AIStackRankMethodology
Index

Methodology v0.2

How the index earns trust.

This methodology explains how the index earns trust. AIStackRank compares AI meeting assistants using fields that can be sourced, dated, normalized, and challenged, making every displayed fact traceable and every ranking reproducible.

Scope
Meeting AI
Catalog
27 products
Review cycle
30 days
Revised
August 2026

01 / Operating principles

Evidence before adjectives.

The index separates observable product facts from editorial judgment. A vendor can be included without being recommended, and a popular product can rank below a smaller one when the declared use case calls for it.

Traceable

Each material field keeps a source URL and review date.

Comparable

Plans and limits are normalized before products are ranked.

Independent

Affiliate terms never change scores, inclusion, or placement.

02 / Inclusion rules

What enters the index.

Included

Products that capture, transcribe, summarize, search, or operationalize business meetings.

Standalone assistants, revenue intelligence products, and native platform assistants can qualify.

Excluded

Generic chatbots, abandoned products, invite-only demos without documentation, and tools whose meeting support is incidental.

A listing also needs an identifiable official product page and enough public information to support a basic record.

03 / Source hierarchy

Not all evidence is equal.

  1. 1
    Official pricing, product, security, and documentation pages

    Primary evidence for prices, limits, integrations, certifications, and plan availability.

  2. 2
    Official support articles and release notes

    Used when the marketing site omits operational detail or a change is still rolling out.

  3. 3
    Structured vendor responses

    Accepted for quote-only plans or clarifications, and labeled as vendor supplied.

  4. 4
    Reputable secondary reporting

    Used to discover a change, never as the sole support for a pricing or security claim.

AI-generated research can suggest pages to inspect. It cannot publish or verify a field on its own.

Source checkedA current official page supports the displayed field.
Plan requiredThe feature or price depends on a broader platform subscription.
Vendor quoteThe vendor does not publish a usable public entry price.
StaleThe field has exceeded its review window and is excluded from scoring.
DisputedA correction is under review and the claim is temporarily withheld.

04 / Pricing normalization

Making unlike plans legible.

The displayed starting price is the lowest generally available paid plan that includes the core meeting assistant. Annual pricing is converted to a monthly equivalent and labeled; monthly-only pricing stays monthly.

  • Per-seat prices remain per seat. Usage add-ons are not folded into the base price.
  • Free trials are not counted as free plans. Ongoing free tiers are.
  • Quote-only products are labeled "Contact sales" and never sorted as free.
  • Native assistants are labeled "Plan required" when access depends on a platform subscription.
  • Foreign-currency prices keep their published currency; estimated USD values are used only for sorting.

05 / Ranking model

A declared score, not a universal winner.

Scores are calculated for a specific audience or workflow. The default order favors broadly useful meeting assistants, while category views can apply different fit inputs without changing the underlying product facts.

Workflow fit25%How well the product supports the use case named in the ranking.
Product coverage20%Breadth and depth of capture, notes, search, and follow-up features.
Pricing clarity and value20%Entry cost, included usage, and how clearly limits are disclosed.
Integrations15%Native connections to calendars, CRMs, collaboration tools, and automation.
Governance and security10%Published controls, certifications, retention options, and admin features.
Freshness10%How recently the fields used by the ranking were checked.

Commercial relationships have a scoring weight of 0%. Paid placement is not part of the ranking model.

Product visit buttons may pass through an internal redirect. When an approved tracking URL is configured, the link is marked as sponsored and the site may earn a commission. The direct vendor and evidence links remain available separately, and commercial status is not an input to scoring.

06 / Update workflow

Automation assists. A review publishes.

01Discover

Scheduled checks flag changed pages, new plans, and candidate products.

02Extract

Structured parsers map candidate values into the product schema.

03Verify

A reviewer compares each material change with the primary source.

04Publish

Accepted edits receive a source, review date, and change-log entry.

Freshness targetPricing and core limits are checked every 30 days. Security and integration fields are checked quarterly or when a material change is detected.

07 / Corrections and limits

What happens when the data is wrong.

Disputed fields are reviewed against the strongest available primary evidence. Confirmed errors are corrected with a dated note; unresolved claims are withheld from scoring until they can be supported.

The current catalog is an initial sourced dataset for the product prototype. Production operation will track evidence at the individual field level, retain historical values, and expose a public correction channel. Prices can vary by region, tax, billing cycle, contract, and promotion, so the vendor checkout page remains the final authority.

Return to the index