Captioning financial services training: FINRA compliance obligations, broker-dealer regulatory training vocabulary, and ADA requirements for Series licensing, AML/BSA, and continuing education

Why financial services training captioning is different

Most industries produce training video because employees need skills. Financial services firms produce training video because regulators require it. This is not a subtle distinction. When FINRA examines a broker-dealer, one of the standard examination document requests is the firm’s written supervisory procedures and evidence that registered representatives have completed the required training under those procedures. When the SEC examines an investment adviser, the examination staff reviews training records as part of assessing whether the firm has implemented an adequate compliance program under the Investment Advisers Act. When the NY DFS examines a covered entity under 23 NYCRR 500, the examination includes evidence that required annual cybersecurity training occurred. When an EU financial services firm is reviewed under MiFID II, the competence and suitability framework requires documented training history.

Training documentation is not incidental to the regulatory process in financial services — it is a primary examination category. And training documentation in a video-based training program includes the caption track. A caption track that fails to accurately reproduce regulatory terminology is a training record that does not accurately document what the training covered. This creates a compliance exposure that does not exist in the same form in other industries: a financial services firm with inaccurate captions on its regulatory training content has a documentation problem that is simultaneously an ADA Title I problem, a FINRA supervision problem, and a SEC books-and-records problem.

Three characteristics of financial services training content make this problem structurally different from the general corporate L&D context:

Vocabulary density. Financial compliance training is the most acronym-dense content category in corporate learning. A 30-minute AML/BSA module may contain SAR, CTR, FinCEN, 314b, BSA, OFAC, KYC, CDD, EDD, FATF, and dozens of firm-specific programme names. A 45-minute Series 7 prep module may contain FINRA, SIPC, UGMA, UTMA, Reg-BI, SIPC, NASD, NASAA, options chain terminology, and multiple types of annuity products (GMAB, GMWB, GLWB, SPDA). General-purpose ASR engines encounter this vocabulary at rates far above what their training corpora prepared them for. The cybersecurity training post covers the analogous problem in security content; financial compliance content has a higher acronym density and a larger proportion of terms that appear nowhere in publicly crawled text.

Regulatory citation accuracy. Financial services training routinely cites the specific rules being trained: FINRA Rule 3110, SEC Rule 17a-4(f), 23 NYCRR 500.14, ERISA Section 404(c), UPIA Section 2(b). These citations must appear correctly in the caption track because the caption track is evidence that the training addressed the cited regulation. A module on Reg-BI where the regulation name is transcribed as “regby” throughout does not demonstrate to an examiner that Regulation Best Interest was covered. This is the accuracy standard that WCAG 2.1 AA’s 99% threshold addresses at the word level; for regulatory citations, the effective requirement is 100% accuracy on the citation string itself.

Examination exposure. Financial services firms do not train hoping they will never need to produce the documentation; they train knowing they will. FINRA examination cycles for broker-dealers typically run every 1–4 years depending on firm size and risk profile. SEC examination cycles for investment advisers are similar. NY DFS examinations occur on a risk-based schedule. The training records, including captions, will be reviewed. A firm that has accumulated several years of inaccurate captions on its regulatory training content has accumulated several years of deficient training records.

The banking compliance captions page covers the platform-level picture for financial institutions. This post goes deeper on the regulatory framework, vocabulary accuracy by content type, and the documentation requirements that make caption accuracy non-negotiable in financial services.

FINRA Rules 3110, 3120, and 1240: training supervision obligations

FINRA’s supervision rules are the primary compliance framework for broker-dealers’ training programs. Understanding which rules apply and how they interact with caption accuracy requires working through each rule’s specific requirements.

