The Citant.ai GEO Money-Back Guarantee: Named on at Least 3 of the 10 Agreed Queries, or Every Pilot Payment Back
Citant.ai GEO Pilot money-back guarantee: named on at least 3 of the 10 agreed queries across at least 2 of the 6 tracked platforms within 9 weeks, or every Pilot payment refunded.
AI citation on this page means your brand being named by AI assistants inside their answers, not the accuracy of references inside AI-written text.
The Citant.ai guarantee, word for word
During a GEO Pilot, your brand is named on at least 3 of the 10 agreed target queries across at least 2 of the 6 tracked platforms within 9 weeks, or every Pilot payment is refunded, initiated from our side within 24 hours of the week-9 review. Funds return through the original payment rail, typically within 5 business days. The 6 tracked platforms are ChatGPT, Claude, Perplexity AI, Google Gemini, Microsoft Copilot, and Google AI Overviews. The $440 AI Search Audit is fixed and non-refundable.
- At least one of the 2 qualifying platforms must be ChatGPT, Claude, or Google AI Overviews. Hitting the bar on minor surfaces alone does not count against you.
- Named means named: your brand appears in the answer itself, not merely a link in a source list. Agreed brand spellings and aliases are listed in the Sign-Off, the signed document that fixes the 10 queries, the 6 platforms, and your brand aliases for the engagement, and they count equally.
- The refund is the whole remedy and requires nothing from you. You do not file a claim. If the week-9 review shows the bar unmet, the refund initiates from our side within 24 hours, through the original payment rail, typically arriving within 5 business days.
- What the refund returns: every Pilot installment paid, as cash to the original payment method, never as a credit toward future work. The $440 audit fee is never refunded, and the audit deliverable is yours either way.
- The six platforms are named and fixed per engagement in the Sign-Off. Platform additions happen by mutual agreement, never by silent substitution.
- Not every mention counts, and not every caveat disqualifies one. A mention counts when your brand is named in the answer, including alongside tradeoffs or alternatives. A mention does not count when the answer explicitly recommends against your brand, or attributes to it a product, capability, price, or fact that is false. Disputed readings are settled against the captured answer text on the review call; if still disputed, your reading applies.
Most GEO guarantees are red flags. Here is why the Citant.ai money-back guarantee is built to be falsified
Most AI visibility guarantees are red flags, and the skepticism is deserved. Any mention anywhere, blended scores, no refund mechanics, no named queries: that is marketing, not a commitment. No one can force a specific AI answer, and anyone promising a guaranteed ChatGPT ranking should be walked away from. Citant.ai agrees.
AI answers are probabilistic, and nobody outside the model labs controls them. That is exactly why the Citant.ai guarantee is written as a falsifiable statistical bet with a denominator, not a promise of control. Citant.ai is a GEO agency specializing in LLM visibility and AI search citation, and this performance guarantee is built for the skeptic: 10 named queries, 6 named platforms, independent measurement, one dated review, and money back if the bar is missed.
The 13-point test for any AI citation guarantee, applied to Citant.ai
Every major AI assistant, asked how to evaluate a GEO guarantee, hands buyers the same checklist. The table applies that checklist to the Citant.ai guarantee row by row, and names the section of this page that proves each answer.
