Valoh
Frequently Asked
The questions buyers actually ask.
Most of these come up in the first thirty minutes of a scoping call. Answers are short, direct, and consistent with the published methodology.
What Valoh is
Isn't this just SEO for AI?
No. SEO and AEO tools are optimization categories — they tell you what to change to lift your visibility. Valoh is in the measurement category. We tell you what's there. We don't tell you what to do about it.
The cleaner analogy is DoubleVerify, IAS, or Nielsen — third-party measurement vendors that exist deliberately on the other side of the optimization line. SEO didn't have a clean equivalent, which is part of why executive trust in SEO numbers has always been thin.
How is Valoh different from Profound, Peec, Otterly, AthenaHQ, or other AI brand-visibility tools?
Those tools sit in the AEO / GEO category — they measure visibility and sell the optimization. Some publish proprietary scores. Most include a remediation engagement or content advisory layer. That's a legitimate product category; it just isn't measurement, by the same standard the MRC applies to programmatic.
Valoh has no optimization product, no managed service, no recommendation engine, and no proprietary aggregate score. Reports contain observed values and stop there.
Will Valoh tell us how to improve our brand presence?
No. That's the line we hold. The moment a measurement vendor starts selling lift, the measurement is no longer trustworthy — the vendor benefits when the score goes up.
If you want optimization advice, hire an AEO vendor or a brand consultancy. We will measure the result, before and after, and that measurement will be defensible because we have nothing to sell on the optimization side.
Methodology questions
Which models do you query?
ChatGPT, Gemini, Claude, and Perplexity, queried independently via API at default settings. Each is reported separately — we do not average values across models.
How are prompts constructed?
Prompts are category-anchored, not brand-anchored — the model is asked about the category, not about your brand. This produces unsolicited mentions, which is what real customer queries look like. Prompt sets are constructed up front, locked for the measurement window, reviewed every six months, and versioned in the report.
Can we influence the prompt set?
No, not during a measurement window. Independence is a methodological requirement; if clients can adjust prompts mid-window, the measurement is no longer comparable to the prior window or to other clients in the category. That's the kind of soft compromise that erodes measurement-firm credibility quickly.
Prompt sets are reviewed publicly between windows, and we welcome category input then.
How often do reports update?
Daily. Every signal is sourced from a logged conversation captured within the prior 24 hours.
How do you validate sentiment and recommendation strength?
Both signals are double-coded. An automated classifier scores every response. A human reviewer codes a 10% stratified sample against a published rubric. Disagreements are resolved and inter-rater agreement (κ) is reported per category per quarter.
Can we audit the underlying transcripts?
Yes. Raw transcripts are retained at full fidelity for 90 days. Clients may request the conversations behind any reported signal value at any time within that window. This is how the report stays defensible to a CFO or an MRC-track auditor.
Engagement questions
Is there a free version?
Yes — a free 7-day pulse. Same five signals, same four LLMs, same methodology, abbreviated to one week. Delivered as a pulse report. We don't pitch optimization at the end. Start your pulse.
How is pricing structured?
Per brand, per category, per measurement window. Two tiers: a 90-day Initial Measurement Window, and Continuous Measurement on an annual cadence with quarterly attestation. Pricing is quoted after a 30-minute scoping call. List prices aren't published because category breadth and competitor-set size move them by enough that the list would be misleading.
What does the scoping call cover?
Brand, category, prompt-set construction, competitor set, retention requirements, delivery cadence. We will not pitch optimization on that call. The output is a quote and a measurement window.
Will Valoh present findings to our team?
One scoping call up front, and one delivery walkthrough at the end of the first window if requested. We don't sell ongoing strategy time on top of the report — that's the consulting wedge that pulls measurement vendors into optimization.
Less common, still important
What about user-personalized AI behavior — memory, custom instructions, prior conversations?
Valoh queries via API at default settings, so personalized model behavior is not captured. This is documented in the Limitations section of the methodology paper. If personalization becomes a meaningful share of consumer AI use, we'll publish a separate methodology for it rather than fold it into the core measurement.
Models update without notice. How do you handle that?
We log the model identifier returned with every response, so version-induced shifts can be attributed correctly. We can't prevent model drift, but we can make sure it's visible in the data and not silently absorbed into a brand's score.
Will you measure private or fine-tuned models for an enterprise?
Possibly. Custom-model measurement is scoped separately because the prompt-set construction and validation work differently. Ask us during the scoping call.
Do you publish category benchmarks?
Aggregate, anonymized category benchmarks are published quarterly for categories with sufficient measurement coverage. Individual-brand values are never published without explicit client permission.
Who is the buyer for this?
CMOs, measurement leads, agency accountability teams, and CFOs who want third-party data they can put in front of a board. Not the SEO desk. Not the content team. Those teams should be buying optimization tools, which Valoh isn't.
Still Reading?