FloQast

Full profile

This profile covers FloQast’s offerings for the close: checklists, account reconciliations, journal entries, variance analysis and the AI features built into them. It addresses three of the 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. It does not cover FloQast’s separate risk and compliance products except where they govern the close features.

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

At a glance

FloQast describes itself as an accounting workflow automation platform "created by accountants for accountants," trusted by more than 3,500 accounting teams to automate close management, account reconciliations, accounting operations and compliance activities, according to its September 16, 2026 release. Its pricing page groups the close offering into two packages: Close Management, which includes the close checklist, account reconciliations, accounting operations, projects, an AI assistant and ERP integration, and Close AI, which adds AI agents, AI matching, automated reconciliations and journal entry posting. FloQast presents Close AI as putting AI to work across matching, reconciliations, journal entries and analysis, "with review and governance built into every step".

Vendor-stated capabilities

Close management

As described on FloQast’s pricing page, the Close Management package runs the close from one checklist, ties out and signs off reconciliations, standardizes operational and close work, tracks one-off projects and includes an AI Assistant to answer questions, prepare and review work. FloQast’s Close AI page says Journal Entry Management automatically creates and posts journals to the ERP, flags potential issues before they reach the ledger, enforces reviews and approvals, requires documentation to be attached and locks approved journals. FloQast’s Journal Entry Management page, read the same day, adds that the product can create journals automatically from data sources and "auto-certify reasonable entries", and that configurable approval workflows "ensure preparers can’t review their own work". The Close AI page also says the product can automate up to 80% of reconciliations, a vendor-stated figure for which the page gives no population or basis.

AI features

Each entry below states what FloQast’s own pages say. The kind of AI uses the categories in What We Mean by AI; where FloQast does not describe the technique, the classification is CoE analysis.

The sources read do not say how the auto-certification and preparer-reviewer controls apply to entries created by Transform agents.

Which AI providers sit behind these features. FloQast’s authorized subprocessor list, version 2.6 of June 2026 names Anthropic, OpenAI and Meta as providers of generative AI services for the FloQast application, Google for file and email hosting and generative AI services, TrueFoundry for AI infrastructure and Reducto for AI PDF parsing. Its change history shows the provider set growing over time: OpenAI added in November 2023, Anthropic in September 2024, Meta in July 2025, TrueFoundry in November 2025 and Reducto in June 2026. FloQast’s data processing addendum (the document titled "FloQast DPA 20260105") commits it to give notice of any addition, deletion or replacement on its list of authorized subprocessors at least 10 days before it intends to engage a new subprocessor, with an opportunity to object on reasonable grounds to the processing of personal data. The sources read do not say which provider serves which feature.

How customer data is used. FloQast’s AI Terms of Use (the document is titled June 2026) state that certain features use third-party AI applications and that inputs and outputs are customer data under the agreement. The terms state that neither FloQast nor its AI providers will use customer data to train or improve their models, and that FloQast has configured its AI provider integrations with zero-day retention so that no customer data is stored by the providers. FloQast’s master subscription agreement (Revision 2024.2) states that its services include these AI services and that the AI terms govern them in addition to the agreement.

Assurance and certification. FloQast’s Trust Center says customer administrators can access its audit reports and compliance resources inside the application, prospects can request them from their account executive, and bridge letters cover the period after a SOC report ends. The same page describes FloQast as "a pioneer and early adopter of ISO 42001," the AI management system standard, and says specialist firms complete penetration tests on high-risk products and infrastructure annually. FloQast’s master subscription agreement lists SOC 1 Type II, SOC 2 Type II, ISO 27001, ISO 27701 and ISO 42001 audits performed annually by third-party auditors, with current reports provided to customers on request. The governance page states that "every FloQast-built agent is audited against the COSO GenAI Framework, so your team can place some reliance on the controls we attest as your technology partner". FloQast’s September 16, 2026 release states that it has hired Lucia Wind, described as COSO’s Board Chair and outgoing Executive Director, who oversaw COSO’s generative AI publication, as its SVP of Risk & Audit Advisory.

Independent market evidence

Filings. FloQast does not appear as an SEC filer: an EDGAR company search for "FloQast" returned no matching company, and a full-text search on October 12, 2026 returned 35 filings, all by other companies. There is therefore no 10-K business description, AI strategy or risk-factor disclosure to report. One filing by another company is relevant: SAP SE’s annual report on Form 20-F for 2025 lists FloQast, Inc. among its equity investments with ownership of at least 5%. SAP also sells close and consolidation software.

Analyst placements. No analyst placement is reported here. FloQast’s September 16, 2026 release says FloQast "is consistently rated #1 across all user review sites"; that is a vendor-stated ranking and is not evaluated here. A placement FloQast cites would be vendor-stated unless the analyst’s own page is read.

