MakeSales

The product is the model and nothing survives its removal. The artificial prospect that joins the call, argues, raises objections and can be closed is the entire deliverable, the persona is generated from the seller's own description of their offer and market, and the scoring across five stages is a model judgement about a conversation that only existed because a model held it. Strip the model layer out and there is no product, not a degraded one.

The pricing structure confirms it from the other direction: every tier is denominated in voice and video minutes, which is compute, so the buyer is purchasing model time directly rather than software with a model attached. This is the fourth instance of the top band on this axis in the index and, like the others, it is an artificial intelligence native product rather than an established platform with a model programme.

The autonomous element here is unusual for this index and worth stating precisely: the model is not acting on the buyer's behalf toward a third party, it is acting toward the buyer as a training counterparty. That inverts the usual risk. The buyer selects a difficulty level from three, supplies the inputs the persona is built from, and starts and ends every session, so control over when the system runs is complete.

What is undescribed is the layer that matters for a training product, the scoring. Whether a human ever reviews the model's judgement of a call, whether the rubric is visible or adjustable, and what recourse exists when the score is wrong are all unaddressed, and the readiness score is presented as a signal of when a seller is ready for a real buyer.

One supplier is named openly, which is more than most vendors in this session manage: the video avatar technology is attributed to a named third party and described as the same technology enterprise sales teams use for training. That single disclosure is the reason this row is not lower. Everything else is unstated.

No model, provider or version is named for the conversational layer that improvises the prospect's objections, none for the scoring layer that produces the readiness number, no hosting location, and no boundary between the two. For a product whose entire output is a model generated conversation and a model generated judgement of the buyer's performance, naming the avatar vendor and nothing else leaves the substance undisclosed.

Four quantified claims carry the marketing and not one has a basis attached. Ten times more sales repetitions a month, a 68 percent close rate displayed inside a mock interface, a 5,000 dollar loss attributed to one fumbled close, and a sample readiness score of 56 out of 100 with stage scores beneath it. The 5,000 dollar figure does the most work, since it is the anchor for the whole price comparison and appears three times, and it is asserted without a source, a segment or a method.

Testimonials, named customers, case studies and independent review listings are all absent, which is consistent with a product in beta and is still what this axis measures. The claim that most users forget they are talking to a machine within 90 seconds is the same shape.

This product sends nothing and contacts nobody, so the axis is graded on what is knowable rather than skipped, and there is genuinely little exposure to record. Every call is between the buyer and a simulated counterparty, no prospect list is uploaded, no message reaches a real person, and the electronic marketing regimes that drive this axis elsewhere do not engage.

The one adjacent question a buyer should note is that a script builder ships in every tier, so copy rehearsed here is intended for real conversations later, and the vendor takes no position on what may be said in them. That is a light note rather than a finding.

The privacy position is a frequently asked question rather than a policy, and for a four line answer it covers more ground than several full policies graded this session: offer details and call transcripts are stated to stay private to the buyer, training on their data is ruled out, and deletion of everything is offered from the dashboard as a self service function rather than an email request. Self service deletion is genuinely uncommon. What is absent is the document itself.

No privacy policy, terms of service or data processing terms are linked anywhere on the page, no lawful basis, retention period, sub processor or transfer mechanism appears, and the sensitivity of what is held is high, since a recorded call carries the buyer's voice and, on the top tier, their video. The named avatar supplier is a sub processor by any reading and appears in a product answer rather than in a register.

There is no third party data in this product at all and that is the finding rather than an absence. No contact database, no enrichment credits, no scraped profiles, no purchased list and no supply chain of unnamed brokers appear anywhere, because the only inputs are the buyer's own description of their offer, their customer and the objections they hear.

The personas are generated from that input rather than modelled on real individuals, so no identifiable person's data underpins the product. Against the vendors graded around it this session, several of which sell records whose origin they will not name, this is the cleanest provenance position available. Off the top band because nothing states whether the underlying models were trained on material the vendor has rights to, which is the one provenance question that does apply to a generative product.

There is no third party platform in the path to be exposed to. Nothing here rents an account, rotates an identity, drives a browser extension, proxies a session, scrapes a professional network or relays mail through infrastructure governed by somebody else's agreement.

The buyer talks to a simulated counterparty inside the vendor's own product, and the only external dependency named is the avatar supplier, which is a service the vendor contracts for rather than a platform it operates against. Judged on the question this axis asks, whether the vendor has built its product on terms it does not control, the answer is a clean no. Recorded plainly rather than treated as a technicality, because in an alphabetical block where account rental was priced in public tier tables, a product with no platform dependency at all is a real structural difference.

The training question is answered directly and without qualification, which is rare enough in this index to be the finding: the vendor states it does not train artificial intelligence on the buyer's data, full stop. That wording is materially broader than the three precisely limited statements recorded elsewhere in this index, each of which ruled out one narrow class of model or one data source and left the general question open.

