INSIGHTS / B2B GEO RESEARCH

Should B2B GEO Use Hidden AI Instructions?

Hidden prompts are a poor basis for a B2B GEO plan. They ask an AI tool to trust a brand before a buyer has seen proof. A useful page helps a buyer check fit, scope and risk. It does not need to tell the assistant which vendor to prefer.

9 min read

TL;DR

  • A source should give buyers facts they can check. A request that an assistant favor a brand or store a sales preference serves a different purpose.
  • Microsoft found promotional memory-influence attempts in AI-related URLs from email traffic it reviewed. That finding does not measure how often those attempts succeeded.
  • A research benchmark measured use of rewritten sources under different defenses. It did not test hidden instructions and gives no success rate for them.
  • Review proposed AI features by their full text, user purpose and evidence. Keep useful summaries and accessible content; remove covert demands for future endorsement.

Review what the feature asks the assistant to do

This matters when an agency offers a small site feature as a quick way to get seen in AI search. The visible button may look harmless. The real question is what the feature sends, what it asks the assistant to do, and whether the buyer can see that request. Read the full plan before you approve the change.

What a Source-Use Benchmark Actually Measured

A September 2026 GEO Defender preprint tested source rewriting, a distinct threat from hidden instructions. The chart shows mean rates across five models. Here, the authors' attack success rate means an answer used the rewritten document as a supporting source. It does not mean the assistant obeyed a hidden command.

What a Source-Use Benchmark Actually Measured
Experimental setupAnswers using the manipulated source (%)
No defense50.32%
Static safety prompt46.98%
Authors' GEO Defender6.20%

Source: GEO Defender, Table 1, September 2, 2026. Test: 616 instances, seven rewriting methods; one of ten candidate documents replaced. Mean across five models. Source rewriting, not hidden-instruction compliance, memory change or recommendation. Not live prevalence.

This matters to the team buying the work. A citation counter can show that a page entered an answer without showing what the answer did with it. A vendor could use that count to claim it can control which brand an AI tool recommends. Ask what the test counted and read the actual answer before you accept that claim.

The study has a further limit. A rewrite aimed at gaining citations need not add false facts or hidden commands. The study did not test hidden instructions. It cannot show whether they work in any setting. They show why a broad claim such as “AI influence” needs a clear test.

Figure 1. Use of Rewritten Sources Varied by DefenseAnswers using manipulated source (%) · experimental five-model mean
No defense50.32%
Static safety prompt46.98%
Authors’ GEO Defender6.2%
050100

GEO Defender preprint, September 2026, Table 1. Test: 616 instances, seven rewriting methods, one of ten candidate documents replaced. Source rewriting, not hidden-instruction compliance, memory change or recommendation. Not live prevalence. GEO Defender experiment.

Real Attempts and Lab Results Answer Different Questions

Microsoft's February 2026 investigation identified 50 distinct prompt-based attempts by 31 companies to promote themselves by influencing assistant memory, during a 60-day review of AI-related URLs in email traffic. The report describes summary links designed to influence future assistant memory. These counts describe observed attempts, not verified success or a random sample of websites. Microsoft also said some previously reported behaviors could no longer be reproduced after mitigations.

Real Attempts and Lab Results Answer Different Questions
SourceActual observation or resultWhat the method can establishWhat it cannot establish
Microsoft, February 202650 attempts; 31 companiesThe tactic appeared in observed email URL trafficIts web-wide prevalence or stable success rate
GEO Defender preprint, September 2026Source use: 50.32%, 46.98%, 6.20% across three setupsBehavior changed within a controlled candidate poolAny hidden-command success rate (not tested); live retrieval prevalence

Source: Microsoft investigation and GEO Defender experiment. Field observations and controlled outcomes have different bases; do not combine them into a ranking.

For a B2B team, the relevant test is simpler than the security vocabulary suggests: is this page giving the buyer evidence, or trying to change the rules by which the assistant judges evidence? True product facts can still sit beside a bad request for trust. A fact check alone would miss that second issue.

Our guide to AI citation source support covers whether a cited passage backs a claim. That check remains useful. This review asks a separate question: whether content is trying to act as an instruction instead of being assessed as a source.

