lemonreach

The cleanest possible evidence for this axis sits in the pricing table: the billing unit is the model output itself, stated against every tier as one generated email equals one credit. The product is literally sold by the generated message. Apply the removal test and the pitch collapses. Describing an ideal customer in plain language and having matching companies and people returned depends on a model. Persona, vertical and company profile generation depends on a model.

Sequence copy written per persona and vertical combination depends on a model. Reply tagging and one line summaries depend on a model. Talking points generated per prospect before a call depend on a model. What remains is a sequencer, a connected inbox, a telephony integration and an enrichment pipeline whose stated purpose is to ground generation it would no longer feed. That is a materially weaker and differently priced product.

Fourth grade at this band in the index, and the pattern established there holds without exception: every one is a model native vendor, and every incumbent that added a programme to a working product lands lower.

This vendor ships the control almost nobody else in this category does, and it markets it as the reason to choose them. Every generated email passes a preview step before it sends, and the operator can edit it there. That is an approval gate at the point of action rather than a report after the fact, and the vendor states its purpose plainly, that the step protects the sending domain while still removing the bulk of the manual work.

The same position is repeated as a design philosophy, that a buyer reviews every campaign before it ships and keeps a hand on the wheel. Around it sit operator set step delays, a schedule the connected inbox dispatches on, and an option to exclude weekends. What is unusual and creditable is that the vendor names the trade off rather than hiding it, telling prospective buyers on its own comparison pages that a competitor is the closer fit for anyone wanting a fully autonomous agent.

Off the top band because the gate covers composition only: nothing is published on an audit record of what was actually sent, no escalation path exists when the system is uncertain, and reply tagging and next step selection run unattended.

One real disclosure exists and it is about place rather than identity: model inference is stated to run on European servers, which tells a buyer where prospect context goes when a message is generated and is more than most vendors in this category offer. Everything else is absent.

No model, provider, version or processing boundary is named anywhere on the pages read, and no documentation describes how a message is produced, what context is assembled to produce it, or what the persona and vertical generation actually consumes. Two features sharpen the gap. Uploaded case studies and documents feed personalisation on the upper tiers, so customer material enters the generation pipeline with nothing said about how. And reply tagging classifies inbound messages from prospects into categories that drive the next step, which is an automated judgement about a person's intent with no stated basis.

Nothing about outcomes is available to verify. There are no named customers, no logos, no testimonials, no case studies, no customer count and no analyst coverage, and a profile on one major software review platform carries zero reviews. The prospect examples used throughout the site, a head of commerce at a Nordic company of a stated size band, are evidently illustrative rather than claimed as real. Two items are recorded specifically.

A verified address success rate of ninety eight percent is published against the prospecting add on with no sample, method, period or definition of what counts as verified, and it is a load bearing claim for a paid feature. And the return calculator on the pricing page produces a headline annual saving figure from values the buyer sets on sliders themselves, which is a projection presented in the visual grammar of a result rather than evidence of anything. Every call to action is a booked demonstration, and the company appears in a database of newly founded ventures, both consistent with an early stage that explains the position without changing it.

Three items sit adjacent to compliance without amounting to a position. Dispatch is described as deliverability aware with an option to exclude weekends, which is a timing control of the kind few vendors publish. The preview gate means a person approves every message before it goes, which places a human decision behind each contact. And a processing agreement is published as a first class footer link alongside a general claim of compliance by default. What is missing is the substance.

No statute is named on any page read, no acceptable use or anti spam policy was located, no consent position appears anywhere, and no unsubscribe or suppression mechanism is described in a twenty row feature matrix that itemises storage duration, support tiers and enrichment sources in detail. The terms and the processing agreement were not read this pass and are flagged, since an acceptable use clause in either would bear directly on this grade.

Four legal documents sit in the footer covering privacy, terms, a processing agreement and cookies, with a cookie settings control alongside, and their contents were not read this pass and are recorded as unexamined rather than absent. Publishing a processing agreement as a first class link rather than offering it on request is better than most vendors of this size manage and is credited.

Adjacent signals from the comparison pages are favourable, describing compliance by default and a processing boundary limited to the union and adequacy recognised countries. None of that can carry a grade above the middle band without reading the documents themselves.

