SolCrys FAQ - 19 questions buyers ask about how we work

How does SolCrys build prompt sets?

We build a Golden Prompt Set (GPS) for every customer category, grounded on four real-world signals: intent volume across major search and marketplace surfaces, trending questions from public community platforms, AI query volume signals, and live follow-up questions from supported surfaces where reliable. Every prompt comes with projected query volume so you can prioritize what your buyers actually ask. You can replace any GPS prompt with your own at any time - typically 40 to 50 percent of your working set ends up customer-supplied. Full methodology: see our Golden Prompt Set methodology page.

Why doesn't SolCrys just use SEO keywords as prompts?

SEO keywords measure what people type into a search box. AI prompts are longer, more conversational, and increasingly bypass traditional search entirely (asked directly to ChatGPT, Claude, Perplexity, or Alexa for Shopping). Synthetic-keyword prompt sets miss the multi-clause, problem-stated questions buyers actually ask AI assistants ('I run a 50-person SaaS team and our current CRM doesn't handle our 6-month sales cycle well - what should I look at?'). Our community-grounded layer specifically closes that gap.

Are SolCrys prompts AI-generated, or sourced from real demand?

Sourced. Every prompt entering the GPS requires evidence from at least one of the four grounding sources, and we prioritize prompts evidenced by two or more sources. We don't ask an LLM to 'generate 100 questions a buyer might ask' and call it research. Synthetic SEO-only prompts are de-prioritized during selection and should not enter the tracked GPS unless they are supported by real evidence.

How often does SolCrys update the prompt set?

The tracked GPS is designed to stay stable as a measurement benchmark. We do not automatically rewrite prompt sets daily or near real time. Prompt changes are handled as review items with the customer during onboarding, regular measurement reviews, customer-requested replacements, or clear category changes such as a new entrant, pricing shift, or major news event. Your tracked prompts do not change unless you approve the change.

Can I bring my own prompts to SolCrys?

Yes. You can replace any GPS prompt with your own at any time. We provide a suggestion panel that flags gaps in your set (for example, 'you don't have any risk/objection prompts') and recommends additions. Your custom prompts are private to your workspace and never used to update other customers' GPS templates.

How does SolCrys measure visibility, and why should I trust the data?

We measure AI visibility by running tracked prompts against supported engines and preserving the evidence behind each response: prompt, engine, capture method, available model or surface signal, timestamp, answer text, citations, and follow-up questions. Where an engine has a reliable API, adapter, SERP, or rendered-surface capture path, the capture method is tagged separately. Retail-assistant validation is scoped as an add-on when access and reliability support it. Full methodology: see our Visibility Measurement methodology page.

Why is SolCrys's consumer-surface capture better than just calling the API?

A consumer chat or SERP surface is not always a thin shell over a public API. It may use a different default model, enable different tools, or render follow-up suggestions and other UI details the API never returns. Google AI Overviews has no public API, so SERP capture must be tagged separately from API-based response capture.

AI engines give different answers each time. How does SolCrys handle that?

By repeated capture and rolling-window aggregation. A single snapshot is a noisy data point - engines like ChatGPT are non-deterministic by design, sampling probabilistically so the conversation feels natural. For active workspaces, we monitor priority prompts daily and aggregate into rolling 7-day and 30-day windows. Single-snapshot moves are labeled as snapshots rather than reported as trend changes.

My visibility score moved 12% this week — is that normal?

Often yes — and the right question to ask isn't "is this within an expected range?" but "what new content was published, and what citations changed?" AEO scores move week to week even when nothing you control has changed, because (a) AI engines are non-deterministic, (b) the source landscape engines query is constantly changing — competitors publish, third-party media cites different sources, model versions roll, content-freshness windows decay — and (c) the measurement layer is a statistical estimate, not a deterministic count. The score is the symptom; the real signals are new content and citation changes. We walk through the 6 sources of variance and the diagnostic path in our guide: Why your AI visibility score moves. Two operating rules of thumb: watch the 7-day rolling average rather than the single-day number, and pull the underlying citation source list before deciding whether to act.