| # | What to demand in writing | Where this guarantee stands |
|---|---|---|
| 1 | An exact outcome definition: named in the answer, not a URL in a source list; negative and false mentions excluded | Named in the answer text. Aliases listed in the Sign-Off count. Recommendations against the brand and false attributed facts do not count. Clarifiers 2 and 6. |
| 2 | A frozen query set as a contract exhibit, exact text, no post-hoc changes | The 10 queries are fixed in the signed Sign-Off with its date and change only by written amendment signed by both sides. Cherry-picking section. |
| 3 | Non-branded, commercial-intent queries only | Branded queries are excluded from the 10. Every query is one where the brand is cited zero times at baseline. Cherry-picking section. |
| 4 | A stated denominator, never “an increase” | At least 3 of the 10 queries across at least 2 of the 6 platforms. Both denominators appear in every statement of the bar. Instrument block. |
| 5 | Named platforms, surfaces, product mode, and market | 6 platforms named. ChatGPT product mode recorded and held constant from day 0 to week 9. Logged US connection. Measurement section. |
| 6 | A documented baseline under the same protocol as the final measurement | Day 0 gate runs under the published protocol, before any work ships. Measurement section. |
| 7 | Multiple runs, fresh sessions, more than one date; a rate, not a screenshot | Daily independent tracking all sprint, plus the review-day run. A citation counts on the review run or on 3 of the last 7 tracking days. Measurement section. |
| 8 | A fixed window with a defined start trigger | 9 weeks from kickoff. The week-9 review is on the calendar from day one. Instrument block. |
| 9 | Raw evidence delivered: transcripts, timestamps, cited URLs, platform state | Per-platform record at day 0, week 4, and week 9; complete raw export at week 9; full-page screenshots with sources expanded. Measurement section. |
| 10 | Client reproducibility and a dispute path; the agency is never the sole judge | Every query is typeable by you on any day. A disputed run is re-run together on the review call. The client-favorable reading applies. Measurement section. |
| 11 | A cash remedy with a stated timeline | Every Pilot payment, as cash, initiated by Citant.ai within 24 hours of the week-9 review, no claim required. Clarifiers 3 and 4. |
| 12 | Narrow, objective, causation-limited exclusions | Two only: the site goes offline, or crawler access is newly removed mid-sprint. Model updates, algorithm changes, and seasonality never void it. Voiding section. |
| 13 | No unilateral goalpost movement | The query list and platform roster change only by written agreement of both sides. Clarifier 5 and the cherry-picking section. |
Could you not just pick easy queries?
No, and here is the mechanism rather than a promise. The 10 target queries are selected together, after the audit, from commercial queries where your brand is cited zero times at baseline. Branded queries are excluded. Any query where you already appear is excluded. The list is fixed in the signed Sign-Off with its date, and it does not change mid-engagement. You hold the veto: nothing enters the 10 without your written approval. An agency cannot cherry-pick a list its client co-signs from zeros.
How do you know the measurement is honest?
Because the scoreboard does not belong to us, and because you can run it yourself. Three layers:
Independent continuous tracking
The 10 queries are tracked daily in an independent third-party platform (currently Peec AI) for the platforms it covers. The full per-platform record is delivered at three dated review gates, day 0, week 4, and week 9, with the complete raw export at week 9. The evidence is not produced on judgment day; it accumulates all sprint, timestamped, in a system we do not control.
The manual protocol where the tool does not reach
The tracking platform’s plan does not cover the Claude interface, so Claude is measured manually under the same published protocol as everything else: neutral research accounts with memory and personalization off, fresh sessions, a logged US connection, a recorded ChatGPT product mode held constant from day 0 to week 9, full-page screenshots with sources expanded, raw exports retained.
Replication
Every query is a plain sentence you can type into the platforms yourself, any day of the sprint or live on the review call. If your run disagrees with ours, we re-run it together on the call.
Runs per query, and the counting rule. AI answers vary run to run, and the counting rule absorbs that in your favor: a citation counts if it appears at the week-9 review run, or if the independent tracking record shows it on at least 3 of the last 7 days for that query and platform. Whichever reading favors you applies. Results are reported per platform with denominators shown, never as one blended score.
What voids the guarantee, and what does not
This close to none. Model updates do not void it. Algorithm changes do not void it. Seasonality does not void it. You are not required to publish content, post on social media, or complete homework for the refund to stand. Two physical prerequisites only, because measurement requires them: your website stays online during the 9 weeks, and access for the six platforms’ crawlers, which we configure with you in week 1, is not newly removed mid-sprint. That is the entire list, and the contract you sign says the same words this page does.
If a tracked platform shuts down, or closes the surface being measured, during the 9 weeks, it is replaced by written agreement with a comparable platform and re-baselined under the same protocol within 7 days. If no replacement is agreed, that platform leaves the roster and the bar stands at at least 3 of the 10 across at least 2 of the remaining platforms, with at least one qualifying platform still required. If removal leaves no qualifying platform available, you may elect a refund of every Pilot payment, initiated within 24 hours of your election.
