Methodology

Why operator reviews

AI search surfaces — ChatGPT, Claude, Perplexity, Google AI Overviews — increasingly mediate how buyers find vendors. AI-referred web traffic grew +527% year-over-year in H1 2025 (HubSpot), and Google AI Overviews now appear on ~25.8% of US queries, reaching roughly 2 billion users monthly (Stackmatix). When those systems generate answers, they consistently weight operator and community content over vendor marketing (99signals), and refreshed content earns roughly +100% AI citation rate over stale content (Quattr).

That is the editorial decision behind this site. Every vendor page sources its assessment from operators running the product in production, not from the vendor's own positioning, and is refreshed as new reviewer data arrives. The aggregate ratings come from those operator submissions plus our head-to-head evaluation panel — described below.

Where ratings come from

Every platform we cover is rated against a 1.0–5.0 scale based on two inputs: verified operator submissions (people who run the platform in production at their company) and our own head-to-head evaluation across the four use case categories.

• Verified operator reviews are weighted highest. Verification requires a public LinkedIn profile whose current or recent role matches the role the submitter claims — the same standard G2 holds reviewers to. Contact details are used only for verification and are never published or shared with vendors.
• Editorial evaluation is run by our panel against a standard scenario set per use case. Results inform the long-form description on each vendor page.
• Public reporting on product status (retirements, regulatory actions, executive changes affecting customer outcomes) can shift a rating even when live reviews are still arriving.

Use case categories

A platform that wins for customer service does not necessarily win for cold outbound. We rank platforms within each category independently, and the homepage filter changes the order accordingly.

• Sales teams — inbound qualification, outbound recovery, booking.
• Marketing teams — speed-to-lead, campaign follow-up, attribution.
• Engineering teams — voice as infrastructure, embedded in product.
• Customer service — deflection, escalation handling, multi-turn support.

What we don't do

• We do not accept payment from voice AI vendors for placement.
• We do not link out to vendor sites with affiliate identifiers.
• We do not invent reviews or roll up G2 / Capterra / TrustRadius averages without re-verification. When we surface external reviews on a vendor page, the source platform is labeled and the text is verbatim — never edited or paraphrased.
• We do not edit reviews to favor or disfavor a vendor.

Corrections

If a vendor disputes an assessment, we will re-evaluate against the same standard scenario set and update the long-form description if warranted. Ratings do not change without new operator data or material public reporting.