SendRoq

Inference writes the copy and scores the leads, and an automation harness does the work.

What the model contributes is real: messages generated around each prospect's profile and observed signals, research performed automatically before contact, and scoring that ranks leads so the customer knows who to approach first. Credits meter it at 100, 300 and 1000 by tier.

What the product fundamentally is sits underneath that. Visiting profiles, sending connection requests, dispatching follow up messages on a schedule, rotating across connected accounts and staying inside published platform limits are all deterministic automation, and they are the mechanism the customer is paying for. The generated copy makes the automation more palatable rather than making it possible.

The scoring deserves a closer look on this axis. Ranking prospects by whether they viewed your profile, follow a competitor or posted on a relevant subject is rule shaped work that does not obviously require a model, and nothing describes what inference adds beyond ordering observable events.

One complication for a reader. The frequently asked questions on the vendor's own pricing page describe the company as a platform to automate workflows and analyse data, which is not this product, so the model claims there should not be read as describing what ships.

Ask what a credit buys and what the scoring model is actually computing.

Running unattended is the selling point, not a side effect.

The vendor states it plainly across its own surfaces: campaigns run continuously without anyone watching a browser tab, and the customer just shows up to the warm replies. That is the value proposition, and it means connection requests and messages reach real people on the customer's behalf, under the customer's name, without anyone seeing them first.

Two passes located no approval step before a connection request or message sends, no sample review before a campaign launches, no audit trail of what was sent to whom, no administrative permission model and no configuration that would place a person in the path.

The one stated constraint is operating within the platform's published daily limits. That is a rate ceiling protecting the customer's account from restriction. It governs how fast messages go out, not whether any of them should.

A unified inbox and, above the entry tier, a shared team inbox handle what comes back. Both are reply management after the fact rather than oversight before the fact.

The combination worth naming is autonomous sending under a real person's identity with generated copy and no review, which places the reputational consequence entirely on the customer.

Ask whether generated messages can be reviewed before they send.

Nothing is named behind any of it, and the page that should explain the product describes a different one.

Two passes across the pricing page, the feature descriptions surfaced in search, the blog and the plan comparison located no model provider, no model name, no version, no statement of what prospect data is transmitted for inference, and no training or retention position.

The capabilities requiring disclosure are message generation written to read as though the customer wrote it, automatic research of each prospect, and scoring that determines who gets contacted first.

The scoring is where the absence matters most. Intent Radar decides which people are approached and in what order, based on their observed activity, and the vendor publishes nothing about how the score is computed, what weight each signal carries, what range it runs on, or how often the judgement is wrong. A customer configuring outreach around it is trusting a ranking they cannot inspect.

Credits meter model usage at three published volumes without ever stating what one credit corresponds to, so consumption cannot be reasoned about either.

Compounding all of it, the frequently asked questions on the vendor's own pricing page describe an artificial intelligence platform for automating workflows and analysing data, which is not this product, so the little that is said about the technology describes something else.

Ask which provider generates the messages and how the intent score is calculated.

The vendor states its own customer count twice and the two figures differ by a factor of two hundred.

On its pricing page, immediately beneath the plans, the claim is that more than fifty thousand marketers trust the product. On a directory listing the claim is more than two hundred and fifty business and consumer startups worldwide. Both are the vendor's own assertions about the same thing and no reconciliation exists. One of them is wrong and a reader cannot tell which.

The performance claims sit in the vendor's own blog rather than on a customer page: three times more replies than cold outreach, seventy eight percent less time spent on manual prospecting, and campaigns running continuously without supervision. No sample, period, baseline definition or methodology accompanies any of them, and they appear inside an article comparing the vendor favourably against named competitors.

Two passes located no named customer, no logo, no case study, no testimonial attributable to a person, and no review on any independent platform.

The imagery is stock photography carrying generic descriptive alt text about people posing against walls, positioned where customer photographs would normally sit.

Ask which customer figure is correct and how the three times reply claim was measured.

Automated contact on a channel with no unsubscribe, and no terms of service published at all.