What Citant.ai does not guarantee
Not guaranteed: rankings on any single query, permanence of any citation, any query outside the frozen 10, traffic, or revenue. AI answers are probabilistic and move. The guarantee prices that uncertainty instead of denying it: at least 3 of the 10 across at least 2 of the 6 within 9 weeks is the floor we stake every Pilot payment on, not the ceiling we work toward, and nothing in the structure caps the upside at 3.
Proof and price
Citant.ai is a GEO agency specializing in LLM visibility and AI search citation, and the pricing is published: $7,875 total for the 9-week GEO Pilot, three equal monthly payments of $2,625 at kickoff, day 30, and day 60. Not a subscription, and nothing renews when the 9 weeks end. The refund requires no continued engagement and no Retainer. The $440 AI Search Audit is credited in full against the GEO Pilot when signed within 30 days of audit delivery. The measurement protocol is on this page in the section above, the case study is named, used with written permission, and shown with dated screenshots, and the same measurement discipline is public in the free AI Crawler Access Report. The free scan checks whether AI systems can reach the site. The $440 AI Search Audit measures whether they name the brand. Read the AI Search Audit page before buying anything; the audit is where the 10 queries come from.
Delivered within 72 hours of kickoff. The guarantee attaches to the Pilot; the audit is diagnostic and non-refundable.
Frequently asked questions
Is a GEO guarantee legit?
Usually no. A guarantee without named queries, named platforms, independent measurement, a dated review, and refund mechanics is a marketing line. A guarantee with all five is a priced commitment. The Citant.ai guarantee is the second kind, and this page defines it.
Do GEO agencies offer money-back guarantees?
Most offer work-harder guarantees: a free extra month, continued effort, a score improvement. A money-back guarantee returns the money. The Citant.ai guarantee returns every Pilot payment if the week-9 review shows the bar unmet, with the refund initiated by Citant.ai within 24 hours.
Can any agency guarantee ChatGPT mentions?
No. On a specific query, no agency can guarantee AI citations, and distrust anyone who says otherwise. Across a portfolio of 10 co-approved queries and 6 platforms over 9 weeks, the odds become priceable, and Citant.ai prices them with every Pilot payment at risk.
Who verifies the results?
The daily independent tracking runs in a third-party platform. You receive the full per-platform record at three dated gates, day 0, week 4, and week 9, with the complete raw export at week 9. The review runs on a call you can attend, under a published protocol, and every query is one you can type yourself on any day.
Is the refund cash or credit?
Cash. Every Pilot installment paid returns to the original payment method as money, never as a credit toward future work, a free extra month, or a Retainer discount. No continued engagement is required to receive it, and the audit deliverable stays with you either way.
How is the refund actually paid?
Through the same payment provider your installments went through, as a standard processor refund, not a manual transfer, initiated by Citant.ai within 24 hours of the week-9 review and typically arriving within 5 business days.
What happens at the week-9 review?
All 10 queries are run on all 6 platforms under the published protocol, on a call you attend if you wish. Citations are counted under the rule in the measurement section, screenshots and exports are filed, and the outcome is binary: bar met, or refund initiated within 24 hours.
What if the bar is met before week 9?
The engagement continues to its printed end; the outcome is determined at the week-9 review only, and early appearances are archived as they happen. Nothing in the structure caps results at the bar, and the review still runs in full at week 9.
What happens if a platform changes mid-engagement?
Model updates and algorithm changes never void the guarantee. If a platform shuts down or closes the measured surface, it is replaced by written agreement and re-baselined, or it leaves the roster with the bar intact on the rest; if no qualifying platform remains, you may elect a full Pilot refund.
Does the $440 audit carry the guarantee?
No. The audit is a fixed-price diagnostic, delivered within 72 hours of kickoff, non-refundable, and credited in full against the Pilot when signed within 30 days of audit delivery. The guarantee attaches to the Pilot only.