Bindago

The pricing modal settles this decisively and in the vendor's favour as evidence. The two AI features, AI Message Personalisation and Intelligent Campaigns, are gated to the top Supra tier alone at 39.99 USD per month; the 9.99 USD Standard tier ships with essential campaign automations, up to three LinkedIn accounts and contact export and NO AI AT ALL, and is sold as a complete product with its own checkout.

The middle Scale tier at 19.99 adds unlimited accounts, campaign analytics and CRM integration and still no AI. That is the Apollo Basic tier pattern in its purest form: the removal test answered by the vendor's own packaging, with a fully functional product sitting below the AI line rather than beside it. Bindago is a recent vendor taking C, which is the correction that matters to the convention: the predictor on this axis is whether the product survives removal of the model, not the founding date.

The operator configures campaigns and the software then executes them on the operator's own machine through their own authenticated session, so the human is present at setup but not at each action. Real configurable governors ship and are marketed: smart delays, daily limits, human like pacing and conditional steps based on acceptance or reply.

Against that, Intelligent Campaigns makes an autonomous judgement about whether to keep contacting a specific person, described as the AI reading replies and continuing the sequence only when it still makes sense so uninterested leads drop off automatically. That is a model deciding to stop or persist in a live human conversation with no documented approval step, no withholding guardrail and no audit trail of agent actions. The built in inbox lets the operator take a thread over manually, which is a genuine escape hatch, but taking over is not the same as being asked.

No model provider, model family or version is named anywhere on the site, in the FAQ or in the legal pages. Bindago is more specific than most about model inputs, stating that personalisation draws on the lead's LinkedIn headline, role and recent posts, which tells a buyer what is being read. It says nothing about what generates the output, where inference runs, whether prompts leave the machine, or how Intelligent Campaigns classifies a reply as uninterested. For a product whose two AI features both act on third party personal data, naming the input and hiding the processor is half a disclosure.

Bottom of band. The homepage carries three unattributed platform statistics, 9,400 plus connection requests sent monthly, 2,500 plus InMails sent monthly and a 41 percent average response rate, with no measurement basis, sample or period. Third party listings repeat a claim that follow up sequences are proven to raise reply rates by up to 65 percent, again with no basis.

The Trusted By Businesses Worldwide strip is eight unlabelled logo image files repeated a second time to fill the row, so no customer is actually named, and the Terms of Use grant Bindago the right to use customer names and logos by default with opt out, meaning the vendor has a mechanism to name customers and names none.

Observed and recorded as observed rather than concluded: the headline 41 percent average response rate is the same figure shown inside the illustrative product mockup on the same page. Review presence is real but very thin, 4.1 out of 5 on the vendor's own site and 4.8 from two reviews on G2. No case studies.

There is no legal posture. No regulation is named anywhere, no GDPR, ePrivacy, CAN-SPAM or CASL, and no consent framework, opt out route or suppression concept is described for the people receiving the messages. What exists is engineering rather than policy, and it is real: smart daily limits, delays, human like pacing, and bulk withdrawal of pending invitations specifically so the operator stays under LinkedIn's own thresholds.

The Terms of Use instruct the buyer not to use the software for spam, fake accounts or harassment and make the buyer responsible for compliance with LinkedIn's terms and applicable law. That is an instruction plus a disclaimer, not a posture. Note the product sends only through LinkedIn, so telephony and email statutes largely do not bite, which is why this lands at C rather than lower.

A split record, and the good half is the load bearing half. The privacy policy carries a dedicated Bindago Desktop Application section that answers the question that actually matters for this product class: LinkedIn username and password are never saved by Bindago and never sent to any third party other than LinkedIn; session cookies are stored locally per authorised account and removed on logout; and all LinkedIn connection data, messages and prospect information are stored locally and never transmitted to Bindago servers or any third party.

It then enumerates exactly what sits on the local disk. That is a scoped, checkable statement about the sensitive category, and it means the vendor holds only email, payment details and a machine identifier. The gaps keep it off A and they are not small: there is no data subject rights section at all, no lawful basis enumeration, no controller or processor designation, no DPA offered, and cross border transfer to Nigeria rests on user consent alone with no standard contractual clauses and no adequacy finding. The website half of the policy is generic template text. Effective date 1 March 2025.

