Leadspicker

The current positioning leads with autonomous agents and the navigation puts them first, but the removal test looks past the marketing to what would remain. Strip every model and the product still holds a lead database, a named multi supplier enrichment waterfall, multichannel sequencing with branching, a unified inbox and record synchronisation, which is a complete and sellable platform.

The corroboration is a customer review describing three years of continuous use built around exactly those components, automatic enrichment and verification, email sequence setup, professional network outreach and the record connection, with no mention of agents at all. The model layer adds plain language configuration, classification and personalisation columns, and an assistant, all of which are priced as separate consumption lines rather than as the substance of the plan. Graded at the middle band on vintage rather than volume, the same reading applied to every established platform in this index that layered a model programme onto a working product.

Agents are sold as running unattended around the clock, discovering and evaluating people from live sources without supervision, and the plain language configuration means the operator describes an outcome rather than specifying steps. Three genuine constraints exist and all three are commercial rather than editorial. Routines are capped per tier at five, fifty and then unlimited.

Credits function as a hard spending ceiling, with automatic top up being an opt in choice rather than a default, so an agent cannot run past a budget unless the buyer allows it. And monthly export limits bound how many records can leave. One real audit artefact is described: the agent keeps a detailed log recording the people it evaluated and rejected as well as those it surfaced, so the buyer can see what was considered and not just what was returned.

Off the upper band because nothing sits at the point of action: no approval step before an agent contacts anyone, no review queue, no escalation path when the agent is uncertain, and no stated containment on what a generated message may say.

Model transparency here goes beyond anything else in this index, and it is structural rather than a paragraph of assurance. Three model providers are named outright in the pricing table rather than left to inference. A model picker exposes the choice to the buyer, and the picker is stated to show the exact cost per row before anything runs.

A dedicated pricing table quotes each model the way its provider quotes it, with the stated purpose of letting a buyer hold those figures against the provider's own price list. Two routes are then published for removing the vendor from the model path entirely: connect a personal account key with either of two named providers and the vendor charges nothing for those messages, or drive the same agent through the vendor's agent protocol server inside an external client so the model used is the buyer's own.

Publishing how to avoid paying you is an unusual thing to do and it makes the underlying disclosure verifiable. Off flawless on two counts: the per model pricing table body did not render on retrieval and is recorded as unread, and nothing states what context is sent to a model, whether prompts or outputs are retained, or what the agents transmit when running on the vendor's own keys.

Two named case studies carry real companies and one carries a number: a major European cloud infrastructure provider reported as improving response rates sevenfold across startup outreach in several languages, and a sales execution vendor identified alongside the acquirer it was bought by. Dedicated sections exist for customer stories, video testimonials and webinars.

Independent presence is genuine and on two separate review platforms, with one review describing three years of continuous use and naming both what works and what does not, which is a shape solicited reviews rarely take. Funding is corroborated externally through a seed round and an investor name, and the site footer carries a public programme disclosure naming a European recovery fund and a national investment agency, which is a verifiable third party relationship rather than a badge.

Off the top band because the headline percentages published across the product pages, covering response rate multiples, conversion improvement and enrichment accuracy, carry no sample, method or period, and because the case studies themselves were not read this pass.

Nothing addressing outbound compliance was located on the pages read. No statute is named anywhere, no acceptable use or anti spam policy appears in the footer alongside the six documents that are published, no consent position is stated, and no unsubscribe or suppression mechanism is described in the pricing table, the product navigation or the frequently asked questions.

The one answer that appears to address the subject is about platform account restrictions rather than about law, promising a reduced risk of restrictions rather than any position on lawful contact. Terms of use and a data processing document are published and were not read this pass and are flagged, since an acceptable use clause in either would bear directly on this grade. The exposure is broad, since the platform both discovers people who have no relationship with the buyer and sends to them across two channels with sending seats sold by the hundred.

The legal estate is unusually complete in structure and was not read in substance, so it is recorded as unexamined rather than absent. Six documents sit in the footer covering terms of use, refund policy, cookie policy, privacy policy, a separate data processing page and a link labelled as a route to decline the sale of personal data, the last of which resolves to a privacy centre inside the application rather than to a static page.