Alongside it sit a statement that offer details and call transcripts stay private to the buyer and a self service deletion route. Off the top band because all of it lives in a product answer rather than in a contract or a policy a buyer could hold the vendor to, no retention period is stated, and the position of the named avatar supplier with respect to recorded voice and video is unaddressed.

The person on the other end of the call is the customer, and they know exactly what they are talking to. Realism is sold openly as the product, the artificial nature of the counterparty is stated in the title, the description, the pricing tiers and the answers, and the claim that users forget they are speaking to a machine within 90 seconds is offered as a feature of a simulation the buyer chose to enter rather than as concealment from someone who did not.

No third party recipient exists anywhere in this product to be misled. Set against the vendors graded immediately before it in this block, several of which manufacture authorship, rent real identities and sell undetectability, this is the clean case the axis exists to distinguish, and recording it that way is what makes the low grades elsewhere meaningful.

The integration surface is empty. A connector, a programmatic interface, a webhook, a marketplace listing, a customer relationship platform destination, an export path for scores or transcripts, and a calendar or conferencing connection were all sought and none was located.

That is a material gap rather than a stylistic one for this particular product, because the whole proposition is a readiness score that tells a seller when they are prepared for a real buyer, and there is no published way to get that score, the transcripts or the trend out of the tool and into anything a coach, a manager or a system of record could use. Scores accumulate in a dashboard and stay there.

One corporate fact is published and it is more than several vendors in this session offer: the operating entity is named in the footer as a limited liability company. Beyond that the picture is blank. No registered address, no country, no processing or storage region, no transfer mechanism and no customer choice appear anywhere.

The gap is sharper here than at a text based tool, because the product records voice on every tier and video on the top one, and where a recording of a buyer's own face and voice is processed is a question a purchaser in a regulated market has to be able to answer.

The security surface is empty. A certification, an audit, a penetration test, a documented control set and a trust page were all sought and none was located, from a product that records the buyer's voice on every tier and their video on the top one and retains transcripts of those sessions. The only security adjacent statements on the page are that data stays private to the buyer and that it can be deleted from the dashboard, both of which are handling commitments rather than controls. Beta status explains this and does not change what the axis measures, and a published control set would move the row one band even with no certification behind it.

Three tiers publish a monthly price, an annual price, the consumption allowance that actually governs use, and an itemised list of what separates each from the one below. Voice at 19 dollars a month carries 100 voice minutes, Voice Pro at 49 carries 300, and Video Pro at 179 carries 300 voice and 200 video minutes plus the avatar. The annual saving is stated as roughly two months and the annual figures are printed rather than implied.

The vendor then does something almost nothing in this index does: it explains why the product costs money at all, stating that every call minute costs it real money in voice and processing, and ties each tier's price to what that tier consumes. Cancellation is stated with its consequence, that access continues to the end of the billing period, and the first call is free with no card.

Two annual only entitlements, minute rollover and custom personas, are disclosed as annual only rather than buried. One inconsistency is worth a buyer noting and does not reach the band below: the annual bonuses are described as four in the pricing block and three in the answers.

Two halves of this axis go opposite ways. Commercial exit is clean and stated: no contract, cancellation from the dashboard, access retained to the end of the paid period, and a free first call before any card is required. Deletion is better than clean, since the buyer can remove everything themselves from the dashboard rather than emailing a request, which puts this vendor ahead of much larger ones on that specific mechanism. Portability is where it stops.

No export function, file format or download path exists for transcripts, scores or the readiness trend, and those are the accumulating asset, so a seller who cancels after a year of practice keeps nothing they can take to a coach or a manager. Deleting everything and exporting anything are not the same capability and only the first is offered.

Nothing is sent and no sender reputation is at stake, so this axis has almost no purchase on the product and is graded on what is knowable, as the sibling rule requires. The nearest equivalent constraint is the minute allowance, which is a genuine consumption governor published per tier and enforced by the vendor rather than by a mailbox provider. The note that matters for a reader comparing this to its category neighbours is simply that the absence here is structural rather than a disclosure failure: there is no sending surface to discipline.

The buyer is named more sharply than most vendors dare, in the title, the description and the positioning: solopreneurs and coaches, explicitly not sales teams, with the product framed as the first practice tool built for that buyer. Naming a narrow buyer and declining the larger one is a real choice and it earns credit. The ladder underneath it has genuine shape, three tiers separated by consumption and by voice against video. Coverage is where it thins.

Language support, region availability, industry fit and any team or seat concept are all absent, and the vendor's own copy pulls in two directions on who this is for, since the scoring interface refers to knowing where reps need work and when reps are ready, which is manager language for a product that says it is not built for teams.