Provenance is fully determinate and needs no disclosure effort because there is only one source: the operator's own authenticated LinkedIn or Sales Navigator session. Bindago ships no contact database, buys no third party data, runs no contributor network and holds no inventory, and the local only architecture forecloses secondary use entirely, since data the vendor never receives cannot be resold or recycled into a directory.

That is a structurally cleaner position than most enrichment vendors reach by policy. Not A because the vendor takes no stated position on the live question its own export feature raises, namely the legitimacy of extracting LinkedIn and Sales Navigator search results to CSV through a third party tool, which is the contested ground in this category and is met with silence.

The most explicit written acknowledgement of platform risk in the index so far, and it still cannot rise above C. The Terms of Use state that LinkedIn automation carries inherent risks including potential account restrictions or suspensions, that the software is designed to operate within LinkedIn's rate limits but cannot guarantee protection against all forms of detection, and that the operators are not to be held accountable for violation of any version of the LinkedIn terms; the liability disclaimer names account restrictions, suspensions and terminations by name.

The vendor's own blog states plainly that all automation tools carry some risk as they are against LinkedIn's terms of service. Naming the exposure beats concealing it. But the same documents transfer the entire consequence to the buyer, and the goal is still framed as reducing detection.

Held at C rather than D on two grounds that distinguish this from the Alsona pattern: the vendor never takes custody of the LinkedIn credential, and the multi account support is sold as managing several accounts an operator already owns rather than as rotation to push past platform limits. Daily limits, pacing and bulk withdrawal are marketed as staying inside LinkedIn's thresholds, which is the opposite of the evasion posture that earns a D.

The cross tenant training question is structurally answered for the mechanical product, because a vendor that never receives prospect data cannot train on it, and the privacy policy says all LinkedIn data stays on the local machine and is never transmitted to Bindago servers or any third party.

But that blanket statement sits in tension with two shipped AI features that must read LinkedIn profile content, headline, role and recent posts, in order to generate text, and with a reply classifier that must read inbound messages. Recorded as observed rather than concluded: the privacy policy carries an effective date of 1 March 2025 and the most likely explanation is simply that it predates the AI features and has not been revised, not that the statement is untrue.

Nothing published says where inference runs, whether prompts leave the device, which provider receives them, or whether prompt content is retained or used for training. Re verify when the policy is next updated, because a single added clause would move this materially.

Bottom of band, and this record introduces a shape the index has not graded before. Three of the four synthetic surfaces are familiar: manufactured effort, where a real named operator sends a message whose implied research was produced by a model; machine generated text inside a live two way exchange, since Intelligent Campaigns reads the human's reply and decides whether the sequence continues; and no Article 50 position, no AI disclosure line and no mention of the obligation anywhere.

The new one is worth naming because it generalises across the whole LinkedIn automation category: the product automates profile visits, post likes and follows, and those are social signals whose entire communicative value is that a human chose to perform them. A like the operator never saw is a smaller deception than a cloned voice but it is the same class of act, and it is sold here as an essential campaign automation on every tier. Held above D because nothing in the marketing celebrates deceiving the recipient, which is the line Alsona crossed.

Three named native integrations, Pipedrive, HubSpot and Zapier, with auto sync of leads on reply or acceptance, plus one click CSV export as the universal fallback. Zapier carries real reach by proxy. Against that: no public API, no developer documentation, no MCP server, no marketplace or app directory listing, and no third party verifiable evidence of any connector at object level, which is the standard Autotouch met for a B in this index. For a desktop application the integration surface is inherently narrower, and the CSV path means a buyer is never trapped, but breadth and verifiability are both absent.

The first A on this axis in the index, and it goes to a 9.99 USD per month desktop tool because local first is a complete residency answer rather than a partial one. The product is an installed application shipped for Windows, macOS on both Intel and Apple silicon, and Linux, distributed through public versioned GitHub releases.

The privacy policy publishes the full data flow rather than a posture: what stays on the buyer's own machine is enumerated item by item, session cookies and authentication tokens, templates and sequences, campaign settings, prospect lists and contact information; what leaves the machine is limited to a machine identifier and a licence key for activation, explicitly excluding LinkedIn data and personal information; and the vendor side jurisdiction is named outright, with the transfer clause stating that data is transferred to Nigeria and processed there.