A separate data processing document and a self service privacy centre are both more than most vendors of this size publish, and the privacy centre in particular is the kind of mechanism that distinguishes a vendor that has thought about data subjects from one that has not. None of that can be graded above the middle band without reading it.

Nothing visible from the navigation identifies a data protection officer, a supervisory authority route, a sub processor register beyond the enrichment suppliers named in the pricing table, a retention schedule or a transfer position.

The supplier chain is published in more detail than anywhere else in this index, and it is published as a table rather than a claim. Six sources are named with the role each plays: a lead database serving cached records, a first pass address finder, a second finder used only when the first returns nothing, a validation service that checks a risky address, a global telephone provider, and a web search service. Each line states whether the source is queried live or served from cache.

The waterfall order is therefore visible rather than summarised, which is the direct inverse of the vendor elsewhere in this index that published a supplier count and named none of them. Off the top band on the half that matters most for provenance rather than for procurement.

The vendor also collects directly, describing agents that gather from professional network profiles, job boards, mapping listings, company websites and industry directories on the basis that anything publicly available can be found, and no licence position, lawful basis or notice covers that layer. The underlying lead database's own provenance is not addressed either, so the chain is transparent from the vendor outward and opaque one step further back.

One architectural claim here is specific, quantified and genuinely favourable to the buyer, and it deserves stating before the criticism. The vendor answers the safety question by stating that more than ninety five percent of professional network actions are automated through its own infrastructure rather than through the buyer's account, unlike many competing tools, which moves the restriction risk onto the vendor rather than leaving it with the customer.

That is the opposite posture to the vendors in this index whose buyers hold the account that gets suspended. Against that, the position stops at assertion. No platform guideline is named, no conformance standard is cited, and the claim that the vendor respects the guidelines of the platforms it works with carries no detail a buyer could check. The remaining share of actions still runs through the buyer's own connected sending seats, which are counted and sold per tier.

And the discovery layer openly gathers from a professional network alongside mapping listings and directories, with the pricing table itself listing live profile and post retrieval as billable actions.

No training position, cross tenant boundary commitment or retention statement for model inputs and outputs appears anywhere on the pages read. What does exist is architectural and is credited here even though the vendor frames it commercially rather than as a privacy control.

Two published routes remove the vendor from the model path: connecting a personal provider account key, after which the vendor states it charges nothing for those messages, and driving the same agent from the vendor's agent protocol server inside the buyer's own client software, where the model is the buyer's. A buyer who cares where content goes has two supported ways to keep it out of the vendor's own model relationships.

That is real, and it is also not a stated commitment about anything, since both descriptions are about billing rather than data handling. On the default path, where the vendor's own provider keys run the agents, nothing states what is sent, what is retained or whether anything crosses between accounts. A processing agreement and a privacy centre are published and were not read this pass.

The synthetic surfaces here are the two this category shares generally: messages composed by a model and sent under the operator's own name, and professional network actions attributed to the operator's profile. The second carries an unusual wrinkle worth recording, because the vendor states that most of those actions are executed from its own infrastructure rather than from the buyer's account, which means a connection request or profile view appearing to come from a named person is performed by a third party's systems.

That is disclosed as a safety feature and never framed as a disclosure question. Nothing published states whether a recipient is told a message was generated, no notice explains how the person was discovered, and the European transparency obligation covering systems that interact with people is nowhere addressed.

Held at the middle band rather than lower because nothing is marketed as concealment: there is no simulated conversation, no cloned voice, no manufactured caller identity and no undetectability claim, and a route to decline the sale of personal data is published.

An agent protocol server is included on every tier, including the entry plan, which is better placement than any comparable vendor in this index has managed, and the vendor documents driving its own agent from external client software rather than treating the protocol as a checkbox. Programmatic access and webhooks ship from the second tier upward and are marked explicitly against each plan.

Integrations are included at every level, a partners directory is published, and the enrichment layer itself is a set of named third party services rather than an opaque internal one. The pricing table doubles as integration documentation, since every data action names the supplier behind it and states whether that supplier is queried in real time or served from cache.

Off flawless because programmatic access and webhooks are withheld from the entry tier while the protocol server is not, which is an odd asymmetry, and because no individual integration carries its own documentation page describing objects, fields or synchronisation direction.