The channel matters here. Connection requests and direct messages on a professional network have no unsubscribe mechanism, no suppression concept and no equivalent of the sender obligations that govern commercial email. A recipient who wants the contact to stop can block or report, both of which count against the sender. So the compliance burden sits entirely on the vendor and the customer choosing not to contact people who have not asked to be contacted, and nothing in the product does that.

Two passes located no suppression list, no do not contact handling, no consent basis, no frequency limit beyond the platform's own daily ceiling, and no jurisdictional guidance for a product sold worldwide.

The legal position is thinner than any record built this session. The footer carries a privacy policy and nothing else. There is no terms of service, no acceptable use policy and no customer agreement of any kind, for a platform that sends messages under a customer's identity and holds access to their professional network accounts.

The one control the vendor does describe, staying within published daily limits, protects the account from restriction rather than protecting anyone from unwanted contact.

Ask what terms govern the service and how a recipient stops contact.

One document, and the most sensitive thing a customer hands over is unaddressed.

The footer carries a privacy policy and no other legal document. It was not opened in this pass. Two passes located no terms of service, no data processing agreement, no subprocessor list, no retention schedule, no deletion commitment, no residency statement, no rights request process and no named privacy contact.

What a customer must supply to use this product is the thing that needs addressing. Automating a professional network account requires the platform to hold either the credentials or an authenticated session for that account, and the top tier connects ten of them. That access carries the customer's entire professional network, their private message history and their identity on the platform. Nothing published describes how it is stored, whether it is encrypted, who at the vendor can reach it, or what happens to it on cancellation.

On the other side sits data about people who never agreed to anything. Profile views, competitor follows and posting activity are observed, scored and retained for individuals who have no relationship with the vendor and no way to know they have been ranked.

Ask how account access is stored and what is retained about scored prospects.

The signals are taken from a platform the vendor has no agreement with, which is why no supplier is named.

Intent Radar is described as detecting who is showing buying signals right now, from profile views, competitor follows and relevant posts. Every one of those is activity visible inside the professional network and observable only by operating there. There is no data supplier to name because none exists: the vendor is reading the platform directly.

That is a different provenance problem from the licensed database question elsewhere in this index, and in some ways a cleaner one to state. There is no undisclosed broker chain. There is also no licence, no interface agreement, no consent basis and no permission from either the platform or the people being observed.

What is built from it is a scored and ranked list of named individuals, assembled from their own activity, held by a company they have never heard of, for the purpose of having them contacted.

Two passes located no statement of what is collected, how long it is kept, how often it refreshes, or what accuracy the signals carry. Competitor follows in particular require monitoring third party company pages to establish who follows them.

The naming is worth noting. Intent Radar implies the purchased intent data category, and what it delivers is platform activity observation.

Ask what is collected to detect competitor follows and how long signals are retained.

Every function of this product is a thing the platform's user agreement prohibits, and the vendor's guidance explains how to do it without being caught.

The automated actions are profile visits, connection requests, direct messages and sequenced follow ups, executed across up to ten connected accounts. The professional network's terms prohibit automated access and third party software that scrapes or automates activity, and the observation layer adds monitoring of profile viewers, competitor page followers and posting activity on top.

The framing is the finding. Sending is described as staying within the platform's daily limits so that the customer's account stays safe while their pipeline grows. That is managing enforcement risk, and it concedes the risk exists. This is the seventh instance recorded this sweep of a vendor marketing survivability rather than permission.

The recommended opening move is worse than the automation itself. The vendor's own guidance proposes an automated profile visit as the first step, specifically because it surfaces in the target's notifications and triggers curiosity at no cost, warming them before the connection request lands. That is machine generated activity designed to be read as a person taking an interest.

One inconsistency: a blog article offers unlimited connected accounts at a flat monthly fee, while the published tiers cap at ten.

Ask what happens to a customer's campaigns when a connected account is restricted.

The word security appears once, as a bullet with nothing after it.

It sits in the middle tier's feature list, between account count and team inbox, unaccompanied by any statement. That is the entirety of the vendor's security signalling. Two passes located no encryption position, no access control description, no logging, no retention schedule, no incident response process, no breach notification commitment, no named security contact and no vulnerability disclosure route.