Naming the jurisdiction plainly rather than staying silent is the 9Lenses precedent, and this goes further by making the operative residency the buyer's own disk. One honest caveat: the no transmission claim is a vendor statement about the behaviour of its own binary and no third party attestation exists to check it.

No certification of any kind, no trust centre, no security page, no penetration test, no documented control set and no vulnerability disclosure route. The privacy policy's security paragraph is the generic template clause that no method of transmission is completely secure.

What keeps this at C rather than D is that the architecture removes the exposure the band is aimed at rather than leaving it uncontrolled: Alsona took D on this axis specifically for holding authenticated LinkedIn sessions on vendor infrastructure with no controls, and Bindago never takes custody of that credential at all.

Distribution through a public versioned GitHub release channel is a modest supply chain positive, since the buyer can see the release history and the artefact, though no code signing evidence was located. Recorded as observed rather than concluded: this is a small vendor with no located security documentation, not a finding that controls do not exist.

A complete and legible purchase path, which is the test this axis applies. Three tiers published with real numbers on the site, Standard 9.99, Scale 19.99 and Supra 39.99 USD per month, each wired to a live Stripe checkout, with a 10 day trial requiring no card, monthly renewal on a stated 30 day cycle, cancellation at any time, and a refund policy published rather than implied.

There is no quote only ceiling and no enterprise black hole, so a buyer can price and complete the entire purchase without speaking to anyone. The tier ladder is fully differentiated and the gates are shown rather than hidden: Standard covers essential campaign automations, up to three LinkedIn accounts and contact export; Scale adds unlimited accounts, campaign analytics and CRM integration; Supra adds AI message personalisation and intelligent campaigns.

Every gated item is displayed on every tier with a tick or a cross and an explanatory tooltip, so the buyer sees what they are not getting as clearly as what they are, which is more candour than most published pricing pages offer.

The second B on this axis in the index, after Alta, and it is earned architecturally rather than by policy. The buyer already physically holds the data, since prospect lists, messages, templates and campaign records sit on their own disk; one click CSV export of LinkedIn search results, Sales Navigator lead lists, network connections and campaign contacts is a headline feature included on every tier rather than metered, which is the direct inverse of the credit gated export pattern that holds Apollo at C; deletion is self service and documented, by uninstalling the application or clearing application data, and logging out of an account removes that account's stored LinkedIn cookies; and the subscription is month to month with cancellation at any time, so there is no contractual lock.

Not A because nothing addresses the vendor side: no stated post termination rights over the account and licence records Bindago does hold, no deletion timeline, no account closure route and no deletion confirmation artefact, which is the piece Alta publishes and this does not.

Graded on the platform appropriate analogue, per the standing rule that an axis is assessed on what is knowable for the product class. Bindago sends no email, so warmup, domain health, placement testing and complaint thresholds do not apply; the equivalent discipline is throttle and reputation management inside LinkedIn, and real controls ship and are marketed rather than merely claimed: smart daily limits, configurable delays, human like pacing, conditional sequencing, reply detection that halts a sequence, and bulk withdrawal of pending invitations specifically to keep the operator under LinkedIn's pending invite threshold.

That last one is a genuine hygiene feature aimed at the exact mechanism that gets accounts restricted. Not A because no threshold, ramp schedule or recommended volume is published anywhere, no guidance exists for warming a new or dormant account, and the only numeric limit visible on the site sits inside an illustrative dashboard mockup rather than in documentation.

Positioning is coherent and the pricing corroborates it, which is the test Aritic failed. The vendor describes its buyer consistently as individual professionals, recruiters, marketers, small businesses, sales teams and agencies, and the commercial shape matches exactly: 9.99 to 39.99 USD per month, Stripe self checkout, a licence tied to one machine at a time, tiering on the number of LinkedIn accounts rather than on seats or enterprise features, and no SSO, team administration, role based access or procurement surface anywhere.

It works with free LinkedIn accounts as well as Premium and Sales Navigator, which widens the base further down rather than up. Not A because there is no evidence of coverage beyond that band, no vertical specialisation, and no disclosed geographic footprint other than the vendor's own jurisdiction.