Overloop

The removal test has a literal answer here and the vendor supplies it themselves. Two products run in parallel with separate logins offered side by side in the navigation and the footer: the current model native platform, and the previous generation still live and still supported.

So what survives removal of the model layer is not a thought experiment, it is a shipping product with its own customers, comprising a database of more than four hundred and fifty million contacts, an email finder and verifier, multi channel sequencing, a conversations inbox, full engagement tracking, reporting and system of record synchronisation.

The commercial design points the same way: credits are consumed when a prospect is sourced or an address is found, and not when a message is generated, so the vendor meters the data and gives the intelligence away. That is a data business with a model layer on top, which is the established platform convention applied on unusually clear evidence rather than a harsh read.

Two human review gates are named explicitly and the permission model underneath them is thin. The vendor states at the top of its pricing page that campaigns run in the customer's voice with review before launch, and separately describes the list building step as preparing matching prospects for the team to review, so a person sits between the model and both the audience and the outbound message.

An exclusion list blocks addresses and whole domains across the entire account so that nobody in the company can contact them, which is a real containment control. Held at the middle band because the roles table published by the vendor discloses the limitation plainly: the two self serve tiers offer a single role, administrator, so every user on a growing team has full control of the account, and separate user and view only roles appear only on the quote priced enterprise step. No audit trail, escalation rule or constraint on what the generator may write appears anywhere.

No base model, provider, family, version or hosting arrangement appears anywhere on the page read. What the vendor does describe is the input side, and reasonably precisely: the writing engine is stated to analyse the prospect's own website and social profiles before composing, and the follow up engine to read every message already sent in the thread. That tells a buyer what the model looks at and nothing about what the model is. Note that the credits system does not meter generation at all, so unlike most vendors in this session there is not even a commercial unit through which the model layer becomes visible.

The evidence infrastructure exists and none of it was reachable from the page read, so this row is held at the middle band and flagged. A success stories section and a separate testimonials page are each published at their own route and linked from the navigation and footer, which is more than a decorative logo row and is the right place for this material to live.

Independent corroboration is real and specific: a major review directory carries a rating of four point four out of five across one hundred and thirty three reviews, with roughly three quarters of reviewers in the small business band, which both supports the vendor's positioning and gives a genuine sample size. Against that, no named customer, quantified outcome, case study figure or measurement basis appears anywhere on the page read, and the vendor publishes no usage or scale claim of its own.

One real suppression control is shipped and no statutory position was located. The exclusion list is a genuine mechanism rather than a gesture: addresses or entire domains can be added, and the vendor states that once added nobody in the company can create or contact a prospect at that address or domain, which is account wide enforcement rather than a per campaign filter. A dedicated regulation page and a terms of use document are each published at their own route and were not read.

Beyond the exclusion list, no statute is named, no consent requirement, unsubscribe obligation or complaint route appears on the page read, at a vendor sending both mail and professional network messages and supplying the contact data for both. Reading the regulation page and the terms is the highest value verification available here.

The structural position is stronger than most vendors this size and the documents themselves were not read. What is verified matters: the operating entity is named in full with a registered street address in Brussels and a published value added tax number, so the company is identifiable and directly subject to the European data protection regime rather than reachable only through a contact form.

A privacy policy, a terms of use document and a dedicated regulation page each sit at their own public route under a legal hub. Most usefully, privacy and processing addendum documentation is listed as an included feature on every plan including the sixty nine dollar entry tier, so the addendum is not gated behind an enterprise contract or an access request, which is the treatment it receives at several larger vendors in this index.

The site is published in five languages, consistent with a European customer base. Off the top band because none of the documents were opened, and no retention schedule, sub processor list, representative or residency commitment was located.

A database of more than four hundred and fifty million business contacts sits at the centre of this product and its provenance is stated nowhere. No source, partner, collection method, jurisdiction, refresh cadence, accuracy rate or lawful basis for holding and reselling those records appears on the page read, which matters more than usual at a company established in a jurisdiction where the basis for processing that data is a live legal question.

One point in the vendor's favour on honesty rather than provenance: the email finder is described with an explicit caveat that it returns a verification status and that availability and results vary by contact, which is a more candid account of what an enrichment lookup actually delivers than most competitors give. The writing engine separately analyses each prospect's website and social profiles with no basis stated for that either.

The professional network exposure is real and the vendor's framing of it is the most restrained in this category. Automated connection requests and messages are generated and sent, and a browser extension turns a profile into a contact record, which is the specific condition this band describes.

What separates it from the vendors graded lower is the language: the vendor's own navigation describes the feature as planning measured outreach steps and reviewing platform limits, which points a buyer toward the platform's rules rather than around them, and is the direct opposite of the marketing that earned the bottom grade elsewhere in this index.

Held at the middle band because acknowledgement is not a position: no conformance statement, permitted use policy or allocation of risk between vendor and customer was located, and the extraction method behind the extension is not described.

Nothing on the page read states whether customer content is used to train or improve any model, whether prompts are retained by any provider, or whether anything crosses between accounts. The material at stake includes connected mailbox contents, the conversations inbox, professional network message history, and the prospect research the writing engine assembles from third party websites and social profiles.

The document most likely to carry an answer is the processing addendum, and the notable fact here is that it is included on every plan rather than gated, so a buyer can actually read it before committing. It was not read and is flagged as the priority verification for this row.

Messages are composed by a model that has read the recipient's website and social profiles, and the vendor's own description of the benefit is that this happens with zero typing required, so the person receiving a researched personal message is receiving something no human wrote or read. No position on article 50 of the European artificial intelligence regulation appears anywhere, which is notable at a company established inside the jurisdiction that regulation governs.

