Just Say No to AI: Why Our Content Will Remain Human-Authored
- Asanka Abeysinghe
- Chief Technology Officer, WSO2
At WSO2, our content is grounded in original thinking and hard-earned lessons from building real enterprise systems with our products. We apply first principles to shape our perspective on our strategy and forward thinking. No model can replace this experience. AI is valuable when used by thoughtful authors, but alone it does not deliver meaningful value to our readers, customers, or community.
Why fully AI-generated content falls short
AI predicts likely sentences but cannot experience decisions, trade-offs, or outcomes. When content relies on models for thinking, it becomes generic. Our readers expect clear insights into what worked, what failed, and why. Polished language is not a substitute for expertise; only firsthand work and verifiable results demonstrate understanding. When content sounds generic, it erodes trust and fails to serve our audience of architects, developers, and technical leaders.
Where AI absolutely helps
AI excels at streamlining the mechanics, allowing humans to focus on ideas. It improves clarity, corrects errors, suggests structure, summarizes lengthy drafts, prompts fact checks, and adapts tone for different audiences. In short, humans author; AI assists.
What “authentic” means to us
Authenticity does not mean casual or unedited. It means our ideas are grounded in real work:
- Real architectures, comparisons, and design decisions from our products and customer engagements.
- Lessons learned from incidents, migrations, and modernizations—what we’d repeat and what we’d avoid.
- Hands-on code, reference designs, and references readers can actually use.
- Clear positions on industry trends based on first principles, not hype cycles.
If a paragraph cannot be traced to a real decision, experiment, or customer outcome, it likely does not belong here.
Our content standard
Before publishing, we check five things:
- The source of truth, what concrete work, data, or experience (first-person stories) it’s based on.
- Explicitness, whether we name the trade-offs, not just the benefits.
- Replicability, whether a reader could reproduce or adapt it.
- Originality, whether it adds something new instead of adding noise.
- Voice, whether it sounds like engineers, architects or technical decision makers talking to peers.
- Judgment, whether the piece shows discernment about when a pattern applies (and when it doesn’t), the risks, constraints, and failure modes.
AI can accelerate steps two through five, but it cannot generate the foundational source of truth and judgment come from humans.
How we (and how we don’t) use AI in content
We use AI to:
- Draft alternate headlines and meta descriptions.
- Improve clarity, grammar, and structure.
- Generate tables, and summaries from our own source material.
- Suggest questions we might have missed in review.
We do not use AI to:
- Invent ideas or experiences we don’t have.
- Generate entire posts without a human author and owner.
- Manufacture benchmarks, references, or customer outcomes.
- Paraphrase competitors or community content and present it as ours.
What readers can expect
When you read WSO2 content, you should gain insight into our work: the architectural decisions behind our products, the realities of running enterprise systems, and practical modernization patterns. You will see actual diagrams, code, and postmortems we have learned from, along with clear attributions for authorship and data sources.
A note to our authors and contributors
Begin with an idea, a blank whiteboard with your notes, diagrams, and data. Use AI to refine your writing and clarify structure. Then review, remove unnecessary details, and add your unique perspective. This final step builds trust and delivers value.
AI amplifies the work of those with authentic insights.