Nothing visible from the navigation identifies a data protection officer, a data subject request route, a supervisory authority path, a sub processor register or a retention schedule, and the platform holds prospect records, professional network history, conversation content and uploaded customer documents.

Three distinct collection layers operate and no source is named for any of them. Prospect discovery returns matching companies and people from a plain language description with no database, partner or supplier identified. Address and telephone verification runs through what is described only as a multi layer validation workflow, which names a shape rather than a chain.

And professional network enrichment automatically pulls recent posts, skills and work history for every prospect plus the posts their employer publishes, with no mechanism, permission or licence position stated. The contrast with the vendor built immediately before this one in the same session is stark, since that vendor published six named suppliers with their roles, their order and whether each was queried live or served from cache. Nothing here identifies who the people in the pipeline are, how they entered it, or what route they have to leave it.

The sending and calling paths are both approached through the front door, which keeps this off the lower bands. Messages leave through the customer's own connected inbox rather than vendor infrastructure, calls run through a named commercial telephony provider integrated as a partner, and a named system of record carries the synchronisation.

Nothing on the pages read offers a browser extension, account rotation, proxy infrastructure, multiple identity sending or any language about avoiding detection. The exposure sits in the enrichment layer. Professional network posts, skills, work history and employer content are pulled automatically for every prospect the moment they enter a campaign, whether they arrived from discovery, a file import or a record synchronisation, and no mechanism is described, no conformance position is published, and the platform's user agreement is nowhere mentioned. Automatic extraction applied to every record, with the method unstated, is the middle band.

One architectural boundary is stated and it is credited: inference runs on European servers and prospect data is processed only within the union or in adequacy recognised countries, which constrains where content can travel even though it says nothing about what is done with it there. Beyond that the axis is unaddressed.

No training position appears, no cross tenant boundary commitment, no retention statement for prompts, generated content or reply text, and no model provider is named so no provider terms can be inspected either. Two inputs sharpen the gap. Uploaded case studies and company documents enter the personalisation pipeline on the upper tiers, and nothing states whether that material stays inside the account.

And professional network history and post content for people who never contracted with anyone is assembled per prospect as the grounding layer for generation, with no position on what becomes of it. A processing agreement is published and was not read.

The manufactured impression of personal authorship is the explicit product promise here, stated more directly than most vendors dare. Sequences are sold as writing copy that reads like the operator wrote it, the enrichment layer is described as the signal that makes every email feel hand written, and the effect sought is stated as making cold feel warm.

Against that sits a genuine mitigation this index should credit: a person previews and can edit every generated message before it sends, so the words that reach the recipient have passed a human, which is more than can be said for the fully autonomous vendors this one positions against.

Held at the middle band because no disclosure position of any kind is published: nothing states whether a recipient is told a message was generated or that their professional history was assembled to ground it, and the European transparency obligation covering systems that interact with people is nowhere addressed. Generated talking points on calls carry the same silence.

An agent protocol server ships on the middle and upper tiers, which remains uncommon enough in this index to be a genuine differentiator, and it is marked as a first class line in the comparison matrix rather than buried. Two named third party systems carry real depth rather than logo placement: a telephony provider whose call steps drop anywhere inside a sequence, arrive with generated talking points, and branch the following steps on the call outcome; and a named system of record with stated synchronisation.

Inbox integration with scheduled automated dispatch covers the sending path, and a separate news enrichment source feeds recent company activity. A dedicated integrations page exists and was not read this pass. Off the top band because no conventional programmatic interface or webhook surface is named anywhere, so a buyer whose requirement falls outside the protocol server and the three named systems has nowhere to go, and because no marketplace or developer documentation was located.

The residency position names the layer that most model native vendors quietly leak, which is what lifts it above the norm. The vendor states that it runs on European infrastructure with model inference on European servers, that it is a Swedish company, and that prospect data is processed only within the union or in countries the union recognises as adequate.