The custody question makes the silence consequential. This platform holds authenticated access to up to ten professional network accounts per customer, each carrying that person's full network, private correspondence and professional identity. Compromise of the vendor would expose all of it across every customer simultaneously.

On the model side, no provider is named, so no training or retention position exists for the prospect research or the generated messages. No accuracy figure, evaluation or error rate supports the scoring that decides who is contacted, and no constraint is described on what the generated messages may say before they go out under a real person's name.

With no approval step recorded on the oversight axis, a generation failure reaches the recipient directly.

Ask how connected account access is secured and what constrains generated message content.

The sequence is designed so the recipient concludes a person is interested in them.

Identity is not falsified: messages go out under the customer's real account, and there is no persona, no rented profile and no synthetic media. That is the one thing in this record's favour on this axis.

Everything else is built on an inference the recipient is meant to draw and that is not true. The vendor's own recommended sequence opens with an automated profile visit, and states the reasoning openly: it appears in the target's notification of who viewed their profile and triggers curiosity before the connection request arrives. The notification says a named human looked at their profile. Nobody did. Software did, on a schedule, as step one of a template.

The messages continue it. The vendor describes them as reading like you wrote it yourself, written around the prospect's profile and their observed activity. A recipient seeing a reference to a post they published concludes the sender read it. A system did.

Two passes located no disclosure of automation anywhere in the flow, and the channel offers no unsubscribe through which a recipient could exit.

The whole mechanism depends on the recipient not knowing what happened.

Ask whether recipients are ever told the profile visit and message were automated.

Two claims of integration exist and both sit inside text describing a different product.

The plan comparison table carries a row for interfaces and integrations. The frequently asked questions state that the product supports integrations with widely used tools across productivity, communication and data platforms. Both appear in template content whose plan names, metrics and product description match nothing this vendor actually sells, so neither can be relied on as a statement about this platform.

Two passes located no named integration, no record system connector, no interface documentation, no authentication scheme, no webhooks and no automation platform listing on any vendor surface.

What exists internally is a unified inbox consolidating replies and, above the entry tier, a shared team inbox. Those are internal surfaces rather than connections outward. The agency tier adds a client management dashboard and white label capability, which serve multi client operation rather than integration.

The practical consequence for a buyer is that replies, campaign outcomes and lead scores appear to have no described path into a record system, so the work this product generates would be transcribed by hand.

One further signal: a feature page indexed by search engines returned a not found error when fetched, indicating pages have been removed without redirects.

Ask which integrations actually exist and whether any interface is documented.

Nothing about the infrastructure was located on any surface.

Two passes across the pricing page, the plan comparison, the blog and the footer found no hosting provider, no region, no data centre, no residency commitment, no tenancy or isolation model, no encryption statement, no backup position, no continuity plan, no uptime commitment and no status page.

What can be established is only how the site is assembled: a hosted page builder serving the marketing surface with the application on a separate subdomain, which says nothing about where customer data lives.

The residency question is not academic for this product. Automating professional network accounts means the platform maintains persistent authenticated sessions and acts continuously on the customer's behalf, so an infrastructure location is also the location from which activity appears to originate. Accounts operated from unexpected geographies are a well known trigger for platform security review, and a customer cannot assess that risk without knowing where the automation runs.

On top of that sits the stored material: account access for up to ten accounts, observed activity about scored individuals, generated message content and reply history.

The product is sold worldwide with no statement of where any of it is processed.

Ask where the automation runs and from which region connections originate.

A one word bullet is the entire security position.

The middle tier's feature list includes the word security with nothing following it, positioned between the account count and the team inbox as though it were a feature being unlocked. Two passes located no certification of any kind, no audit report, no international standard, no penetration test, no trust portal, no security page, no encryption or access control description, no subprocessor list, no vulnerability disclosure route and no named security contact.

The only published legal document is a privacy policy. There is no terms of service.

What makes this the sharpest gap in the record rather than a routine one is what the product custodies. To automate a professional network account the platform must hold that account's credentials or an authenticated session, and the top tier connects ten. Each carries a person's full professional network, their private message history and their identity. A customer is being asked to hand that over to a company that publishes one legal document and a single unexplained word about security.

