Revenue.io

The current identity is a unified revenue AI platform and the intelligence layer is everywhere in the marketing: real time guidance, an ask anything assistant, scorecards, roleplay against synthetic buyers, conversation agents. The removal test still lands in the second band because the base tier is a dialer with live coaching and the AI concentration sits in the top package, while the telephony platform predates the intelligence layer by a decade under the prior brand name still visible in the video channel handle. The intelligence is load bearing at the top of the ladder and optional beneath it, and the vendor's own contract makes AI use voluntary by its terms.

The oversight model is stated contractually rather than as interface marketing, and it is the fullest such statement in this index. The machine's autonomous surface is bounded to analysis and data writes, conversation agents read transcripts and write structured fields, while selling actions route through the rep with real time guidance.

The published AI addendum then binds both sides: automated decisions with detrimental impact on individual rights are prohibited without human supervision, customers must maintain human oversight mechanisms and audit trails, every user holds a contractual opt out from AI features that leaves core functionality intact, customers may demand exclusion from newly introduced AI functionality, material behavior changes trigger notification, and an AI incident reporting duty runs both directions with a designated safety contact. Where the other top grade on this axis was earned in product controls, this one is earned in binding terms.

The deepest disclosure architecture recorded in this index, carried in a published AI addendum to the master services agreement. The vendor states its position under the European AI regulation as a deployer that does not develop or train models, classifies all its AI features at the limited risk tier, discloses that foundational capability comes from third party general purpose models supplied by AI subprocessors on a public subprocessor list, and commits to maintaining model lineage documentation covering origin, provider, versioning, configuration and evaluation methodology, available to customers on written request.

Training sources are addressed in their own section, identifiable customer data never trains models, with the aggregated and de identified carve in disclosed rather than hidden. One irony worth recording: the addendum page itself carries a noindex directive while the marketing pages are indexed, so the strongest transparency artifact on the site is the one withheld from search.

The reference set is enterprise grade and specific: a global hardware company selling into one hundred fifty plus countries reports four hundred percent more opportunities, a logistics firm reports eight times more conversations and ten times more meetings, an equipment rental major reports every call coached and scored, and named individuals with titles from a payments company and a real estate technology firm carry quoted endorsements, one stating the platform was chosen twice over a named competitor.

Directory ratings are printed from three sources including the platform marketplace at four point nine. Held off the top band because the headline numbers are multipliers without published method and the strongest claims live in vendor authored case studies.

The telephone consumer protection statute appears as a named compliance badge on the homepage beside the privacy regimes, and compliance is productized for regulated buyers, the financial services page markets AI that scans live conversations and alerts reps to compliance risks in the moment, an unusual inversion where the monitoring machinery serves the customer's own regulatory exposure.

Held off the top band on two counts: no published statutory guidance, consent framework or suppression documentation was located at the depth the top grade requires, and the dialer's local presence feature, local numbers in forty plus countries for higher pickup, is the caller identity proximity tactic this axis treats as a demerit when marketed as a connect rate lever.

The operational privacy apparatus is thorough: a global privacy policy, a data subject request form, a data processing addendum referenced in contract, a public subprocessor list, an employment privacy notice, badges naming the European regulation and both California statutes plus health data compliance, and a per user right to opt out of AI driven profiling written into the AI addendum with three stated exercise routes.

The workspace data position is unusually strict, mail and calendar data obtained through the consent framework is contractually excluded from all AI processing and training under the platform's limited use policy. Held off the top band because no certification of the privacy program itself and no transfer framework participation appeared on the pages read.

The model is first party, the platform processes the customer's own calls, messages and meetings rather than selling contact data, and the provenance position is actually written down: a dedicated contract section on training data sources and model lineage states that identifiable customer data never trains or fine tunes any model, that foundational capability comes from third party general purpose models whose providers maintain their own lawful data provenance and licensing practices, and that lineage documentation is available on request.

Enrichment enters through a named commercial data partner in the integration row. The strongest licensing statement in this lane, held off the top band because the upstream model providers' provenance is vouched for rather than evidenced.

The sanctioned architecture taken to its limit: the product is built inside the CRM platform itself rather than syncing against it, distributed through the official marketplace where it holds a four point nine rating, and the mail platform relationship is governed by a stated conformance position written into contract, workspace data obtained through the consent framework is subject to the limited use policy by name, with a triple stated carve out excluding that data from every AI and training pathway, referencing the specific sections of the service agreement and privacy policy that implement it.