FINRA Rule 3110: Supervision

FINRA Rule 3110 requires each member firm to establish and maintain a system to supervise the activities of each associated person that is reasonably designed to achieve compliance with applicable securities laws and regulations, and with applicable FINRA rules. The training materials provided to registered representatives and associated persons are subject to the firm’s supervisory system under Rule 3110. This means:

• Training materials must be reviewed and approved by a qualified principal before being presented to registered representatives
• The content of training materials must accurately represent the regulatory requirements being trained
• Records of the supervisory review of training materials must be maintained
• Evidence that associated persons completed required training must be documented and retained

For video-based training with caption tracks, the supervised training content includes the caption track. If a compliance video is reviewed and approved under Rule 3110 in its English-language audio form, but the caption track contains material inaccuracies — particularly inaccuracies affecting regulatory terminology — the caption version of the training has not been through the same supervisory review as the audio version. This creates a Rule 3110 gap for associated persons who consumed the training primarily through the caption track.

The practical implication: regulated firms should include caption review in their training material supervisory workflow. A compliance officer or qualified principal reviewing a training video for Rule 3110 purposes should also review the caption track for material accuracy on regulatory vocabulary. The QA methodology post covers the DCMP spot-check protocol that provides a documented accuracy review; in a financial services context, that review is the principal review for the caption version of the training material.

FINRA Rule 3120: Supervisory Control System

FINRA Rule 3120 requires that each member firm designate a registered principal to test and verify that the supervisory procedures established pursuant to Rule 3110 are reasonably designed to achieve compliance. The annual report required under Rule 3120 must include an assessment of the adequacy of the firm’s supervisory system. A training program that produces inaccurate captions on regulatory content — and has no quality control mechanism to detect those inaccuracies — is a gap in the firm’s supervisory control system that a Rule 3120 review should surface.

Rule 3120 annual reports are not typically drafted with caption accuracy in mind. But the supervisory control review that underlies the report should assess whether the training program for registered representatives is adequately supervised across all its delivery modes — including the captioned versions of training content that associated persons with hearing disabilities are receiving. A firm that has never reviewed caption accuracy as part of its Rule 3120 assessment has a gap in its supervisory control review.

FINRA Rule 1240: Continuing Education

FINRA Rule 1240 requires each member firm to establish a written training plan (the Firm Element) to provide ongoing training to registered persons. The Firm Element training must cover securities products, services, and strategies offered by the firm, as well as applicable regulatory requirements. The annual training curriculum under Rule 1240 is one of the primary contexts in which financial services video training is produced and distributed at scale.

Rule 1240 also covers the Regulatory Element, which requires registered persons to complete FINRA-mandated continuing education within specified deadlines. While the Regulatory Element content is produced by FINRA rather than by the firm, firms that supplement Regulatory Element content with their own training materials must ensure those materials meet the same documentation and accessibility standards as the rest of the firm’s training program.

The Firm Element training plan must be reviewed annually and updated to reflect regulatory changes, changes in firm products and services, and risk areas identified by the firm or by FINRA. Each year’s training plan update requires updating training content, which requires updating caption tracks on video-based modules. A firm with a 200-module annual training curriculum, of which 80% is video-based, produces a significant volume of caption updates per year simply from the annual Firm Element review cycle.

ADA Title I and WCAG 2.1 AA requirements for employer training

ADA Title I applies to employers with fifteen or more employees. All financial services firms above that threshold — which is virtually every regulated broker-dealer, investment adviser, bank, and insurance company — must provide accessible training to employees and applicants. WCAG 2.1 AA Success Criterion 1.2.2 requires synchronized captions for all pre-recorded audio and video content used in training. The 99%+ accuracy standard is the DCMP threshold most broadly applied in educational and corporate training contexts.

For a financial services firm, the ADA Title I obligation has three distinct application contexts:

Registered representative training

All training provided to registered representatives — Series 7 prep, annual Firm Element modules, product training, compliance updates — must meet ADA Title I caption requirements. This includes training that registered representatives are required to complete as a condition of maintaining their registration or satisfying FINRA Rule 1240 obligations. The overlap between the regulatory training obligation and the ADA accessibility obligation means that failing to provide accurate captions is simultaneously a FINRA compliance failure and an ADA failure.

The practical significance: registered representatives who use captions include not only those with hearing disabilities, but also those working in noisy trading floor environments, non-native speakers of English who benefit from simultaneous reading and listening, and those who review training content outside of standard office environments. The caption track that ADA Title I requires is the same track that serves this broader population. A firm that treats captions as a niche accommodation rather than a standard training delivery component is misallocating the caption investment and likely producing lower quality output as a result.

