Trintech

Full profile

Trintech’s close platforms include Cadency and Adra. It describes an assistant, Beacon, embedded in Cadency, a set of AI agents delivered through its AI Platform, and a separate set of AI capabilities in Adra. This profile sets out what Trintech says these features do, what independent evidence exists, and what a finance team should establish before relying on them.

Which AI this covers: AI built into Trintech’s finance platform, as Trintech documents it: generative features, agents and established machine learning or matching automation. It does not cover general-purpose assistants a team might use alongside Trintech. See What We Mean by AI.

This profile covers three of the four kinds of AI described in What We Mean by AI: AI built into finance platforms, AI agents and automated actions, and established machine learning and rules-based automation. Product statements come from Trintech’s own pages, releases, brochures, compliance documents and published contract terms. Trintech’s product documentation and release notes are in its Success Center, which could not be read as text, so the profile rests on what Trintech publishes openly.

At a glance

Trintech presents Cadency as a financial close platform for large enterprises, covering balance sheet reconciliations, transaction matching, close management, journal entry, governance, risk and compliance, and intercompany accounting. It describes the Adra suite as automating reconciliations, transaction matching, close task management and journal entry with AI-assisted workflows, and Adra’s native AI options as supporting reconciliations, transaction matching, task management, variance analysis and support. Trintech, headquartered in Plano, Texas, describes its platform as automating reconciliation, transaction matching, close management, journal entry, intercompany accounting and compliance. Trintech also lists ReconNET, Frontier, Accurate and DATAFlow among its products; this profile does not cover them.

Vendor-stated capabilities

Close capabilities

• Reconciliation and matching. Trintech says Cadency notifies accountants of items that need attention together with applicable risk factors, and creates exceptions automatically for unmatched items, which are routed through configurable notification and resolution processes.
• Close management. Cadency orchestrates system and close activities in a workflow designed to show bottlenecks and create actions for issue resolution.
• Compliance and audit support. Supporting material is captured in an online "eBinder," and auditors have secure access to a centralized record.
• Intercompany. Cadency provides a workflow for comparing and reconciling transactions between entities as part of the close.
• Journal entry. Cadency Journal Entry lets customers configure viewing and editing rights, validates entries for accuracy and completeness before they move to the next workflow stage, and connects to SAP, Oracle and NetSuite through certified pre-built connectors.
• ERP integration. Trintech lists pre-built Cadency connectors for SAP, Oracle ERPs and NetSuite, along with APIs and data-as-a-service.
• Adra. Trintech describes Adra Matcher for automated multi-way transaction matching, Adra Balancer for balance sheet reconciliations, Adra Task Manager for close checklists, Adra Analytics for reports and dashboards, and Adra Journal Entry for creating, posting and reporting journals. It lists a pre-built NetSuite connector for Adra, an embedding in Microsoft 365, and an integration with Planful that it presents as running "from close to consolidate to report and plan".
• Consolidation support. Trintech’s consolidation page describes automated matching across source systems and consolidation inputs, variance and flux analysis across entities, currencies and reporting hierarchies, close tracking across entities and regions, and automated intercompany matching and elimination to remove internal activity from consolidated results. The page names no Trintech product or module as the consolidation solution. Its opening line reads "Automate, standardize, and control the entire FP&A consolidation and close process," followed by "From multi-entity eliminations to management reporting".

Security and assurance statements

Trintech states that it has held ISO/IEC 27001 certification since 2021 and names a certification body’s directory where third parties can check its status. It states that re-certification is by audit every three years, with an annual surveillance audit. It states that it maintains an SSAE 18 SOC 1 Type 2 attestation and an annual SOC 2 Type 2 attestation covering security, availability and confidentiality, with monthly bridge letters between audit periods. Its SOC 2 report covers October 1 to September 30, and its SOC 1 reports cover three twelve-month windows ending September 30, March 31 and June 30. Trintech says it updates its SIG and CAIQ security questionnaires quarterly. Its Trust Center lists TX-RAMP Level 2 certification for Cadency and ReconNet.

