Binocs

The removal test leaves an empty page. Every deliverable the company sells is generated, from due diligence reports and investment memos through market sizing, competitive grids, target scoring, growth strategy frameworks, offering memoranda and credit assessment memos.

The architecture is explicitly the product, described as specialised agents that collaborate in a structure mirroring a real investment team or committee, and the human expert layer is offered as an optional delivery mode rather than a component, which confirms the models are doing the work rather than assisting someone who is.

Two delivery models are published as a deliberate choice rather than a progression, one a fully self serve engine and the other a hybrid pairing the agents with expert human support for tailored analysis and strategic problem solving, so a firm selects its own level of machine autonomy per engagement. Naming the human layer as a distinct commercial mode is more honest than most vendors manage, since it concedes there is work the agents alone should not be trusted with.

The tension sits in the word optional: the default self serve path produces investment memos, credit assessments and offering documents with no human review at all, and nothing describes a confidence signal, a flag for thin source material, or any gate before that output reaches a committee.

Outputs are citation backed, so an analyst can trace an assertion to the document that produced it rather than accepting it, and that is the single most effective control available for generative analysis and the property that earns Daloopa and Reflexivity their standing. The optional expert layer provides a second checking mechanism where a firm elects to use it. Against that, the published measurement is not credible as stated.

Claiming over 98 percent accuracy and zero hallucinations simultaneously is internally inconsistent, neither figure carries a methodology, test set or definition, and for a product generating investment memoranda a specific and defensible error rate would be worth more than an absolute claim that cannot be true.

Founded in 2022 and no customer is named anywhere, no client count is published, and no funding was located. The published claims are unverified and one pair of them will not survive scrutiny: accuracy is stated at over 98 percent while hallucinations are stated at zero, and those two figures contradict each other, since 98 percent accuracy is a description of the 2 percent that is wrong.

Cost advantage is claimed at up to 90 percent below traditional consulting, which is a benchmark comparison rather than a measured result. The team page shows a small organisation weighted toward engineering and quality assurance.

No data boundary statement was located, and the exposure is the one recorded for ToltIQ and before it for Rogo. Serving private equity firms, investment banks and strategy consultancies on one platform means the same asset will be analysed for opposing parties, sometimes simultaneously, and a sell side memorandum generated for one client concerns a company a buy side client may be screening. Nothing describes separation between engagements or whether curated data and learned patterns are shared across them.

A privacy page is maintained and no data protection agreement, retention schedule, subprocessor list or deletion commitment was located within accessible material. The payload is confidential transaction documentation, since offering memoranda and target financials are material non public information about live deals, and the platform additionally ingests them alongside curated third party data. Nothing published states how long a deal workspace persists, whether it is segregated per client, or what happens to a target's documents when a bidder withdraws.

A service organisation control type two certification is stated and a dedicated security page is maintained, which puts this ahead of most vendors in the deal analysis lane, including ToltIQ, where no attestation appears at all. That matters commercially as much as substantively, since a private equity firm will not route a target's confidential data room through a third party without one. Held at B because only one framework is named, no trust centre or report request process is described, and no detail on encryption, key management or engagement segregation was located.

No supervisor, statute or instrument is named, and one output makes that omission consequential. Generating sell side confidential information memoranda means producing the marketing document for a securities transaction, which in most jurisdictions is regulated activity with rules on accuracy, balance and the presentation of projections, and credit assessment memos feed lending decisions with their own documentation standards. The company holds no licence and needs none as a technology supplier, but nothing addresses the obligations attaching to the documents it produces or how a firm using them discharges them.

No consumer decision applies, so the axis adapts, and the sharpest exposure is at the front of the funnel. Thesis aligned sourcing with fit scoring and prioritised shortlists means the model decides which companies a fund looks at in the first place, so any skew in coverage or in what the training material treats as an attractive business determines which firms attract capital and which are never seen.

Market sizing and competitive benchmarking embed further assumptions presented as analysis. One capability is circular enough to note: the platform assesses a target's exposure to disruption by artificial intelligence, which is a model judging which businesses models will displace. Nothing published describes source coverage, scoring criteria or how thinly evidenced markets are handled.

No guarantee, indemnity or falsifiable accuracy commitment was located, and the published accuracy claims work against the vendor here rather than for it, since an absolute assertion of zero hallucinations is not a commitment anyone could enforce or the vendor could honour. Citation backing does give the user a real means of checking any individual assertion before relying on it, and the expert layer offers a paid route to human verification.

Nothing describes correction of an error found after a memo has gone to committee, notification when a source was misread, or what the vendor owes a client whose diligence rested on a faulty conclusion.

The agent architecture is described in some detail, with specialised roles collaborating in the shape of an investment team, but nothing underneath it is disclosed. No model provider is named, no hosting arrangement or subprocessor list appears, and the curated third party data that feeds market sizing, benchmarking and competitive analysis is referred to only as curated, so a buyer cannot establish whose data supports a market estimate or whether the licences behind it permit the use being made of them.

One integration capability is described, capturing and syncing enriched deal data into a customer relationship system for pipeline tracking, and no such system is named. Nothing else appears: no data room, document repository, portfolio monitoring platform or market data provider is identified as a connection, the curated third party data underpinning market analysis is unattributed, and no developer documentation or interface reference was located. For a product whose value depends on ingesting a firm's deal flow, how that material arrives is left entirely undescribed.

No hosting provider, region selection, residency commitment or private deployment option was located. The company operates from India with a presence in the United States and sells to funds and banks in both regions, so where confidential transaction documents are processed and stored is a question a client's own information security review will raise before any deal material moves.

A free entry point is published and self serve, so a prospective buyer can use the product before speaking to anyone, which in a category where every competitor gates access behind a demo request is a materially better position and lets a firm test output quality on its own material. Cost is also framed against a named alternative, at up to 90 percent below traditional consulting engagements, which gives a buyer a reference point for the value case. What is absent is the rate itself, with no indication of what the paid tiers cost or whether charge falls per report, per user or per deal.

The buyer set is wide for a company this size, spanning private equity, private credit, venture debt, venture capital, investment banks, banks and non bank lenders, corporate development teams and strategy consultancies. Coverage follows the deal lifecycle end to end rather than a single moment, from thesis aligned sourcing and inbound screening through commercial diligence and financial analysis to value creation planning and exit preparation, with a separate private credit line handling covenant tracking and portfolio monitoring after the deal closes. The limits are function and evidence: this is analytical work rather than operations, and no market or client is demonstrated.