Employee communications and all-hands training

Compliance training for non-registered employees — operations staff, technology staff, human resources, finance — also falls under ADA Title I. Annual cybersecurity awareness training (required under NY DFS and other regulations), Code of Conduct training, HR policy training, and management training must all meet the same WCAG 2.1 AA caption accuracy standard. The vocabulary in this content includes firm-specific product names, internal system names, and department-specific terminology that is as prone to ASR failure as the regulatory vocabulary in registered-representative training, but less visible to compliance reviewers who focus primarily on Series licensing and regulatory content.

Externally licensed compliance content

When a firm licenses compliance training content from a third-party provider — ComplianceWire, Navex One, Skillsoft, Intuition, Kaplan Financial, or similar providers — the ADA Title I obligation to provide accessible training falls on the employer, not on the content vendor. The third-party compliance training post covers this gap in detail. In financial services, this gap is acute because the majority of AML/BSA training, reg-BI training, and FINRA-approved continuing education content is licensed rather than produced in-house, and third-party providers’ caption accuracy on financial compliance terminology is typically below the WCAG 2.1 AA threshold.

The compliance matrix post covers how ADA Title I, Section 508, and WCAG 2.1 AA interact across employer contexts. For financial services, the key point is that no exemption exists for the complexity of the content or for the regulatory origin of the training obligation. A FINRA-mandated training module must be both FINRA-compliant in content and ADA-compliant in delivery.

SEC Rules 17a-3 and 17a-4: captions as compliance documentation

SEC Rules 17a-3 and 17a-4 establish the books and records requirements for broker-dealers registered with the SEC. These rules specify what records must be created, how they must be maintained, and how long they must be retained. Training documentation is an explicit category in these rules, and the caption track on a training video is part of the training record.

Rule 17a-3: Records to be made

SEC Rule 17a-3(a) lists the records that every registered broker-dealer must make. The rule includes records of training programs provided to registered representatives and associated persons. Specifically, Rule 17a-3(a)(12) requires records of each registered person’s compliance with training requirements, including completion dates and content covered. A firm that provides training in video format with caption tracks must maintain records that accurately document the training content delivered — including, implicitly, the caption version of that content that was delivered to associated persons who consumed it via captions.

The practical implication of Rule 17a-3 for captioning: the training record is not just the fact that an associated person completed a module. The training record is the documented content of the module that was completed. If a FINRA examination requests the actual training materials reviewed by associated persons during the examination period, those materials include the caption files. A caption file that contains material inaccuracies in regulatory terminology is a training record that does not accurately document what was trained.

Rule 17a-4: Retention requirements

SEC Rule 17a-4 sets retention periods for the records required under Rule 17a-3. Training records generally fall under Rule 17a-4(b), which requires retention for at least three years, with the first two years in an accessible location. For training records associated with a registered person’s employment or registration status, longer retention periods may apply based on the associated person’s tenure at the firm.

Rule 17a-4(f) additionally requires that records maintained on electronic media must be stored in a non-rewriteable, non-erasable format (WORM compliance) and must be available for examination by FINRA, the SEC, and other regulators. For firms that store training video and caption files in an electronic recordkeeping system, this requirement applies to the caption files as well. A caption file that has been modified after the training was completed, without a version log, may not satisfy the WORM compliance requirement for the original training record.

This creates a specific requirement for caption version control in financial services. The caption file that was delivered to an associated person for training completed on a specific date must be the version that is retained in the compliant recordkeeping system. If the caption file is subsequently corrected for accuracy (a common operational event when captions are reviewed post-delivery), the correction must be documented as a new version, and the original version must be retained alongside it. The version control framework described in the caption glossary maintenance workflow — specifically the per-job glossary version logging and the version history retention — applies directly to this requirement, extended from glossary versioning to caption file versioning.