On AI specifically, the Trust Center says AI features are opt-in, that users can see how the AI reached a recommendation, and that every model update goes through a request-for-change and rollback process. It states that customer data is never used to train general-purpose AI models and that people oversee AI outputs. It says Trintech monitors its models monthly for accuracy, bias and drift and keeps logs, approvals, audits and data protection impact assessments for every model and use case. It describes its AI governance framework as built with ISO/IEC 23894 and the NIST AI Risk Management Framework in mind. These are Trintech’s statements; CoE | Vendor Intelligence has not seen the certificates, the reports or the governance records.

AI features

The kind of AI and the suggests-or-acts reading in this table are CoE classifications based on Trintech’s descriptions; Trintech calls several of these features agents even where the documented function is to produce a draft for review. Trintech’s Agentic AI page describes the agents as "Workflow execution, not task suggestion," completing finance work end to end with review and approval points. The "who controls it" column reports only what Trintech documents; beyond the Trust Center’s general statement that AI features are opt-in, Trintech does not say feature by feature who switches each one on.

Across the agents, Trintech describes a move "from AI that surfaces insights to AI that executes the work itself," within existing controls, approvals and audit trails. Its Agentic AI page says the agents learn from how teams match, reconcile and resolve work, and operate across the Trintech platform or alongside ERP and third-party systems.

Independent market evidence

No current SEC filings by Trintech. EDGAR’s company index lists no current registrant named Trintech: its only entries are for Trintech Group PLC and its ADRs, and that company’s last filing was a Form 15-12G terminating its registration, filed January 13, 2011. There is therefore no 10-K business description, AI strategy statement or risk-factor disclosure to read, as there would be for a public vendor.

Lender filings. Trintech, Inc. appears as a borrower in Morgan Stanley Direct Lending Fund’s quarterly report for the period ended June 30, 2026, which lists first-lien debt in Trintech, Inc. maturing on January 28, 2033, together with an unfunded revolver and a delayed-draw term loan. A lender’s schedule of investments records the lender’s holdings; it says nothing about Trintech’s products, AI features or operating results.

Analyst placements. CoE | Vendor Intelligence did not read an analyst firm’s own public page on Trintech for this profile, so no analyst placement is shown.

Regulators and standards bodies. No regulator or standards-body publication about Trintech’s AI features was found. Trintech’s statements about its certifications appear above as its own statements; for ISO/IEC 27001, Trintech names the certification body’s directory where its status can be checked.

Independently verified outcomes

No independently verified outcomes were found. CoE | Vendor Intelligence read the Trintech documents listed on this page and one lender filing that mentions Trintech, and searched the web for independent studies or audits of Trintech’s AI features; the search returned user-review listings, comparison articles and a return-on-investment paper prepared with Trintech, and no outcome measured or attested by a party independent of Trintech.

The outcome figures Trintech publishes are vendor-stated, and CoE | Vendor Intelligence has not verified them. Its Cadency page shows a 75% reduction in time spent preparing and reviewing journal entries, and its Cadency Journal Entry page gives "up to" 40% for the same measure; neither states the population, baseline or method. Its Adra AI tip sheet says that teams have more than doubled their automatic match rates "in just one month," without a population or baseline.

Pricing and packaging

None of the Trintech documents cited here states a price for Cadency, Adra, Beacon or the AI agents, and Trintech’s agent announcements direct existing customers to their account manager and new customers to an AI specialist. In October 2026, Trintech’s website had no pricing page and its FAQ stated no price. On this evidence, Trintech’s pricing is available by quotation, so there is no list price to compare with negotiated terms. The documents also do not say whether Beacon, the agents or the Adra AI capabilities are included in an existing subscription, sold separately, or limited by edition or region.

