The three kinds of company behind one label
Monitoring platforms track your mentions across ChatGPT, Gemini, Claude and Perplexity and report trends; you act on the findings yourself. Consultancies and agencies audit and advise, and sometimes execute, on retainers; quality varies widely with the practitioner. Execution platforms, the newest type and where Aethon sits, combine the measurement with the work: content, structure, entity repair and citation earning shipped as part of the product, with the answers themselves as the proof. Most disappointment in this category comes from buying one type while expecting another.
What the work actually is, whoever does it
Strip the branding and every legitimate LLM optimization company does versions of five things: build a prompt set from real buyer questions, measure the answers across the major models, fix entity contradictions and wrong facts at their sources, publish content that answers directly, and earn citations in the sources the models trust. If a pitch cannot map to those five, it is either monitoring alone or vapor.
Costs across the landscape
Monitoring platforms run from free tiers to several hundred dollars monthly. Consultancy retainers typically start in the low four figures. Execution platforms price between the two; Aethon starts at $199/month with the initial audit free. Anchor any decision to the value of one incremental customer that an LLM recommendation brings you per year, which in most B2B and high-ticket categories covers months of any option.
Choosing: the four-question filter
Ask every candidate the same four questions. Do you test natural buyer phrasings across all four major models, or keywords in one? Do you show the citations behind each answer? Who does the fixing, you or us? And can you show before-and-after answer screenshots from a real engagement? The pattern of answers sorts the landscape fast, and anyone guaranteeing specific LLM placements exits the conversation immediately, since model output cannot be bought. The free Aethon audit is a working sample of the category: your questions, measured answers, and the fix list, before any spend.
