the only asset a small founder has, which is being believable. The rule is simple and enforced in every prompt in the product: the AI may only use what you gave it. Your updates, your revenue data, your articles, your voice notes. If a fact is not in there, it does not appear in the post.
Step 1: Feed the tool real material: updates, customer emails, transcripts, published a…
Feed the tool real material: updates, customer emails, transcripts, published articles. Everything you write here becomes context. The AI quotes you rather than a generic founder.
Step 2: Set the voice once per project. Register, language and the things you never say.
Set the voice once per project. Register, language and the things you never say.
Step 3: Review every draft before it goes out. Drafts land in review, not in the publis…
Review every draft before it goes out. Drafts land in review, not in the publish queue.
Good to know
No invented anecdotes, no fabricated metrics, no imaginary customers. If the input has no numbers, the output has no numbers. Output always follows the project language, even when you type the update in a different one. No em-dashes and no AI marketing clichés. It is stripped after generation, not just requested in the prompt. Revenue posts are built from your synced provider data, so the figure on the image is the figure in your account.
Does it post anything automatically without me?
Only what you approved. Generation creates drafts in review; publishing needs your approval or a schedule you set.
Can I turn the AI off entirely?
Yes. Write and upload manually and the AI never touches your content.
Related