Nothing addressing residency was located on the pages read. No hosting provider, region, country, data centre or residency option appears in the navigation, the pricing surface or the frequently asked questions, and no regional choice is offered or sold. The only geographic facts published are corporate: a Czech operating entity, a Prague base, and a footer disclosure naming a European recovery fund and a national investment agency as programme funders.

Those establish where the company sits and not where the data does. The gap is more conspicuous than it would be for a vendor of another origin, because a European company selling into European buyers under a European regulatory frame is exactly the case where a stated residency position is expected, and because the enrichment chain routes records through at least five named third party services whose own locations are not addressed either.

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.

The word safety does appear in the frequently asked questions, but the answer beneath it concerns the risk of professional network account restrictions rather than any security control. What sits behind that silence is substantial: connected sending accounts on two channels sold by the hundred, a unified inbox holding conversation content across both, harvested records on people who never contracted with anyone, and provider account keys where buyers have connected their own.

Among the most complete pricing pages in this index, and the clarity is deliberate rather than accidental. Four tiers publish euro figures at one hundred and ninety nine, four hundred and seventy nine and one thousand five hundred and ninety nine a month plus a quoted enterprise plan, and five separate capacity dimensions are given a number against every tier: credits, email sending seats, professional network sending seats, routines and team members.

The sentence that settles the grade is the credit definition, which reads that one credit is always one euro cent so the credit figure and the money figure are the same number in different units. Vendors in this category routinely obscure exactly that.

Beneath it sits a per action price table with a figure for every operation, from two credits for a verified address to fifty for a telephone number down to four hundredths of a credit for a profile lookup, alongside a list of actions that cost nothing at all. Pricing is stated as per delivered result, so an unenriched row costs nothing and where several suppliers are tried only the one that returns the value is charged. Monthly export limits are published per tier.

Credit expiry is spelled out across three billing shapes, and separately bought credits are committed to survive cancellation, which is the direct inverse of the forfeiture on cancellation recorded elsewhere in this index.

Extraction routes are real and more than one. Programmatic access and webhooks ship from the second tier upward, an agent protocol server ships on every tier including the entry plan, and unlimited records can be uploaded from a spreadsheet, which implies the same path outward. Cancellation is self serve from the billing section with the subscription running to the end of the paid cycle.

The commitment that lifts this above the middle band is unusual and specific: separately purchased credits are stated to remain with the buyer until spent and explicitly not to be taken away on cancellation. Very few vendors address what happens to prepaid value at the exit, and the ones that do in this index usually forfeit it.

Off the top band because no export function is named as such and no file format or schema is specified, because no retention window or deletion timeline is published for lead records, campaign history or inbox content, and because the two programmatic routes are withheld from the entry tier.

Two real hygiene controls ship and both sit before the send rather than after the bounce. Address validation through a named third party service is included free on every row rather than metered, and a verified address is charged only when a valid one is actually returned, so the commercial design rewards not sending to bad addresses. The multi supplier waterfall behind address finding means a row that cannot be resolved simply is not, rather than being guessed at.

Sending seats are capped per tier on both channels, which bounds volume structurally. Beyond that the axis is unaddressed: nothing is published on authentication guidance, warmup, sending pacing or intervals, complaint rate thresholds, blocklist monitoring, bounce handling policy, list hygiene beyond the initial validation, or unsubscribe and suppression mechanics.

Segmentation is worked in three directions rather than asserted once. By role, four buyer types each carry a dedicated page covering sales teams, agencies, revenue teams and marketing teams. By use case, six further pages cover agency operations, enrichment, signal tracking, account based motions, inbound and outbound.

And each pricing tier names its intended buyer directly, running from a solo founder or growth operator through regular sales teams to mature revenue organisations with advanced workflows. Team member limits of one, ten, one hundred and unlimited give that ladder a numeric shape rather than an adjectival one, and agency operation is treated as a first class motion with tooling described for running many client accounts on one platform.

Three named competitors are addressed directly, covering an enrichment specialist and two large contact database vendors, which turns the positioning into something a buyer can test. Off the top band because no upper ceiling, unsuitable use case or disqualifying condition is stated anywhere, and because no geographic boundary is named despite the operating entity and its funding both being European.