Investment Advisers Act Rule 204-2

For registered investment advisers (RIAs) rather than broker-dealers, the analogous books and records rule is Investment Advisers Act Rule 204-2. This rule requires investment advisers to maintain records of their compliance program, including their training programs under the investment adviser compliance program. RIAs that provide video-based training to their supervised persons — investment adviser representatives, portfolio managers, operations staff — must maintain those training records in the compliant format required by Rule 204-2. The same caption documentation principles apply: the caption track is part of the training record, and material inaccuracies in the caption track create a documentation gap.

FFIEC IT Examination Handbook and annual training requirements

The Federal Financial Institutions Examination Council (FFIEC) publishes IT Examination Handbooks that guide examiners from the OCC, FDIC, Federal Reserve, NCUA, and CFPB in assessing financial institutions’ information security programs. The Information Security Booklet and the Management Booklet both include requirements for annual information security training for all employees with access to sensitive customer data or critical systems.

Training requirements under the Information Security Booklet

The FFIEC Information Security Booklet (updated 2016, with subsequent interagency guidance) requires financial institutions to implement security awareness training for all personnel commensurate with their roles and responsibilities. The training must be provided at least annually and must cover topics appropriate to the institution’s risk profile. For employees in IT, operations, or customer-facing roles, the training must address specific threats and controls relevant to those roles.

For video-based security awareness training, the FFIEC examination process will assess whether the training is appropriate in content, was delivered to the required personnel, and is documented. Captions on video-based security awareness training serve the same dual purpose as in FINRA-supervised training: they satisfy the ADA Title I accessibility obligation for employer training, and they contribute to the documentation record that FFIEC examiners will review. A bank or credit union that has deployed security awareness training without accurate captions has both an accessibility gap and a training documentation gap — the documentation gap becoming visible if an examiner requests the training materials as delivered.

Third-party training content in financial institutions

Most banks and credit unions use third-party vendors for a significant portion of their annual compliance training — Bank Secrecy Act training, cybersecurity awareness, HMDA training, TRID training, and Fair Lending training are commonly licensed from providers such as Bankers Edge, Abrigo, Compliance Coach, or Wolters Kluwer. The FFIEC examination process assesses the adequacy of the training content and the documentation of completion. As with FINRA-supervised firms, the financial institution bears the obligation to ensure the training content meets accessibility requirements, regardless of whether the content was produced internally or licensed from a vendor.

For the cybersecurity training vocabulary problem specifically, the cybersecurity training captioning post covers the ASR failure rates on MITRE ATT&CK, CVE identifiers, and NIST framework vocabulary. Financial services cybersecurity training adds an additional vocabulary layer: financial-sector threat vocabulary (Business Email Compromise, wire fraud, account takeover, synthetic identity fraud, SWIFT network vocabulary) that intersects cybersecurity terminology with financial crimes vocabulary, creating a compound vocabulary challenge that general-purpose ASR handles poorly.

MiFID II and EU financial services training record requirements

The Markets in Financial Instruments Directive II (MiFID II) and its implementing regulation MiFID II Delegated Regulation (EU) 2017/565 require investment firms operating in the EU to ensure that their staff have the knowledge and competence required for the services they provide. Article 25 of MiFID II covers suitability and appropriateness assessments; the staff competence requirement that supports those assessments is implemented through the ESMA Guidelines on Knowledge and Competence (ESMA Guidelines 2015/1796, updated 2018).

MiFID II staff competence and training records

The ESMA Guidelines require investment firms to establish and maintain ongoing training programs for relevant staff, to assess staff competence against defined criteria, and to maintain records demonstrating that staff meet competence requirements. The training records that support competence assessments include the content of training modules completed, the dates of completion, and evidence of the training’s relevance to the staff member’s role and the investment services they provide.

For video-based training delivered to EU financial services staff under MiFID II, the caption track on the training video is part of the training record in the same way it is under FINRA and SEC rules in the US context. EU national competent authorities (NCAs) — the FCA in the UK pre-Brexit, BaFin in Germany, AMF in France, AFM in the Netherlands — may request training records during supervisory reviews. A training record with an inaccurate caption track does not accurately document the MiFID II training content delivered to the staff member.