Can I reproduce a specific data point SolCrys shows me?

Yes. Every chart drills back to the exact prompt, engine, available model or surface signal, capture method, timestamp, and captured response. Copy the prompt text, submit it on the engine yourself within a short time window, and compare substance (allowing for known engine non-determinism). If a result ever looks wrong, request the response evidence and we can show what was captured.

What happens when an engine changes its default model?

We monitor provider announcements and model deprecations on an ongoing basis, with internal alerting for default-model changes. When a configured model or surface changes materially, we update tracking as soon as the change is verified and disclose the change so you can interpret any trend-line discontinuities. Model signal is stored where the engine discloses it.

How do I see SolCrys data on my own brand before committing?

Request our Free AI Visibility Audit. By default, we auto-generate the prompt set directly from your industry's Golden Prompt Set (GPS) - the recommended path, because every GPS prompt is grounded in our four-source evidence model, so the baseline is more intent-aligned than prompts drafted cold. The Free Audit should be treated as a one-time baseline and paid-workflow preview, with sample Deep Analyses and sample recommended actions showing what the paid workflow looks like on a comparable brand. Brand-specific Deep Analysis and your own recommended-action plan are part of paid engagements.

How long does SolCrys onboarding take?

For a standard workspace, the GPS for your category is pre-built and you can review or replace prompts quickly. Our team uses onboarding to tune prompts, configure competitors, validate the engine allowlist, and confirm which buyer surfaces matter most for your category.

How often does SolCrys re-run prompts?

The Free AI Visibility Audit is a single-snapshot read. Ongoing workspaces use daily monitoring for priority prompts and rolling 7-day and 30-day windows that separate real movement from normal engine noise. We do not want buyers to confuse a one-time audit with trend-grade measurement.

How does SolCrys compare to other AI visibility platforms?

We don't write public competitor comparisons by name. Instead, here is the honest framework: there are six methodology questions every buyer should ask any vendor in this category - about prompt sources, capture method, model/surface disclosure, randomness handling, per-data-point reproducibility, and response to engine default changes. Our methodology pages answer all six. Run the same questions with anyone you're evaluating.

What questions should I ask any AEO vendor before signing?

Send these in writing and score 0 / 1 / 2 (no answer / vague / specific): 1) Where do prompts come from - specific evidence sources, not 'AI-generated'? 2) API, SERP, or rendered-surface capture, and how is each tagged? 3) Which model or surface signals are stored, with a disclosure process for material changes? 4) How do you handle engine non-determinism - repeated capture, rolling windows, and snapshot-vs-trend labeling? 5) Can I reproduce a specific data point with full metadata? 6) What happens when an engine changes its default?

A vendor below 8/12 isn't yet ready for production work. We try to score well on all six and we expect you to score us against this same standard.

What's the single biggest red flag when evaluating any AEO vendor?

Any answer that boils down to 'the platform decides which model gets used' - that means data is non-reproducible, and you can't verify a result you cannot replicate. A close second: refusal to walk you from any chart to the underlying captured response. Both correlate with vendors hoping buyers won't ask harder questions.

Which AI engines does SolCrys cover?

Coverage depends on the buyer surfaces that matter for your category and on whether each engine is technically reliable to track. We prioritize major AI answer engines and retail assistants where customers have real exposure, and we confirm the current engine allowlist during evaluation. We don't silently add or remove engines without disclosure.

Is SolCrys's data capture compliant with engine terms of service?

Yes. We comply with each provider's terms and use the access methods each provider supports - public-content access patterns on consumer surfaces, official APIs for the API channel, and SERP capture infrastructure for surfaces with no official API (most notably Google AI Overviews and AI Mode). For enterprise customers with strict compliance requirements, we provide written documentation of the access methods used per engine under NDA.

Can I export my SolCrys data?

Yes, where export is part of the customer workflow. Structured response data can include prompt, engine, timestamp, response text, citations, and extracted entities. If export is important to your team, we confirm the current format and access method during evaluation.