The agency tier compounds it, since an agency connecting ten client accounts is passing along custody of assets that are not even its own.

Ask what the security bullet refers to and how account credentials are protected.

The prices are clear and everything printed underneath them belongs to another product.

The published part is genuinely good. Three tiers at 59, 119 and 249 dollars monthly, with a yearly option saving twenty percent, a free trial, and cancellation stated as available at any time without hassle. Connected account counts are stated at one, three and ten, which is the metric that actually determines fit. The vendor even criticises competitors in its own blog for requiring a demonstration booking before revealing prices, and its pricing page is consistent with that criticism.

Credits are stated per tier at 100, 300 and 1000, and nowhere on any surface examined does the vendor say what a credit is or what consumes one. Three numbers with no unit behind them cannot be compared or budgeted.

Below the tiers sits a comparison table headed as a full feature comparison, and it describes something else entirely. The plans are named Basic, Pro and Business rather than Solo, Team and Business. The metrics are included users, maximum users, monthly content limit, prompt to campaign and weekly reports, none of which appears in the actual product. All six buttons in it lead to a contact page rather than to signup.

A blog article separately offers unlimited connected accounts at a flat monthly fee, contradicting the tiers.

Ask what one credit buys and which plan structure is current.

Cancellation is easy and nothing comes with you.

The commercial exit is stated twice on the pricing page as cancel any time without hassle or fees, with a free trial at the front. That is a clean commitment and it is the whole of what is offered.

Two passes located no export mechanism, no format, no scope statement, no timeline, no deletion commitment, no retention position after cancellation and no interface through which a customer could extract anything programmatically.

What accumulates is specific to this product and none of it has a described route out. Campaign structures and sequence designs. Reply threads sitting in the unified inbox, which are real conversations with real prospects and in many cases the only record of them, since messages sent through a professional network account live in that account rather than in a record system. Accumulated lead scores. Connection history establishing which relationships were built through which campaign.

For the agency tier the exposure multiplies, since an agency leaving takes responsibility for ten clients' campaign histories with no stated way to hand any of it over.

The absence of any named record system connector, recorded on the ecosystem axis, closes the alternative route out.

Ask whether campaign history and inbox conversations can be exported before cancellation.

The mechanics are correct for the channel and they exist to protect the sender rather than the recipient.

This axis translates awkwardly away from email, where the analogue is account standing rather than inbox placement, and by that measure the vendor does the recognised things. Sending stays within the platform's published daily limits, which is the single most important constraint for avoiding restriction. Volume distributes across connected accounts, with rotation described in the vendor's own guidance. The recommended sequence opens with a low intensity touch before the connection request, which is the correct escalation pattern rather than leading with the highest risk action.

Those are the right mechanics and several tools in this category ignore all three.

What is absent: no numeric limits are published, no per account daily ceiling is stated, no account health monitoring is described, nothing explains what happens when a connected account is restricted or how the platform detects that it has been, and no measured data of any kind supports the approach.

The framing deserves restating from the terms axis. Every one of these controls protects the customer's account from enforcement. None of them limits how many people receive unrequested contact, which on a channel with no unsubscribe is the constraint that would matter to anyone on the receiving end.

Ask what the per account daily ceilings are and how account restriction is detected.

The tier ladder describes the buyer clearly and the frequently asked questions describe a different one.

The ladder works. One connected account for founders and solo operators doing their own outreach, three for a small team with a shared inbox and campaign analytics, ten for agencies running outreach across multiple clients, with a client management dashboard and white label capability at the top. Account count is the right axis to segment on for this kind of product, and the agency tier is genuinely differentiated rather than being the middle tier with a higher number attached.

The stated market is business to business startups and sales teams, and the positioning line about intent led outreach for sales teams done guessing is consistent with that.

The contradiction sits on the same page. The frequently asked questions describe the audience as marketers, support teams, content creators and smaller businesses looking to scale without extra staff, which is a different buyer for a different product, and the section header describes the questions as covering marketing analytics.

What is missing throughout: no geography, no industry vertical, no company size band beyond account count, and no statement of which regions the observation layer covers, which matters because platform activity signals are only as available as the platform's presence in a given market.

Ask which markets the intent signals cover.