Conquest Planning

A definitional edge case, and the vendor's own words resolve it in both directions at once. The engine was the product from founding in 2018 rather than added to a platform that worked without it, so the go empty test is passed decisively: remove the strategy engine and no product remains, which is what separates this record from the three established planning platforms in this pocket that shipped their first agent in 2026. What prevents the top grade is what the engine is.

The company consistently describes it as an expert system and as an auditable, deterministic calculation engine, and its stated differentiator is precisely that it is rules based and reproducible rather than probabilistic. Under this index's convention that encoded human expertise executed deterministically is not inference, the core sits alongside the human built calculation engines graded elsewhere in this pocket. Genuinely inferential features exist around it, including natural language interaction and generated content, and they are the newer layer rather than the substance.

Supervision is the product's stated selling point rather than a paragraph on a governance page, and three distinct published capabilities support the grade. Every recommendation is stated to be traceable, with the vendor putting the point in the sharpest form found anywhere in this index, that a firm never has to tell an auditor the answer came from the model. A compliance trail is described as built into the expert system rather than layered over it.

And firm leadership receives real time analytics on plan quality, adviser activity and client engagement, which is supervision at the book level rather than the individual output level and is something no other vendor in this pocket offers. Determinism reinforces all three, because a reproducible engine can be re run to show how a recommendation was reached.

The one place the question stays live is the self directed channel, where a household receives planning with no human involved at all, and there the answer is that reproducibility substitutes for review rather than accompanying it.

An architectural answer to model risk is published and no measurement accompanies it. The engine is described as deterministic and auditable, which if accurate removes rather than mitigates a whole class of risk, since a reproducible system returns the same recommendation from the same inputs every time and can be re run to demonstrate how it got there.

Set against the rest of this pocket, where nothing at all is published, that is a substantially stronger position and the strongest of the structural approaches found. Held below the top grade for two reasons. The claim is made in marketing language with no technical account of what a traceable recommendation actually shows a reviewer, and the newer generative capabilities layered around the deterministic core, including natural language interaction and generated content, are not deterministic and nothing published addresses their behaviour at all.

This clears the bar by three separate routes, any one of which would carry the grade. Named institutional customers: major banks in two countries, a life insurer, a clearing business, a national brokerage and a digital wealth manager are identified by name.

A quantified business outcome attached to customer firms: onboarding of households bringing more than one million dollars in investable assets reported rising from 12 percent to 26 percent, which is an outcome measured on the customer's business rather than on time saved.

And the strongest form of independent party with money at stake this index recognises, since several of the institutions using the platform have also invested in the company, including a major bank, a global custody business, a large retirement provider and a global asset manager's venture arm. A named adviser at a named firm is quoted, more than two million plans have been created, and the company appeared on a global fintech ranking in 2026.

Searched for any statement on whether household financial profiles, plans or interactions are retained, used to improve the engine, or used to train, tune or evaluate any model, and located none in either direction. Two million plans built across three countries is a substantial body of real household financial data, and the platform's institutional deployment model raises a question no single tenant vendor faces, namely whether anything learned from the customers of one bank can inform behaviour seen by the customers of another. For an engine described as deterministic the answer may well be that nothing is learned at all, which would be a clean and easily stated position, and it is not stated.

Recorded as an unverified absence. Searched for a retention schedule, a deletion commitment, published processing terms or any account of the division of responsibility between the vendor and a deploying institution for customer data, and located none in the material opened.

The question is structurally more complex here than for a vendor selling directly to advice firms, because when the engine sits embedded inside a bank's own digital estate a customer may never know a third party processed their financial profile at all, and nothing published describes how that relationship is disclosed or governed. Queued check: a privacy policy and any processing terms.

Recorded as an unverified absence, and one where the absence is almost certainly in my reading rather than in the vendor. Searched the material opened for a named attestation, an information security standard, an audit report or a dedicated trust page, and located none, and no security page was reached.

A platform deployed inside major retail banks, a life insurer and a clearing business has by definition passed those institutions' vendor security reviews, which are among the most demanding in any industry, so published credentials very probably exist. This row should be read as work not yet done rather than as a finding. Queued check: a trust or security page.

