Send the answer that worries you
Use the form for AI search visibility audits, citation reviews, German-English source consistency checks, and monthly observation work. The best starting point is a real query, a real answer, and the business record you expected the engine to understand. A short note is enough if it includes the company, the market, and the claim that feels wrong.
Frequent questions
What do you usually work on?
I usually start with a small set of real AI search queries around a German SME, product category, supplier search, or comparison phrase. I save the answers, inspect the cited sources, and map which claim came from which public page. From there I write a repair plan for pages, profiles, directories, and source language.
Which topics do you take on?
I take on German SME visibility, B2B supplier descriptions, export-facing company profiles, product and service categorization, citation share, and German-English evidence mismatches. I am most useful when a company is already visible somewhere, but AI search is describing it with the wrong role, geography, category, or proof.
How fast do you usually reply?
I usually reply within two working days. If the request is messy, I may answer with a narrow first read instead of a formal proposal. A useful first reply should tell you whether the problem looks like source selection, claim support, language mismatch, or entity confusion.
How are consultations formatted?
I usually work from saved answer records, source notes, and a written findings document. A consultation can be a one-off audit, a deeper source repair plan, or continuing observation where the same query groups are checked over time. Calls are used when they help explain the evidence.
What does pricing usually look like?
Project work usually ranges from $3,800 to $14,000, depending on the number of query groups, engines, languages, and source types reviewed. Monthly observation usually ranges from $2,200 to $6,500 and depends on how many query groups, engines, and languages are watched over time.
What do you not take on?
I do not write fake reviews, manufacture citations, hide bad evidence, or promise a specific ranking inside an AI answer. I also do not treat a single good answer as proof that the issue is solved. If the public record is thin or contradictory, I will say that before recommending edits.
Start with the sentence the machine already repeats.
If we can find the source behind it, we can decide what deserves repair.
Contact Mira