Comparison
Provenir vs ArboRule: the incumbent, or the one you can try.
Provenir has been doing this at a scale we have not. What it does not have is a way for you to find out whether it fits before a procurement cycle.
The short answer
Choose Provenir when
- You are lending across borders. They claim 120-plus data partners offering thousands of services and customers in 60-plus countries — one contract instead of a bureau relationship per market.
- Your bank cannot use multi-tenant SaaS. They document a dedicated virtual private cloud per client on AWS, a VPN so on-premises resources can participate in cloud decisioning, and a split where development runs in their cloud and production on your infrastructure.
- Governance is a procurement requirement. SOC 2 Type II, ISO 27001 and ISO/IEC 42001 for AI management, with compliance as a named platform module.
- You want the vendor's data science team involved. Provenir AI proposes models and rules, and custom models are built in partnership with their team rather than left to you.
- Scale itself is reassuring. They state 4-plus billion decisions a year across 120-plus financial services providers, with offices in six regions.
Choose ArboRule when
- You want to know the price. Ours is on the pricing page; Provenir publishes none and every route is sales-gated.
- You want to read the API docs. There is no public developer portal or documentation site for Provenir, so nothing technical can be verified independently.
- You want to build something today. Sign up, draw the flow, run real cases through Sandbox — no meeting required, and no four-week implementation.
- The team is small. Priced by decision rather than by seat or by implementation, with no professional-services engagement to schedule.
- You want your own data relationships. We orchestrate calls on your credentials rather than reselling data, which keeps the bureau contract yours.
Side by side
Every row links to where it came from.
Claims about Provenir are taken from Provenir’s own documentation, pricing and repositories — never from a review site, and never from us. Follow the link in the left column to check any of them.
| Criterion | Provenir | ArboRule |
|---|---|---|
| The productprovenir.com/platform read 2026-07-30 | different model A decision intelligence platform consolidating data, proprietary AI models, analytics and decisioning agents, sold as four modules: AI decisioning, data marketplace, compliance and case management. | One product, not a module set. Decisions, data connections, history, cases and review are the same system at one price. |
| Pricingprovenir.com/contact read 2026-07-30 | not in the product Not published anywhere on the site. Every path is sales-gated behind a meeting request; there is no self-serve signup and no free trial. | US$ 500 a month on the Startup plan, including 50,000 decisions across Sandbox and Live. Priced by decision, not by seat. Building and testing is free up to 10,000 decisions a period. |
| Documentationprovenir.com/platform read 2026-07-30 | not in the product There is no public developer portal or documentation site. Everything technical — the API, the editor's capabilities, versioning semantics, trace contents, retention — is behind a sales conversation. | Public, including the decision and history endpoints, request and response shapes, API key scopes and a per-flow OpenAPI 3.1 document. |
| Where it runsprovenir.com — unified cloud solution read 2026-07-30 | in the product The current site says API-first and cloud-native with mission-critical SLAs. An older page still live documents three platform models on AWS with a dedicated virtual private cloud per client, a VPN so on-premises resources can participate, and a hybrid option where production runs on your own infrastructure. | Hosted only. There is no self-hosted, on-premises or bring-your-own-cloud option, and no published certification — the controls are an append-only audit log, role-based access, scoped API keys, configurable retention and SSO/SCIM. |
| Time to liveprovenir.com — unified cloud solution read 2026-07-30 | different model They claim new use cases can launch in as little as four weeks, citing a customer live within ten weeks at 57% below anticipated cost. Implementation is a project, with people attached. | Signup to a running Sandbox flow is the same afternoon. Going live is publishing a version and pointing traffic at it. |
| Authoringprovenir.com — decisioning read 2026-07-30 | in the product A visual drag-and-drop interface for creating, modifying and publishing decisioning flows, positioned so business and technical users collaborate with minimal code and business teams change strategy without an IT queue. Decision tables are not named as an artifact anywhere on the site. | A visual canvas with 22 node types. Rule, Decision Table, Scorecard and 2D Matrix are grid editors a policy owner can work in; operand fields also accept Python, and the Code node is Python only. |
| Rollout and testingprovenir.com — decisioning read 2026-07-30 | in the product Promotion between development, QA and production with built-in testing, simulation and code management, single-click deployment, and champion-challenger testing for refining fraud strategies. Explicit rollback and backtesting against historical decisions are not published. | Integer traffic weights across published versions, summing to 100, sticky per grouping_id so an application keeps the version it started on. No shadow mode: a challenger serves real traffic or none. Build a dataset from decision history — past requests as inputs, recorded outputs as expectations, every external call pinned to the response the provider actually returned — then run it against any version. Up to 500 cases over a 366-day window, executed serially. |
| Decision historyprovenir.com/platform read 2026-07-30 | different model Claimed clearly — every decision traceable, explainable and auditable, with full audit trails for regulatory compliance, and field-level audit in case management. Retention, query interface and trace contents are not published. On adverse action they publish a technical position on deriving reason codes from SHAP values rather than a product feature. | Every run stores its input, output, the version that produced it and a node-by-node trace, including each external call's URL, status and latency and each model's SHA-256. Queryable at GET /history/api/v1/decisions. Retention is per-workspace, five years by default. |