Tracking is extensive and undisclosed to the tracked party, covering opens, clicks and a conversion pixel placed on the customer's own pages. Two things sit on the other side of the ledger and are worth recording: campaigns are reviewed by a human before launch, so a person has at least seen the message, and the exclusion list gives a recipient's domain a route to permanent suppression.

Solid breadth and one disclosure almost nobody makes. Three native system of record integrations are published with real depth described, covering field mapping, record types, user mapping and optional logging of sent mail through each platform's blind copy address.

A published interface, a workflow broker reaching more than a thousand applications, a scheduler, a messaging platform and a browser extension complete the set, alongside a help centre, a public status page and five language versions of the site. The disclosure worth naming is a limitation the vendor volunteers inside its own integration description: prospect replies cannot synchronise from this platform into one of the three named systems of record.

Publishing a specific gap in your own connector, where a buyer will meet it, is rare enough to be creditable. Off the top band because the interface is gated above the entry tier, one system of record is reserved for the quote priced enterprise step, and no webhooks, protocol server, developer documentation subdomain or marketplace was located.

Nothing on hosting provider, country, region, data centre or residency option appears on the page read. What is available is jurisdictional rather than technical and it is unusually complete for this axis: the operating entity is a Belgian company with a published Brussels street address and tax registration number, the site is served in five European languages, and the enterprise tier adds bank transfer and the European payments scheme as settlement options for teams that need them.

That establishes where the company sits and where its customers are, and says nothing about where mailbox contents, prospect records or generated content are actually stored. Third party listings describe the platform as hosted in Europe, which is consistent with everything above but is not a vendor published statement and should be verified before it is relied on.

One real third party assessment, named precisely, available on every plan, and scoped far more narrowly than the marketing around it suggests. The vendor states that it is certified for handling sensitive data and identifies the assessment by name and tier, which is the enumeration this axis rewards and which several vendors in this session avoid.

The important qualification is what that assessment actually covers: it is a cloud application security review conducted for access to a particular provider's restricted data scopes, not a general audit of the platform, so describing it as certification for handling sensitive data is broader than the certification's own scope. Beyond it there is no general security audit, international standard, penetration test, trust centre, security page or vulnerability disclosure route located. Real controls that do exist are modest and mostly commercial: a public status page, role separation on the top tier only, and shared or dedicated sending addresses by plan.

A buyer on either self serve plan can compute their exact bill and know precisely what exhausts it. Two tiers carry real per seat monthly figures at sixty nine and ninety nine dollars, and every gating quantity is published against them: credits per user per month, connected mail accounts per user, active campaigns per account, available roles, which integrations unlock at which step and which support tier applies.

The decisive disclosure is the credit rate card, published in plain terms in the questions section: sourcing a prospect from the database costs one credit and finding a valid address costs one credit, whether the prospect was sourced or imported. A per unit rate for the thing that actually consumes the allowance is the single most common omission on this axis and it is here. A fourteen day trial takes no card and signup is self serve with no demand call.

Two honest caveats keep this from being flawless rather than from being top band: the enterprise step is quote priced, and six named services covering technical setup, secondary domains, onboarding, deliverability review, priority support and account management are listed with no rate against any of them.

Independent review commentary also indicates that credits cannot be topped up mid month and that no warning fires when they run out, which is a product constraint rather than a disclosure failure but is worth a buyer knowing.

One ungated product capability and no published contractual position. Comma separated export appears in the feature list under the heading of capabilities available on all plans, so bulk retrieval of prospect data is available at the entry price rather than reserved for larger customers, and a published interface on the middle tier and above gives a programmatic route as well.

Against that, no export format guarantee, post termination retrieval window, deletion timeline, retention period or data return obligation was located on the page read. What a departing customer would want back is substantial: prospect records, campaign and sequence histories, the conversations inbox spanning mail and professional network threads, engagement history and the exclusion list they have built. The terms of use and the processing addendum were not read and are flagged as the documents that would settle this.

The standout is a disclaimer rather than a feature, and it is the most honest sentence about deliverability recorded in this index. Describing its own built in warmup, the vendor states that it sends low volume traffic through a third party network and then adds that deliverability and domain reputation still depend on provider policies, configuration, content, recipient behaviour and sending history.

That is a vendor refusing to let its customer believe warmup is a guarantee, and naming the five factors that actually decide the outcome, at a point in the page where every competitor in this session instead promises higher rates or in two cases absolute delivery.

Real mechanics sit around it: bulk address verification in real time, automatic detection of both soft and hard bounces alongside out of office replies, an account wide exclusion list, shared sending addresses for secondary domains on the lower tiers rising to a dedicated address on the top one, a technical setup and deliverability review offered as a service, and three free public diagnostic tools covering domain health, spam scoring and sending infrastructure. Off the top band because no complaint rate threshold, unsubscribe header commitment, per mailbox volume ceiling or feedback loop handling is published.

One coverage disclosure here is close to unique in this index: the entire site is published in five languages, with parallel routes for each, which is a verifiable statement about who the vendor serves rather than a claim about global reach. European settlement options on the top tier point the same way.

Each tier carries a plain sentence naming its intended buyer, running from individuals through growing teams to large teams, and six solution pages address the distinct jobs a buyer might be shopping for, from lead generation through cold mail to a sales development agent. Two named competitor comparison pages place the product. Independent review data corroborates the positioning precisely, with roughly three quarters of reviewers in the small business band across a real sample. Off the top band because no customer count, seat band, region or country list, industry breakdown or statement of who this is wrong for appears anywhere.