AI Visibility Intelligence FAQ Quick answers to the questions buyers ask most
Quick answers to the questions buyers ask most about Sparkloft's AI Visibility Intelligence (AIVI). Each answer stands on its own.
What is AI Visibility Intelligence (AIVI)?
AI Visibility Intelligence (AIVI) is Sparkloft Media's four-phase system for measuring and improving how AI platforms represent a destination, organization or brand. The phases are Benchmark, Strategize, Orchestrate and Re-measure. AIVI reports a 0 to 100 score plus a separate Brand Accuracy Audit of the facts AI states.
What is generative engine optimization (GEO), and how does AIVI relate to it?
Generative engine optimization (GEO) is the practice of improving how AI platforms mention and describe a brand in generated answers. AIVI is Sparkloft's method for GEO: it measures a brand's current standing, identifies the sources AI relies on, and assigns each fix to an owner.
Which AI platforms does AIVI measure?
Six core platforms on every engagement: ChatGPT, Google Gemini, Anthropic Claude, Perplexity, Google AI Overviews and Google AI Mode. Microsoft Copilot is added for business, meetings and B2B audiences. Meta AI, Grok and sector-specific AI tools are optional additions.
How is the AIVI score calculated?
The score blends four pillars on a 0 to 100 scale, using responses to questions that do not name the brand: Mention, Citation, Position and Sentiment. Scores are reported overall, by platform and by topic, with confidence ranges.
How many prompts and runs does a benchmark include?
A benchmark uses 25, 50 or 100 prompts per audience, depending on tier, each run three times on every platform. A Standard benchmark of 50 prompts on six core platforms produces up to 900 responses per audience.
Do we need an existing baseline?
No. Establishing the baseline is the first deliverable. Sparkloft needs one 90-minute prompt-strategy workshop per audience, input on priority audiences, markets and competitors, approval of the prompt panel, a short list of key facts for the Brand Accuracy Audit, read access to Google Analytics 4, and introductions to your web, SEO and PR partners.
How long does the Benchmark phase take?
Four to six weeks from kickoff to findings readout. Two or more audiences take the longer end of that range, and the client's turnaround on approving the prompt panel drives the schedule.
How do you handle inaccurate information about us in AI answers?
The Brand Accuracy Audit tests 20 to 50 facts about the brand, classifies each AI answer as correct, incorrect, outdated or missing, and traces errors to the page that likely seeded them. Each fix goes to the owner of that source: the client's web team, a partner, a PR team or a reference site's correction process.
Can Sparkloft guarantee that AI platforms will recommend us?
No. AI answers vary from run to run and platforms change often, so no one controls them. AIVI measures a baseline, directs work at the sources platforms rely on, and re-measures with the same prompts to show what changed. Targets are set after the baseline exists.
Does AI referral traffic show the whole picture?
No. Many AI answers produce no click, so referral traffic in Google Analytics 4 undercounts AI influence. AIVI reports it as a supporting signal next to the score, share of voice, accuracy rate and source presence.
How is an AIVI benchmark priced?
Benchmark pricing is scoped per audience and by panel size (25, 50 or 100 prompts). Contact Sparkloft for a quote. Re-measurement, strategy, orchestration and Sparkloft-delivered work are scoped once cadence and scope are agreed.
How does Sparkloft use AI responsibly in this work?
Sparkloft uses only disclosed, platform-compliant tactics: no fake accounts, undisclosed brand posts, purchased reviews, hidden text or prompt injection. Creator content follows FTC disclosure rules. Prompt panels contain no personal data, staff review every AI-assisted output, and Sparkloft follows each client's AI use policy.
Has Sparkloft run AIVI for clients?
Sparkloft is in the process of implementing AIVI with several clients.
How AIVI compares with software-only AI visibility trackers
| Feature | AIVI | Typical software-only tracker |
|---|---|---|
| Scoring method | Pillars, weights and run counts published to the client | Often a proprietary score with undisclosed construction |
| Prompt set | Built from demand research and client workshops; client approves it | Often generic or auto-generated |
| Runs | Three per prompt per platform, with confidence ranges | Often a single run per snapshot |
| Output | Gaps assigned to a lever and an owner, with implementation tracking | A dashboard number |
| Independence | Does not sell the client's CMS or website platform | Varies by vendor |
Keep reading
Start with a baseline
No existing baseline is needed; establishing one is the first deliverable of AI Visibility Intelligence (AIVI).