Known limitations

• Model providers and their locations. Trintech’s list of sub-processors, dated August 2026, names Microsoft Azure for Beacon’s model-related functionality and OpenAI as a failover for Beacon, and gives the United States as the location of both. It states that Azure supports Flux’s model-related functionality effective October 6, 2026, replacing Dell. The same list shows Cadency hosted in the United States, the United Kingdom, Germany and Australia. The Beacon brochure states "Your data stays in your system with no risky transfers" and offers a single-tenant option in a private cloud. Where Beacon and Flux process prompts and data is a question to put to Trintech.
• Sub-processor changes. Trintech’s data protection addendum for Adra, which its Adra terms page lists without a legacy label, and its data protection addendum for Cadency (version June 2025), which its Cadency terms page labels "Legacy (through 2025)," each require written notice of an intended sub-processor addition or replacement and let the customer object within 15 days of the notice if it reasonably determines that the change is contrary to the addendum or applicable privacy laws; neither addendum states a notice period before the new sub-processor is engaged. If the customer does not accept Trintech’s recommended change, or Trintech makes none within 15 days of the objection, the agreement is deemed terminated insofar as the new sub-processor would be used, with a refund of prepaid fees for unused services. The addenda concern the processing of personal data; whether a change of model provider is notified under this clause is a question for Trintech.
• Adra model processing is not described. The sub-processor list names model-related sub-processors only for Cadency features and none for Adra. How the Adra AI capabilities are delivered is not documented in the materials read.
• Training, learning and usage data. The Trust Center’s commitment covers training of general-purpose models. Trintech also says its agents learn from how teams match, reconcile and resolve work, and that the Accruals Intelligence Agent tracks its own accuracy by preparer and account. Both data protection addenda, the Adra addendum and the Cadency addendum labeled "Legacy (through 2025)," state that Trintech "may process Services usage data as a controller to provide, maintain, and improve the Services, to the extent permitted by Applicable Privacy Laws," in accordance with Trintech’s privacy policy. The documents read do not say whether usage data includes prompts and outputs, or whether one customer’s data shapes models or rules used for other customers.
• Breach notice, backups and deletion. The Adra security addendum, which applies to production environments unless the order says otherwise, commits Trintech to notice within 72 hours of becoming aware of a security breach that could expose customer data, keeps backups for up to 30 days, removes customer data and encryption keys within 30 days of a written request or termination notice, and sets a recovery point of 4 hours and a recovery time of 12 hours. The Cadency security addendum (version March 2023), which Trintech’s Cadency terms page labels "Legacy (through 2025)," sets 48 hours for breach notice and, for production environments unless the order says otherwise, backups for up to 28 days, the same 30-day removal, and a recovery point of 1 hour and a recovery time of 4 hours. Both addenda say Trintech provides its SOC 1 and SOC 2 reports on request, limited to the most recent applicable audit and one request in any rolling twelve months except after a security breach with the potential to expose subscriber data. None of the four addenda refers to AI features or model training, so they should not be read as AI-specific terms.
• Assurance scope. The Trust Center describes the SOC 1 and SOC 2 reports and their periods without saying which products or features they cover. Whether the AI features and agents are in scope is not stated.
• Packaging and availability. The June 2026 release describes its two agents as available; the September 2026 release gives no availability status for its three, and the documents read state no edition, region or language limits for the AI features. The Beacon brochure mentions voice and multilingual use without listing languages.
• Nature of the evidence. The capability statements on this page come from marketing pages, press releases and brochures rather than technical documentation, which Trintech keeps in its customer portal.

CoE analysis

Trintech’s AI features fall into three groups, and each needs different evidence before finance relies on it.

Drafts that a person approves. As Trintech describes them, Beacon’s journal entries, Flux and Variance narratives, and Accruals recommendations arrive as proposed work for a reviewer. The control is the review, and an approval counts as evidence when the reviewer could have caught what the draft got wrong, as discussed in Review Is a Control Only When It Can Catch the Error. A recommended accrual or a generated journal entry should be supported like any other entry before it is booked, including evidence that nothing was left out, as set out in Treat AI Output Like an Unsupported Journal Entry.