| External dataprovenir.com — data read 2026-07-30 | in the product A data marketplace of pre-built integrations to credit, identity and fraud data behind a single API, claiming 120-plus partners offering thousands of global data services. Note the number counts partners, not individual sources. | 28 packaged providers, six databases, and any HTTPS/JSON API through a Custom Connection. You bring your own credentials for each: the platform orchestrates the calls, it does not resell the data. |
| Human reviewprovenir.com — case management read 2026-07-30 | in the product Case management integrates manual intervention into automated decisioning flows, with an exception queue, fraud alerts and referrals across the credit lifecycle, plus configurable reports and screens by status, decision metric, lead source and product line. | A Manual Review node suspends the decision and puts it in a reviewer queue; Case nodes record work without blocking the run. Case queues carry SLA policies pinned onto each case at creation, with business-hours clocks and escalation rungs. |
| Modelsprovenir.com — AI read 2026-07-30 | in the product Both, plus their own. Provenir AI auto-learns from data and proposes models and rules, custom models are built with their data science team, and they claim expanded explainability for Python and other models running on-platform. | ONNX models up to 32 MB, immutable once uploaded and re-verified by SHA-256 at scoring time, executed in-process. One graph input, CPU only. AI and Agent nodes call OpenAI, Anthropic or Google models inside the decision, on your own API key. No key, no AI. |
This is not a like-for-like comparison, and pretending otherwise would be silly
Provenir states 4-plus billion decisions a year, 120-plus financial services providers, 60-plus countries and six offices. They have been selling to banks for a long time, they hold SOC 2 Type II, ISO 27001 and ISO/IEC 42001, and they employ people whose job is to build your model with you.
ArboRule has none of that. No certification, no data science team you can borrow, no offices in six regions, no implementation practice. If your evaluation criteria are institutional — vendor scale, references at your size, a signed audit report — the comparison is short and Provenir wins it.
What we would push back on is the assumption that those criteria are the ones that predict whether the policy will actually be good. They predict whether the purchase will be safe, which is a different question, and the two get conflated constantly in this category.
Vendor scale predicts how safe the purchase is. It does not predict how quickly your risk team can change a cut-off.
The thing you cannot check
Provenir publishes no developer documentation. There is no public docs site, no API reference, no schema. Their site was recently redesigned and several older technical pages now return 404, so the most specific deployment and model-format detail survives only on legacy URLs.
That means the claims in the middle of the table above — that every decision is traceable, explainable and auditable; that promotion between environments has built-in testing and simulation; that champion-challenger testing exists — are marketing copy that we have quoted accurately and cannot verify. They may all be entirely true. You will find out during an evaluation, not before one.
Compare that with what you can establish about ArboRule in ten minutes without talking to anyone: the price, the endpoints, the request and response shapes, the API key scopes, the retention default, and that there is no certification. Some of those facts are unflattering. Publishing them is the point.
Data: whose relationship is it
Provenir's strongest asset is the data marketplace — 120-plus partners reachable through one API and one contract, across 60-plus countries. Building that is slow and expensive, and for a lender entering several markets at once it is the single best reason to choose them.
ArboRule takes the opposite position, and it is a position rather than an absence. We ship 28 packaged providers plus any HTTPS or JSON API through a Custom Connection and six database connectors, and you use your own credentials for every one. We do not resell data and we do not sit in your commercial relationship with a bureau. That means more procurement work for you at the start, and it means changing provider is your decision, priced by them, without renegotiating with us.
Which is better depends entirely on whether you already have those relationships. If you do, an intermediary adds a margin and a dependency. If you do not, an intermediary is a year of work you skip.
Questions
How much does Provenir cost?
Not published. There is no pricing information anywhere on provenir.com, no self-serve signup and no free trial; every route is a meeting request. Expect an enterprise contract with implementation attached — they cite new use cases launching in as little as four weeks and a customer live within ten.
Does Provenir have public API documentation?
No. There is no public developer portal or documentation site, so the API, the editor's capabilities, versioning behaviour, trace contents and retention periods cannot be checked independently. Everything technical is behind a sales conversation, which is worth planning for if your evaluation depends on answering technical questions early.
What is a good Provenir alternative for a smaller lender?
It depends which part you need. If it is the global data layer — bureaus and identity providers across many countries under one contract — nothing small replaces that, and Provenir or Taktile is the honest answer. If it is the decisioning itself, with versioned policies, weighted rollout, queryable history and a referral queue, then ArboRule covers that at a published price with public documentation and no implementation project.
Last verified July 30, 2026 against Provenir’s public documentation. Products change and pages move; if something here is out of date, tell us at hello@arborule.com and we will correct it.