EAA and EN 301 549 accessibility obligations

For EU-based financial services firms, the European Accessibility Act (EAA) and EN 301 549 create an additional accessibility obligation for digital training content. The EAA applies to private-sector organisations providing services in EU member states; EN 301 549 specifies the technical accessibility standard, which includes WCAG 2.1 AA as the benchmark for audio/video content. A financial services firm subject to MiFID II staff training requirements and operating in the EU is simultaneously subject to EAA accessibility requirements for its digital training delivery. The caption accuracy obligation under EAA/EN 301 549 is the same 99%+ accuracy standard that applies under ADA Title I in the US context.

The compliance matrix post covers the ADA/Section 508 side of the compliance framework; EAA and EN 301 549 apply the same logic to EU-context financial services training. A global financial services firm operating in both the US and the EU faces the same caption accuracy requirement from two regulatory directions simultaneously.

DORA and operational resilience training

The EU’s Digital Operational Resilience Act (DORA), which applied from January 2025, requires financial entities in the EU to implement ICT risk management frameworks that include staff training on digital operational resilience. DORA Article 13 requires that ICT training programmes be established and updated annually. For video-based ICT resilience training, the caption and documentation requirements follow the same pattern as MiFID II Firm Element training: the caption track is part of the training record, and the vocabulary includes technical terminology (API security, cloud dependency mapping, incident classification, RTO/RPO, BCP/DRP vocabulary) that compounds the ASR accuracy challenge already present in financial compliance content.

NY DFS 23 NYCRR 500: annual training documentation and examination evidence

New York Department of Financial Services Cybersecurity Regulation 23 NYCRR 500 applies to all covered entities — banks, insurance companies, and other financial services firms — that operate under a DFS license in New York. The regulation has been amended since its 2017 enactment; the current version (as amended in 2023) includes specific requirements for annual cybersecurity awareness training and for the documentation that covered entities must produce during a DFS examination.

Section 500.14: Annual cybersecurity awareness training

23 NYCRR 500.14 requires each covered entity to provide cybersecurity awareness training for all personnel on at least an annual basis. The training must be updated to reflect risks identified in the covered entity’s risk assessment and must include coverage of social engineering, phishing, and other cybersecurity threats relevant to the covered entity’s operations. The 2023 amendments require that training address current cybersecurity threats, which in the financial services context includes Business Email Compromise, business account takeover, ransomware, and wire fraud vocabulary.

For video-based cybersecurity awareness training subject to Section 500.14, the caption track carries both the ADA Title I accessibility obligation (for firms with fifteen or more employees) and the training documentation function. The training completed under Section 500.14 must be documented in a way that can be produced during a DFS examination. A caption track that misrepresents the cybersecurity vocabulary covered in the training — transcribing BEC as a different acronym, or misrendering phishing scenarios with incorrect technical vocabulary — creates a documentation gap in the Section 500.14 training record.

Section 500.17: Notice and reporting requirements and examination evidence

23 NYCRR 500.17 requires covered entities to submit annual certifications to DFS attesting that the covered entity has complied with the regulation’s requirements. The annual certification requirement means that DFS examination staff can assess compliance with Section 500.14 training requirements by reviewing the annual certification and requesting the supporting documentation — training completion records, training content, and evidence of the training’s adequacy relative to the covered entity’s risk assessment.

The DFS examination process for cybersecurity regulation compliance typically includes a document request for annual training materials and completion records. A covered entity whose annual cybersecurity training is delivered via video must be prepared to produce the training video and its caption track as part of that document request. The caption track as submitted to the DFS examination process should accurately represent the training content. Covered entities that have not reviewed caption accuracy on their Section 500.14 training materials are producing training records without knowing whether the records accurately document what the training contained.

New York State law overlay

Beyond the DFS Cybersecurity Regulation, New York financial services firms are also subject to the New York Human Rights Law (NYHRL), which prohibits disability discrimination in employment and applies to employers with four or more employees — a lower threshold than ADA Title I’s fifteen-employee minimum. For financial services firms in New York, the caption accessibility obligation for employer training arises under both ADA Title I (fifteen+ employees) and the NYHRL (four+ employees). The state-level compliance post covers the NYHRL and other state laws that create caption obligations beyond federal requirements.