Regulators and standards bodies. No regulator or standard-setter statement about FloQast was found. Two standards appear in FloQast’s own materials. ISO 42001 is described above as a vendor statement and as an annual audit in FloQast’s master agreement; no certificate or certification body was identified in the sources read. COSO’s publication Achieving Effective Internal Control Over Generative AI, released February 23, 2026, is the basis of FloQast’s governance product. That publication adapts COSO’s 2013 framework rather than creating a separate one, and it states that it contains general information only and that its views may differ from those of regulators. COSO’s own release announcing the publication quotes Lucia Wind as COSO’s Executive Director and Chair.

Independently verified outcomes

No independently verified outcomes were found. CoE | Vendor Intelligence searched SEC EDGAR full text, FloQast’s customer stories, which are vendor-stated, and public sources for academic or audit-firm studies of FloQast. FloQast’s website states that a Total Economic Impact study it commissioned from Forrester Consulting in January 2026 found 275% ROI and payback in under six months for a composite enterprise organization with 95 accountants; the study itself sits behind a registration form and was not read. These are vendor-stated figures from a study FloQast commissioned, modeled on a composite organization, and they are not independently verified outcomes.

Pricing and packaging

FloQast publishes no prices: its pricing page states "No Per-User Fees," says packages scale with business outcomes rather than seat count, and ends each package with "Contact Sales". Pricing is by quote as of that read. The same page places the AI Assistant in the Close Management package; AI agents, AI matching, automated reconciliations and journal entry posting in the Close AI package; and Risk & Compliance as a separate package available alone or with either close package. The page does not describe usage units, consumption limits or overage terms for the AI features.

FloQast’s master subscription agreement (Revision 2024.2) is the most recent dated version among the agreements listed on FloQast’s site, and no FloQast page read says which version current contracts use. That agreement makes the services available on a per-user basis, or per client where the order says so. The pricing page states "No Per-User Fees" and that packages scale "with your business outcomes, not your seat count". The two documents are not reconciled on the pages read; which basis a given order uses is a question for FloQast. Under the agreement, each order is non-cancelable and fees are non-refundable unless the order or agreement says otherwise, the subscription term is set in each order, and the agreement itself contains no automatic-renewal clause. Unless an order says otherwise, invoices are payable within 30 days of receipt. A customer may withhold a disputed portion if it notifies FloQast in writing within 30 days of receiving the invoice, works with FloQast in good faith to resolve the dispute and pays the undisputed portion.

Known limitations

• The customer carries the review obligation. FloQast’s AI Terms of Use require the customer to review and approve all outputs, state that the AI services are not a substitute for human review, warn that outputs may be inaccurate or biased, provide outputs as-is and exclude FloQast’s and its providers’ liability for inaccuracies or biases.
• The AI providers change. The subprocessor list records generative AI providers added in 2023, 2024, 2025 and 2026. FloQast’s data processing addendum gives at least 10 days’ notice before a new subprocessor is engaged, and its service level terms say FloQast operates on a continuous release cycle. The sources read do not say how customers are told when the model behind a feature changes.
• Availability is documented by package, not by release status. None of the FloQast pages read labels an AI feature as preview, beta or early access, but the September 16, 2026 release says FloQast "will soon operationalize" COSO’s framework in the platform as the COSO GenAI Module. FloQast’s master agreement provides beta services as-is, without warranty, excluded from the defined services, and lets FloQast suspend or end them without notice. Edition, region and language restrictions for the AI features are not documented in the sources read.
• Which parts of the governance offering are available now is not stated. FloQast’s product page and its September 2026 release describe the COSO governance capability differently.
• Contract terms on liability and data at exit. FloQast’s master agreement excludes each party’s liability for loss of use, inaccurate data, lost profits and consequential, indirect, special, incidental, reliance or punitive damages, and caps each party’s total liability for direct damages at the fees paid or owed for the services in the prior twelve months under the applicable order, subject to listed exceptions that include gross negligence or willful misconduct, indemnification, breach of the confidentiality obligations, and breach of the security measures or the DPA (Section 7.c). On termination, FloQast returns or destroys customer confidential information, which includes customer data, if asked within 30 days, other than copies kept in routine backups. No FloQast document read states how long FloQast itself retains AI inputs and outputs, which its AI terms treat as customer data.
• Integration prerequisite. ERP integration is part of the Close Management package. Agents that post depend on that connection and on the access it grants.

CoE analysis

This profile does not score or rank FloQast. What follows is what a finance team should establish before relying on these features, using the five questions in the close and consolidation playbook: evidence, judgment, authority, control and review.