A software supplier holds no licence and requires none, and no penalty attaches. The position is worth stating carefully because this vendor engages with its customers' obligations more deliberately than most, marketing an auditable trail and the ability to prove plan quality to regulators, and because one delivery mode raises a question the others do not: in the fully self directed channel a household receives suggested strategies with no adviser involved, and whichever institution deploys that channel owns the regulatory characterisation of what the customer just received. Searched for any statement about where the vendor considers the boundary between planning output and regulated advice to sit, and located none.

Searched for a fairness evaluation, any testing of recommendations across household types, or any account of how strategies are ranked, and located none. The exposure here is larger than for any adviser tool in this pocket and follows directly from the vendor's own ambition.

A single engine is stated to deliver the same planning depth from a mass market saver to an ultra high net worth household, and in the self directed channel it recommends strategies to people with no adviser to sense that something is off.

If the strategy ranking suits some household shapes better than others, the effect lands hardest on exactly the customers reached through the channel designed to widen access, and it lands at the scale of a thousand institutions rather than one advice firm.

Searched for a warranty, an accuracy commitment, a correction obligation or any allocation of responsibility between the vendor and a deploying institution, and located none. The traceability the vendor markets is a defence for the institution rather than a remedy for the household, since being able to show an auditor how a recommendation was reached establishes that a process was followed and says nothing about whether the outcome was right.

The self directed channel is where the gap matters, because a customer who acts on a suggested strategy without an adviser has no professional standing between them and the engine, and nothing published describes what recourse exists, who carries it, or how a person would even raise that a recommendation was wrong for them.

Searched the company material, the funding coverage and the trade press for a model provider, a model family, a version or an inference host behind the natural language interaction and the generated content features, and located none. The deterministic core is arguably self disclosing, since a rules based expert system built in house has no external model dependency to name and the vendor says as much by describing what it built.

That leaves the newer layer entirely unaccounted for: a platform that now offers conversational interaction and dynamic content generation is calling something to do it, and nothing published says what, which is the more consequential silence precisely because it sits beside such a clear account of the part that is not modelled.

The architecture is built to be embedded rather than connected to, which is the deeper of the two commitments, and the deployment record proves it works. An open two way interface lets an institution construct its own client and adviser experience over the planning engine while the engine continues to do the reasoning, so the product can sit invisibly inside a bank's own digital estate rather than appearing as a separate application.

Data gathering is designed to draw on what the institution already knows about a customer before asking them anything. The evidence that this is real rather than a claim is the installed base: more than a thousand financial institutions, including organisations whose integration and procurement requirements are as demanding as any in financial services. Queued check: an enumerated integrations directory, which would add named systems to what is currently an architectural account.

Recorded as an unverified absence, and residency is a live question for this vendor rather than a formality. The company is headquartered in Canada, holds most of the Canadian adviser market, and operates in the United States and the United Kingdom, so it serves three jurisdictions whose rules about where customer financial data may sit differ materially, and Canadian institutions in particular frequently require domestic residency.

Searched for a hosting region, a residency commitment, a per market data location statement or any account of deployment topology, and located none in the material opened. Queued check: a security or architecture page.

Searched for a rate, a tier, a per adviser figure or any account of how enterprise licensing is structured, and located none. The model is enterprise sales to institutions, where an absent public price is the norm rather than an evasion, since a thousand seat bank deployment and a small advice firm are not comparable purchases.

One change makes this worth revisiting: the company announced a direct channel for independent advisers in 2026, and a self serve channel of that kind usually arrives with published pricing where the enterprise business never did. Queued check: any pricing published for the independent adviser channel.

The broadest coverage located in this pocket by a wide margin, and it is broad along four independent dimensions rather than one. Institution type: major retail banks in two countries, a life insurer, a clearing and custody business, a national brokerage and a digital wealth manager are all named users, which is a genuinely different buyer set rather than a size range within one.

Geography: three national markets with different tax systems and retirement regimes, holding a substantial majority of the adviser market in the home country. Client segment: the same platform is stated to deliver equivalent planning depth from a mass market saver to an ultra high net worth household. Delivery channel: adviser led, hybrid and fully self directed from one engine, which very few planning vendors attempt at all. More than 60,000 advisers across over 1,000 institutions is the scale evidence behind it.