Vocabulary accuracy: why financial compliance content fails ASR

General-purpose ASR models — including Whisper, Google Speech-to-Text, and Amazon Transcribe as used in most LMS platforms and third-party caption services without domain customisation — are trained on corpora that underrepresent financial regulatory vocabulary. The result is systematic substitution errors on the terminology that financial compliance training content is specifically designed to teach. These are not random errors; they are predictable failures on specific term types that appear frequently in financial services training.

The following sections document the substitution patterns by content type. These patterns are not hypothetical — they are the errors that surface in DCMP spot-check reviews of financial services training caption tracks produced by general-purpose ASR without a domain glossary. The error rates listed represent word error rates on the named-term categories specifically, not on the full transcript; overall WER on a 30-minute module may be 8–15% even if the filler-vocabulary WER is below 2%.

Series 7, 63, and 66 exam prep content

Series licensing prep content is among the highest-vocabulary-density training in financial services. A 45-minute Series 7 prep module may reference 60–80 distinct regulatory terms, acronyms, and proper nouns. Common substitution failures on general-purpose ASR:

• FINRA → “Finnra,” “Finra” (inconsistent), “fin RA” (spaced)
• Reg-BI (Regulation Best Interest) → “regby,” “reg B I,” “Reg. B. I.” (with periods that disrupt meaning)
• UGMA (Uniform Gifts to Minors Act) → “Ugma,” “U-G-M-A” (spelled out rather than pronounced as a word)
• UTMA (Uniform Transfers to Minors Act) → “Ultima,” “utma,” “UTMA” (sometimes correct, often not)
• SIPC (Securities Investor Protection Corporation) → “sipik,” “sick,” “S-I-P-C” (spelled out)
• NASAA (North American Securities Administrators Association) → “NASA” (the space agency), “NASSA,” “na-SAA”
• GMAB / GMWB / GLWB (variable annuity guarantee types) → “GM AB,” “GM WB,” “GL WB” (always spaced, sometimes further garbled)
• RIA (Registered Investment Adviser) → “ria” (lowercase, which in training context is indistinguishable from the given name)
• B-D (broker-dealer) → “be D,” “BD,” “B D” (inconsistent rendering)
• SPDA (Single Premium Deferred Annuity) → “speed A,” “S-P-D-A,” “spda”

In Series 7 prep content, these terms are not peripheral — they are the examination vocabulary. A module on retirement account types that refers to UGMA and UTMA accounts throughout, and has those terms misrendered in the caption track, is not providing examinees with a reliable study resource. For the ADA Title I accessibility obligation, this is the equivalent of a Braille study guide that misrenders the examination vocabulary.

AML/BSA and financial crime vocabulary

Anti-money laundering (AML) and Bank Secrecy Act (BSA) training is the highest-volume mandatory training category for banks, broker-dealers, and money service businesses. AML/BSA training must be updated annually to reflect changes in regulatory guidance and emerging financial crime typologies. The vocabulary in AML/BSA training combines standard regulatory citations with financial crime terminology and law enforcement vocabulary:

• FinCEN (Financial Crimes Enforcement Network) → “Fin-sen,” “Fin sen,” “fincen” (inconsistent capitalisation)
• SAR (Suspicious Activity Report) → “spar,” “S-A-R” (spelled out), “sar” (lowercase, indistinguishable from “saw” in context)
• CTR (Currency Transaction Report) → “C-T-R,” “ceeter,” “sitter”
• 314b (USA PATRIOT Act Section 314(b)) → “three fourteen B,” “314 B,” “three-one-four B” (none of which render as “314b”)
• OFAC (Office of Foreign Assets Control) → “O-Fak,” “Ofac,” “oh-fack”
• CDD (Customer Due Diligence) → “C-D-D,” “CDD” (sometimes correct), “see-dee-dee”
• EDD (Enhanced Due Diligence) → “E-D-D,” “Eddie,” “EDD” (inconsistent)
• FATF (Financial Action Task Force) → “fat F,” “FATF,” “the FATF” (inconsistent)
• Smurfing → Usually misidentified as “surfing,” “stuffing,” or even “Smurf-thing” in context-free transcription
• Structuring → Usually correct as a standalone word, but misidentified in compound phrases like “structuring transaction” where “structuring” as a financial crime term may be rendered as “restructuring” by a model that treats financial context as reconstruction vocabulary