Separate what suggests from what acts. The AI Assistant, AI Variance Analysis and Detect flag, explain and recommend. Transform agents and journal entry posting create and post entries in the ERP. The second group needs the treatment described in Your AI Agent Is an Unlisted Signer: under whose credentials an agent posts, what limits apply, and whether the person who builds an agent can also approve its output. FloQast says Journal Entry Management can auto-certify "reasonable entries"; which entries qualify, and whether entries created by agents can be auto-certified, are questions to put to FloQast.

A sign-off before an agent runs is not a review of each run. FloQast says an accountant signs off before an agent runs, that agents are tested against numbers the team already approved, and that the team can "review every result". The sign-off before go-live approves the design. Review of each run is a separate control, and FloQast’s AI Terms of Use place it on the customer. A team relying on an agent’s entries needs evidence of that review for each run that could catch an error, as set out in Review Is a Control Only When It Can Catch the Error.

Read "COSO-aligned" as the vendor’s description. FloQast’s pricing page says its built-in COSO compliance "ensures your AI usage is always compliant and will pass audit". The COSO publication describes itself as general information whose views may differ from regulators’. Whether a control passes audit is a conclusion for management and the external auditor, not a property a vendor can guarantee. The hiring of COSO’s chair is a fact to know when weighing that language, not evidence against the product.

Ask what "reliance" means in practice. FloQast invites customers to "place some reliance on the controls we attest". The COSO publication treats AI reliance as dependence on AI outputs as evidence for an ICFR control, which then calls for the same evidence standards expected for ICFR. FloQast’s master agreement lists annual SOC 1 Type II and ISO 42001 audits. Before relying, ask which report, issued by which auditor and for which period, covers the AI features, and whether it covers changes to prompts, models and providers. Internal Control Over Generative AI, Applied to the Close sets out the change controls to look for.

Test matching and anomaly detection on your own closes. Automatic matching and learned anomaly thresholds earn reliance through their accuracy and completeness on your data, not through the vendor’s description. Run them against prior periods whose answers are known, and treat unexplained output the way Treat AI Output Like an Unsupported Journal Entry describes.

Keep your own inventory. FloQast’s agent inventory lists Transform agents as they are built. The COSO publication’s roadmap calls for an inventory of all active and planned GenAI uses, including shadow AI. A vendor’s register covers the vendor’s agents; the organization’s inventory has to cover the rest.

Community Observations

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

Member input under review

None yet.

Relevant diligence questions

1. Which FloQast AI features are enabled in our environment today, which were on by default, and how will we be told when a feature, model or AI provider behind it changes?
2. For each Transform agent, what can it post or change in our ERP, under whose credentials, can its entries be auto-certified, and what stops the person who built it from approving its output?
3. Which AI provider processes our data for each feature, in which region, how long does FloQast itself keep AI inputs and outputs, and what evidence supports the zero-retention and no-training terms?
4. Does a current SOC 1 or SOC 2 report cover the AI Assistant, AI Matching, Detect and Transform agents, including change management for prompts, models and providers, and what exactly does FloQast attest about agents it builds?
5. What evidence will the governance module produce for our external auditor, which parts are available now, and has any auditor accepted that evidence?
6. What accuracy and completeness results can we see for AI Matching and Detect on our own historical data before we rely on them?

Evidence gaps and unanswered questions

• Product documentation. FloQast’s help center returned a bot challenge to text retrieval and could not be read, so this profile relies on product pages, legal documents and releases rather than user documentation.
• Consolidation and intercompany. The FloQast pages read describe close, reconciliation, journal entry and variance tools; whether FloQast offers consolidation or intercompany elimination was not established.
• Provider mapping. Which AI provider serves which feature, and where that processing takes place.
• Contract version. Which version of FloQast’s master agreement and data processing addendum governs current contracts, given that the agreement read makes services available on a per-user basis, or per-client where an order so specifies, and the pricing page says there are no per-user fees.
• FloQast’s own retention. How long FloQast keeps AI inputs and outputs, and whether they are kept outside the customer’s data region.
• ISO 42001. The certificate, its scope and the certification body.
• Assurance scope. Whether FloQast’s audit reports cover the AI features and agents, and what FloQast means by attesting to agent controls.
• Governance module timing. Which parts of the COSO module are generally available, given the different descriptions.
• The automation figure. The basis for "up to 80% of reconciliations".
• AI pricing mechanics. Whether agent or AI usage is metered, capped or charged separately within a quote.
• Outcome evidence. The Forrester study FloQast commissioned sits behind a registration form and was not read; its figures would remain vendor-stated if used.

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.