Naming the inference path specifically matters here, because a platform that generates every outbound message is sending prospect context to a model on every send, and that is precisely the route by which European data usually ends up processed elsewhere. Citing adequacy decisions rather than a vague assurance also gives the claim a real mechanism behind it. Off the top band on three counts. No hosting provider, region or data centre is named.

No sub processor register exists, which matters because enrichment, verification and telephony all involve third parties. And the claim appears in competitive comparison copy rather than in a policy, trust page or the published processing agreement, none of which were read this pass.

Nothing on the pages read addresses security. No certification is claimed, and there is no trust centre, audit report or period, penetration test, dedicated security page, vulnerability disclosure route, status page or enumerated control set, not even the transport encryption and access control basics most vendors of this size at least assert.

A third party technology scan detected a bot verification service on the site, which is visitor screening rather than a security posture and is recorded as an observation about the website rather than about the platform. What sits behind the silence is an authorised connection to the customer's own mailbox with automated dispatch, conversation content across email and telephony, uploaded company documents on the upper tiers, and assembled professional histories for people who never contracted with anyone.

Three tiers publish figures on both billing shapes at one hundred and ninety nine, three hundred and forty nine and five hundred and forty nine euros a month, with the quarterly totals spelled out beside each and an annual option stated to save around a fifth. The unit that governs value is defined in plain words against every tier, one generated email equals one credit, with monthly allowances of seven hundred and fifty, three thousand five hundred and eight thousand.

A prospecting add on is separately priced at seventy nine euros a month with its own credit allowance and its own per action rates of two credits for an address and four for a telephone number. A twenty row comparison matrix marks every inclusion and exclusion explicitly. Campaign storage duration is published per tier, which almost no vendor puts on a pricing page at all. Off the top band on two counts.

Nothing states what happens when the credit allowance runs out, whether unused credits roll over, or what cancellation involves. And every call to action on the page is a booked demonstration, while the vendor's own comparison pages sell self serve signup and published plan pricing as the specific advantage over two named competitors.

One disclosure here is genuinely unusual and cuts both ways. Campaign storage duration is published on the pricing page as a tier feature, at six months on the entry plan and twelve months above it. Almost no vendor states a retention period at all, let alone on the page a buyer compares plans on, and that transparency is credited. It is also an adverse term stated plainly: campaign history expires, so a buyer on the entry plan loses the record of anything older than half a year.

Beyond that the axis is thin. No export function is named anywhere, no file format or schema is specified, no deletion timeline or data handling on cancellation is published, and no conventional programmatic interface exists to pull records out. The agent protocol server on the two upper tiers is a real route for a technically equipped buyer, and a file import path is described for migrating in from a competitor, which implies the same path outward without stating it.

Four controls are published and they cluster before the send rather than after the bounce. Addresses pass a multi layer validation workflow before entering a campaign. Messages leave through the customer's own connected inbox rather than shared vendor infrastructure, so reputation consequences stay with the party making the decisions. Dispatch is described as deliverability aware and runs on an operator set schedule with an option to exclude weekends.

And the preview gate is framed by the vendor itself as protecting the sending domain, which is an unusual and honest way to justify a control that also slows the product down. Off the upper band because the discipline stops there.

Nothing is published on warmup, authentication guidance, sending volume ceilings beyond the credit allowance, pacing intervals within a day, complaint rate thresholds, bounce handling policy, blocklist monitoring or suppression, and no unsubscribe mechanism appears anywhere.

The tier structure gives the segment a numeric shape rather than an adjectival one, capping campaigns, personas and verticals at four each on the entry plan and lifting them to unlimited above, and setting campaign storage at six months there against twelve higher up. What earns the upper band is rarer than that.

On its own comparison pages the vendor states plainly when a competitor is the better choice, writing that anyone wanting a fully autonomous agent spanning calling and revenue operations should look at the named alternative instead. Publishing the case against yourself is close to a disqualifying condition, which is exactly what this axis rewards and which almost no vendor in this index offers.

The same pages carry a dated caveat advising buyers to confirm competitor pricing at source, which is unusual care. Off the top band because no industry, company size band or geographic boundary is named anywhere despite the operating entity and the residency claim both being European, and because the illustrative persona used throughout implies a Nordic commerce focus that no page states.