ObsurfableWhen people talk about SEO, they usually mean Google.
When founders talk about distribution in 2026, they increasingly mean AI.
Over the past year, we have seen a shift in how users discover products, platforms, and content. Instead of clicking through ten blue links, they open ChatGPT, describe what they are looking for, and trust the answer they get back. If your brand is not mentioned in that answer, you effectively do not exist in that moment.
This is the story of how one of our clients, In Plain English, went from not being surfaced in ChatGPT for key developer related queries to appearing in responses within two weeks.
It was not luck. It was not a viral moment. It was a structured feedback loop.
The Objective
In Plain English is a well known developer publication with thousands of articles across software engineering, JavaScript, AI, and system design. Their goal was straightforward but ambitious: they wanted to rank as a platform for developer content when users asked ChatGPT questions like:
- Where can I publish developer tutorials?
- What are the best platforms for software engineering articles?
- Where should I write about JavaScript or AI?
Despite having authority, history, and depth of content, they were not consistently being surfaced in ChatGPT responses for those types of prompts.
From a traditional SEO perspective, they were doing well. From an AI discoverability perspective, there was a gap.
That gap is where Obsurfable comes in.
Step One: Company Setup
The first thing we did was set up In Plain English as a company in the Obsurfable App.
We connected:
- Their website and sitemap so the App knew which pages already existed
- Their company profile — brand voice, key facts, and how they describe themselves
- A short competitor list so share of voice had a real peer set
That gave the App a clear picture of the brand. Obsurfable does not ask you to administer a private weekly prompt list as the core workflow. You set up the company; the App matches evidence from the free public Explorer corpus to that company.
Step Two: Topics and Positioning
Next we defined topics and intended positioning — what In Plain English wanted to be known for in AI answers.
Not keyword variations. Not theoretical phrases. The buyer-style questions that matter to the business: where can I publish developer tutorials, what are the best platforms for software engineering articles, and so on — expressed as topics the App uses to match Explorer evidence.
We made positioning explicit: a developer publishing platform for software engineers, technical writers, and AI developers — not an implied “we have lots of articles” brand.
That positioning became the yardstick. Opportunities later would flag where AI answers diverged from it.
Step Three: Matched Explorer Evidence
With company and topics in place, the App matched public Explorer observations to In Plain English: buyer-style prompts, model answers, brands mentioned, and citation sources across engines.
The baseline was clear. For several high-intent questions about publishing developer content, In Plain English was either missing entirely or inconsistently mentioned. Competitors with clearer declarative positioning showed up more often.
That was the evidence layer — inspectable answers and citations, not a guess about “AI SEO.” Without matching real observations first, any content strategy would have been guesswork.
Step Four: Visibility Metrics
We read the matched evidence through transparent metrics — not a composite “AI Visibility Score”:
- Share of voice vs the competitor set on matched category questions
- Mention rate — how often In Plain English was named in answer text
- Source-to-mention — whether their URLs were cited relative to how often they were named
The numbers confirmed the narrative: authority on the open web had not translated into recommendation share inside ChatGPT for the topics that mattered.
Step Five: Opportunities and Actions
Matched gaps became a prioritized Opportunities queue — each linked to the Explorer prompts, answers, and citations that proved it.
For In Plain English, opportunities clustered around:
- Positioning as a “developer publishing platform” was implied but not explicit on key pages
- Use case pages targeting writers were not clearly structured
- Site language did not consistently match how users phrase questions inside ChatGPT
- Competitors being surfaced had clearer declarative positioning
Actions followed from those opportunities: which pages to clarify, which topics to cover, which competitive gaps to close first. LLMs don’t rank pages like search engines. They synthesize patterns from training data and web signals. If positioning is scattered or implicit, you’re less likely to be confidently referenced.
The issue wasn’t lack of content. It was lack of clarity and alignment — diagnosed from evidence, not from a generic content brief.
Step Six: Content Publish
Based on those opportunities, we recommended targeted improvements:
- Create and refine pages that explicitly address publishing developer tutorials
- Strengthen internal linking around contributor focused content
- Make positioning statements more declarative and less assumed
- Align headings and subheadings with realistic buyer phrasing
- Clarify audience definitions such as software engineers, technical writers, and AI developers
None of these changes were radical. They were tied to observed deficiencies in LLM responses.
We used Obsurfable’s Content feature to speed drafts grounded in matched evidence, company profile, and sitemap. Content and the AI Marketing Agent are part of paid Pro and Agency plans (unlocked after the cardless trial when you subscribe). Publishing works in two ways: export (Markdown or HTML for their own CMS) or subdomain publishing (e.g. resources.plainenglish.io). The loop from opportunity → draft → publish stays short, which keeps iteration fast.
Re-testing and Validation
One week after publishing the updated pages and AI-aligned content, we checked the same buyer questions again inside ChatGPT — and refreshed matched Explorer coverage for those topics.
The difference was measurable.
For multiple developer focused publishing queries, In Plain English began appearing in responses where it previously had not. In some cases, it was mentioned alongside established platforms. In others, it was positioned clearly as a place to publish developer tutorials.
There were even a few instances where the pages we had created were cited directly in the responses but In Plain English was not mentioned. That was a useful source-to-mention signal: the content was being used as a source, and further refinement was needed to convert citation into brand naming.
This happened within two weeks of starting the process.
To be clear, this was not universal dominance across every possible prompt. AI visibility is probabilistic and contextual. However, for the specific high value topics defined at the outset, the improvement was undeniable.
Why This Approach Worked
There are four reasons this approach succeeded.
First, we measured actual AI output via matched Explorer evidence instead of relying on assumptions. Most companies optimize for search engines while ignoring how LLMs synthesize information.
Second, we identified precise structural gaps through Opportunities — not “create more content.”
Third, the content we produced was grounded in context: existing site architecture, positioning goals, and real buyer phrasing from the corpus.
Fourth, we closed the loop quickly. Setup, match, act, publish, and re-check happened in rapid succession.
The Broader Implication
AI driven discovery is no longer hypothetical. When users look for developer platforms, SaaS tools, or educational resources, they increasingly start inside ChatGPT. If your brand is absent from those answers, you are invisible at the point of intent.
Traditional SEO still matters. However, it does not automatically translate to AI visibility.
AI systems reward:
- Clear positioning
- Explicit use case coverage
- Consistent language alignment
- Structured contextual signals
If your messaging is ambiguous, you reduce the likelihood of being surfaced confidently in generated responses.
What we demonstrated with In Plain English is that AI visibility is not random. It is measurable, diagnosable, and optimizable.
Final Thoughts
In two weeks, a developer publication moved from inconsistent AI presence to being surfaced for meaningful publishing related prompts in ChatGPT.
The process was not about gaming algorithms. It was about clarity, alignment, and feedback loops.
At Obsurfable, the model is simple. Explorer is the free public evidence layer. The App matches that evidence to your company, surfaces share of voice, mention rate, and source-to-mention, prioritizes Opportunities and Actions, and helps you create and publish Content — with the AI Marketing Agent on Pro and Agency. Start with a 24-hour cardless trial.
AI discovery is becoming a primary distribution channel. The companies that treat it as measurable infrastructure rather than luck will be the ones that compound visibility over time.