The Counter-Article Playbook for Lost AI Citations

Key Takeaways

• Finding out a competitor got recommended instead of you is useless information on its own. The value is in turning it into a specific brief you (or a writer) can execute the same day.
• A real counter-article brief starts from the exact prompt and the exact source that won, not a vague "write about why we're better."
• The strongest counter-content isn't a comparison page trashing the competitor, it's the better, more specific, more current version of whatever source AI actually cited.
• Ship narrow and fast. One sharp post answering the one question you lost beats a broad "why choose us" rewrite that answers nothing specifically.
• Track whether the counter-article actually changed the answer. Publishing isn't the finish line, a re-run showing the citation flipped is.

---

Start from the loss, not from your product

The instinct when you find out ChatGPT recommended a competitor instead of you is to write something that makes your product look better in general. That's the wrong brief. The right brief starts from the exact thing that happened: this specific prompt, asked this specific way, pulled this specific source, which named this specific competitor and not you.

Everything after that should be narrow and traceable back to that one loss. A broad "why we're the best choice" post doesn't fix a specific citation gap, because it isn't the direct answer to the question that surfaced the gap in the first place.

---

The five things a real brief needs

1. The exact prompt. Word for word, not a paraphrase. "Best AI visibility tool for a solo founder" and "how do I know if I'm invisible in ChatGPT" are different prompts with different winning content shapes.
2. The engine and date. Answers shift over time and differ by engine, note both so you can tell later if a re-check found a real change or just normal drift.
3. The winning source. The actual URL that got cited. Open it. Read what it says and how it's structured, this is the format you're competing against, not guessing at. If you're building this brief for more than one loss at a time, here's a systematic method for finding these sources instead of hunting for each one by hand.
4. Why it won, specifically. More recent, more specific numbers, a direct answer format, more reviews, a Reddit thread with real community consensus. Name the actual reason, not "they have better content."
5. What you're going to publish. One clear content type: a direct-answer post, an updated comparison page, a Reddit reply in the same thread that won, a review-site push. Pick one, don't hedge across three.

---

What to actually write

The highest-performing counter-content in this category isn't attack content. It's the more useful, more current, more specific version of whatever source won. If a competitor's blog post answered the question with a 2025 list that's now stale, your brief is "write the 2026 version, more current, more specific, with real numbers." If a Reddit thread is what won, your brief probably isn't an article at all, it's a genuine, specific reply in that same thread or a closely related one, since that's literally the source the model trusted.

Avoid the trap of writing a "GetIntel vs Competitor" page as your default response to every loss. Comparison pages help when the buyer's actual question is comparative. They don't help when the buyer's actual question was something else entirely and a competitor just happened to be the one named as the answer.

---

Ship it fast, ship it narrow

The value of a counter-article decays if you sit on it. Once you have the five pieces above, this should be a same-day or same-week publish, not a project. Narrow scope is what makes that possible: you're writing the direct answer to one specific question, not revising your whole content strategy.

---

The step almost everyone skips: checking if it worked

Publishing the counter-article isn't the finish line. The actual test is whether the same prompt, re-run a few weeks later, shows a different answer. Most teams publish and move on without ever closing that loop, which means they genuinely don't know if any of this content strategy is working.

This is the exact loop GetIntel automates: when it finds a lost citation, it drafts the counter-article brief (the specific prompt, the winning source, the reason it won, and a ready-to-ship draft) so you're not building the brief from scratch, and it re-checks the same prompt on a schedule so you find out whether the citation actually flipped.

See your current lost-citation list and get a drafted brief for the ones that matter most.