Features that act. The Data Access Agent retrieves data from other systems, including login-only and MFA-gated sources, and delivers it into the close. Beacon, by Trintech’s description, executes commands once a person approves them. Both belong in the organization’s authority register: whose credentials the agent uses, what it may do without a further approval, and how that limit is enforced in the systems it touches, as discussed in Your AI Agent Is an Unlisted Signer.

Scoring and matching. The Risk Rating Engine and AI-assisted matching rules are closer to established automation, but a rule or score that a model proposes still changes what reviewers see and what auto-matches. Finance should know who confirmed each rule and when it last changed.

Model changes. Trintech states that model updates go through a request-for-change and rollback process, and its data protection addenda (for Adra, and for Cadency in the addendum labeled "Legacy (through 2025)") give 15 days to object, on stated grounds, to a new sub-processor. Its August 2026 sub-processor list, for example, states that Azure supports Flux’s model-related functionality effective October 6, 2026, replacing Dell. Internal Control Over Generative AI, Applied to the Close explains how existing change-management and vendor-oversight controls extend to changes of this kind. Generated variance and fluctuation narratives are also claims about the numbers, and the presentation questions in A Chart Is a Claim apply to them as much as to charts.

CoE | Vendor Intelligence does not score or rank Trintech against other vendors. The useful comparison is between what Trintech documents and what a finance team needs to evidence for its own controls.

Community Observations

Not yet published. Community Observations are added after a licensing review.

Member input under review

None yet.

Relevant diligence questions

1. Which AI features are switched on in our Cadency or Adra tenant today, which need us to opt in, who in our organization can enable them, and how will we be told when a release changes what they do?
2. Where are Beacon and Flux prompts, customer data and outputs processed, in which regions, and does the 15-day sub-processor objection right in our data protection addendum apply to every change of model provider?
3. Is our data, including the usage data the addenda let Trintech use to improve the Services, used to train, tune or improve any model or rule, including the learning the agents do from our matching and reconciliation activity, and is it combined with other customers’ data?
4. Do the SOC 1 and SOC 2 reports Trintech provides on request include Beacon, the agents and the Adra AI capabilities in scope, and which controls in those reports cover model changes?
5. For each agent, what can it do without a person’s approval, which steps does it hold for review, whose credentials does it use, and what does the audit trail record about its inputs and the model version used?
6. How are Beacon, the agents and the Adra AI capabilities packaged and priced: included in the subscription, sold as add-ons or metered by usage, and limited by edition, region or language?

Evidence gaps and unanswered questions

• Trintech’s product documentation and release notes are in its customer portal and were not read; this profile relies on public pages, releases, brochures, the sub-processor list and published contract terms.
• Beyond naming model sub-processors, no public document read describes the models behind Beacon, the agents or the Adra AI capabilities, how they are evaluated, or how model versions are recorded.
• No independently verified outcomes and no analyst page were found.
• Trintech’s consolidation page describes matching, flux analysis, close tracking and intercompany eliminations that support consolidation; neither that page nor the product documents read describe a consolidation ledger or a group consolidation reporting module, so teams evaluating consolidation should ask Trintech what is in scope.
• Trintech’s Cadency terms page labels the Cadency data protection and security addenda read for this profile "Legacy (through 2025)"; its Adra terms page lists the Adra addenda without a legacy label. Which addenda apply to Cadency orders signed after 2025 is not stated on the pages read.
• The documents cited do not state who owns Trintech.
• Pricing, packaging and edition limits for the AI features are not published.

Right of reply

Organizations referenced on this page may respond under our Editorial Independence & Right of Reply policy.

Review status

This profile was last reviewed on October 12, 2026. Vendor-stated information is rechecked monthly. It is not legal, accounting or procurement advice.