Authorization runs through the platforms' own consent flows. The professional network appears only as an integration logo with no automation product attached, and no unofficial channel machinery exists anywhere on the surface.

The AI addendum is the deepest stewardship artifact recorded in this index: a governance program aligned to the international AI management standard and the national AI risk framework, a named governance board that reviews all material model updates and subprocessor driven changes before deployment, disclosed testing for bias, hallucination, toxicity, drift, adversarial vulnerability and misuse, periodic red teaming, an AI incident notification duty owed to the customer without undue delay, a designated AI safety contact, synthetic data rules that bar identifiable customer data from test sets, a prohibition on sensitive data as input, and a commitment that identifiable customer data never trains models. The honest caveat is disclosed in the same document, aggregated and de identified customer data and output may be used to improve the AI, a carve in stated plainly rather than buried.

The first contractual authenticity position recorded in this index: the AI addendum forbids the customer from misleading any person into believing machine output is solely human generated, obliges the customer to provide all legally required disclosures about AI use, and states the vendor may implement watermarking, cryptographic provenance indicators or metadata tagging to meet transparency requirements under the European regulation.

Where one competitor this session markets voice cloning so mail reads like the rep wrote it, this vendor binds its customers in the opposite direction. Held off the top band because the recording side is thinner, call and in person meeting capture are core products and no consent handling for recorded parties is described on the pages read.

Native residence on the CRM platform with marketplace distribution, an integration row spanning the two meeting suites, both mail clients, both calendar systems, the marketing automation incumbent and a named commercial data provider, an ecosystem page, a public status page, a support site, and mobile applications on both operating systems that extend the dialer and meeting capture to the field.

Held off the top band because no public programmatic interface documentation, developer program or partner directory of the kind the top grade requires surfaced on the pages read, the depth is on one platform rather than across many.

The architecture pitch is that activity data lives inside the customer's own CRM organization with no shadow system, which answers part of the residency question structurally, but the recordings, transcripts and AI processing necessarily run on vendor infrastructure whose regions, hosting providers and residency options are not described on any page read. No deployment choices, data centre locations or residency commitments are published, so the buyer learns where the CRM writes land but not where the conversation media lives.

A service organization control type two examination is claimed with a footer badge, a health data compliance seal is displayed, a live third party application security verification badge links to the assessor's confirmation page, a dedicated security page exists in the legal footer, a public status page runs, and enterprise security and compliance controls are listed in every pricing tier.

Held off the top band because no trust centre hosting the actual artifacts, no international certification and no penetration testing disclosure appeared on the pages read, the badges assert where the top grade requires the documents.

The packaging disclosure is complete and the price disclosure is deliberately absent: three named tiers are published with a full feature comparison table, the seat floor of fifteen is stated twice, support levels per tier are spelled out, and the pricing questions section then says outright that the company builds a custom quote on the demo call rather than publishing a list price. Candor about withholding is better than pretending, but the axis measures whether a buyer can price the product before a sales conversation, and here they cannot at any tier.

The native architecture is itself half an exit story, every call, email, meeting and insight is written into the customer's own CRM records in real time, so the activity history survives the vendor relationship by construction. The other half is unaddressed: the AI addendum states the company is not required to retain inputs or outputs once a customer stops using the AI, a retention disclaimer rather than a portability commitment, and no export mechanism, deletion timeline or media retrieval process for recordings and transcripts appears on the pages read.

Email and text sequencing ship in the middle tier and the platform sends through the customer's connected mail accounts under the platforms' consent flows, with none of the category's volume machinery anywhere on the surface, no mailbox inventory, no warmup service, no purchased domains, no rotation marketing. The telephony side carries its own reputation question, local presence numbers in forty plus countries, which this index records under the compliance axis. The clean email position is architectural rather than stated, no sending policy or authentication guidance is published.

The segment statement is precise and repeated: purpose built for teams on the major CRM platform with fifteen or more seats, from mid market to global enterprise, and the fit logic is even explained, that seat count is where consolidating dialing, coaching, engagement and AI pays off fastest.

Industry pages cover financial services and technology, role pages span development reps through marketing, local presence numbers in more than forty countries state the calling footprint, and the reference logos run from payments to medical devices to publishing. Held off the top band by the absence of geographic, language or regulatory coverage statements beyond the calling footprint.