How we produce this content.
The Veesker blog runs on a deliberate workflow that combines AI drafting with human editorial oversight. Here is exactly how it works — no marketing spin.
- 1
Topic selection — human
The Veesker team picks every topic. Pauta is shaped by our roadmap, recurring questions from customers and contributors, and themes from the wider Oracle and database ecosystem. No topics are auto-generated.
- 2
Draft generation — AI
Each post is drafted by a Claude agent running on a scheduled job — typically Mondays (deep-dives) and Thursdays (manifestos). The agent works from a curated prompt, our editorial guidelines, and reference material we provide. Every draft ships in both English and Portuguese.
- 3
Review and editing — human
Every draft is reviewed by the Veesker team before publication. We check for technical accuracy, factual claims, alignment with our positioning, tone, and voice. We edit, rewrite, or reject as needed. Drafts that do not meet the bar do not ship.
- 4
Publication — human-approved
Posts go live only after explicit approval from the Veesker team. The byline on every post is Veesker — we take editorial responsibility for what we publish, regardless of how the first draft was produced.
Why we disclose this
- It matches our brand. Veesker is an Oracle development platform for the AI agent era. Pretending humans wrote every word would be incongruent with the tools we build and the workflows we advocate for.
- It lets readers calibrate trust. Knowing how the sausage is made is more useful than a generic byline.
- The infrastructure is open. Our publishing pipeline lives in the Apache 2.0 source. Anyone can verify how posts move from draft to live.
Spotted an error?
If you find a factual mistake, technical error, or unclear claim in any post, please open an issue on GitHub. We treat blog corrections the same way we treat code bugs — every report gets reviewed.