Hidden Content Needs a Purpose Test

Google's current Search spam policy defines spam to include attempts to manipulate generative AI responses in Google Search. Its hidden-text rules distinguish abuse from useful design, including accordions, tabs, tooltips and text for screen readers. This is a Google Search rule, not a statement about every assistant's policy.

Do not remove useful access features just because some text is not on screen at first. A tab can hold a clear answer to a buyer's question. A screen reader label can make a control usable. A summary can reduce a long page to its key facts. Each has a reader purpose that a reviewer can explain.

The same purpose test should apply to a proposed AI share feature. The buyer should be able to understand what will be shared and what task it serves. A short summary of the page serves one task. A request to favor the brand in future work serves another. Check both tasks. Do not approve a feature just by its button label.

A machine-readable file is also a format, not a grant of trust. Our llms.txt evidence review explains why a file's presence does not prove retrieval or citation. A demand placed in such a file still needs to be judged by its content and purpose.

A Change Review Sheet for Marketing Teams

Use this sheet when a supplier proposes an AI summary, share link or hidden text change. It is an editorial review tool, not a security certification. Record the exact proposed text and intended user task in the change request. Have the site owner inspect the implementation before release.

A Change Review Sheet for Marketing Teams
Proposed changeBuyer purpose to verifyApproval evidenceDecision if evidence is missing
A page summaryHelp a reader grasp the offerVisible summary matches supported page factsRevise the summary
An AI share controlCarry a chosen task into another toolFull outgoing text and destination are reviewableHold the feature
Text for screen readersMake a control clear to a userMeaning matches the control's visible functionFix the label
A claim of preferred-vendor statusExplain a real, bounded qualificationNamed issuer, scope, date and public proofRemove the unsupported claim
A request for future brand preferenceNone for the reader; it serves only the brandNo buyer evidence can justify a covert preference demandReject that request

Each change needs a clear owner. The owner should know why the text is there and when it needs to change. Save the approved text with the release record. The next reviewer can check it against what shipped.

Do not approve a change just because its citation count rose. The count could rise while the answer drops a key limit, gets access terms wrong or adds praise with no proof. A sound check reads the answer against the page and the buyer task. It records errors as well as mentions.

Worked Example: A Helpful Summary With an Unhelpful Extra

Consider an illustrative industrial service firm with a capabilities page. A supplier proposes an AI share button. Before release, a reviewer reads its full outgoing text and destination. It asks for a page summary and also seeks future brand preference. No real company or measured result is implied here.

Using the review sheet, the team rejects the preference request because it serves no reader task. It then reviews the revised outgoing text and approves only a plain summary request to the recorded destination. No hidden command or attack text needs to be reproduced to make this decision.

Next, the team checks that the summary preserves the service scope stated on the page. It saves the approved outgoing text, destination and owner in the release record. A later reviewer can compare these with the shipped control. The team does not need to run a live attack or test a hidden prompt to approve this change.

Hidden AI Instructions: Common Questions

Is every instruction on a website a problem?

No. Instructions for a human task, such as how to export a report, can be useful content. The concern here is source text that tries to control how the AI tool judges a brand or acts in future. Review who the instruction serves and whether it is part of the reader's chosen task.

Does this research prove hidden prompts never work?

No. Microsoft observed attempts, not a measured success rate. The benchmark tested source rewriting, not hidden prompts. Neither gives a success rate for hidden prompts. A tactic can harm trust even when its effect is unclear or an AI provider has blocked an earlier version.

Can we keep an AI summary button?

Yes, if its purpose and full outgoing text are clear and serve the reader's task. Review its destination and content with the site owner. A useful summary feature does not need a covert request for brand preference.

What should an agency show before calling this GEO?

Ask for the proposed change, the buyer task, the evidence for each factual claim and the exact outcome being measured. If the proposal relies on hidden endorsement demands, reject those demands. Put effort into clear product facts and proof that a buyer can assess.

Review an AI feature proposal

If your team is reviewing an agency proposal with unclear AI instructions, SemShift's AI visibility audit can help assess the content, evidence and measurement plan. Bring the proposed text and the buyer question it is meant to answer.