AML/BSA training with these substitution errors creates a documentation problem that is particularly acute: the caption track on a SAR-filing procedures module that renders “SAR” as “spar” throughout does not document that the SAR filing requirement was trained. The 99% accuracy post covers why these individual errors compound across a training module; in AML/BSA context, a 12–15% proper-noun error rate on a 30-minute module may produce 50–80 regulatory term substitutions per module.

ESG, fiduciary, and ERISA vocabulary

Investment management training for portfolio managers, investment advisers, and plan sponsors combines ESG vocabulary (Environmental Social Governance, AUM, NAV, PE ratio, fixed income vocabulary) with fiduciary and ERISA-specific terminology:

• ERISA (Employee Retirement Income Security Act) → “Erisa” (sometimes correct), “E-R-I-S-A” (spelled out)
• ERISA 404(c) → “Erisa four O four C,” “Erisa 404 C,” “Erisa four-oh-four-c”
• UPIA (Uniform Prudent Investor Act) → “Upya,” “U-P-I-A,” “UPIA” (inconsistent)
• QPAM (Qualified Professional Asset Manager) → “Q-Pam,” “Ku-pam,” “QPAM” (rarely correct)
• AUM (Assets Under Management) → “aum,” “OM,” “OHM” (the physics unit)
• EBITDA → “Ebita,” “Ebidah,” “E-B-I-T-D-A”
• NRR (Net Revenue Retention) → “narrow,” “N-R-R,” “NRR” (inconsistent)
• Pari passu → “pari pass-you,” “parry passu,” “parity passu”
• Basel III / Basel IV → “Basil the third,” “Basil III” (the herb, not the regulatory framework)

For plan sponsor training on ERISA fiduciary obligations — a high-stakes content type because ERISA 404(c) compliance requires documented procedural prudence — inaccurate captions on training content are a fiduciary documentation gap. The training record that demonstrates the plan sponsor understood and implemented prudent expert standard procedures should accurately reproduce those procedures, including the ERISA section numbers that define them.

Earnings call and financial reporting vocabulary

For listed companies with investor relations training programs, financial reporting vocabulary presents an additional ASR accuracy challenge. The executive communications post covers earnings call captioning in detail; the vocabulary failures documented there — non-GAAP rendered as “non-gap,” EBITDA as “Ebita,” forward-looking statement safe harbor as “forward-looking statement safe harbour” (the British spelling) or further garbled — apply equally to training programs for finance staff and investor relations teams that cover earnings release procedures, quiet period protocols, and Regulation FD compliance.

Domain glossary requirements for financial services content

The vocabulary failure rates documented above are consistent across general-purpose ASR providers. The solution is a domain glossary that provides the caption model with phonetic variants, confidence scores, and context restrictions for financial regulatory vocabulary. The glossary architecture post covers the structural design of a caption glossary; for financial services content, the glossary must cover at minimum:

• FINRA, SEC, and relevant regulatory agency acronyms with phonetic variants
• Series exam vocabulary (UGMA, UTMA, SIPC, NASAA, all annuity product type acronyms)
• AML/BSA terminology (FinCEN, SAR, CTR, OFAC, CDD, EDD, FATF, 314b)
• ERISA and fiduciary vocabulary (ERISA section numbers, UPIA, QPAM, 404(c))
• Firm-specific product and system names
• Regulatory rule citations in the format used by instructors (“Rule 3110,” “17a-4,” “500.14”)

With a domain glossary of 150–300 terms covering these categories, financial compliance training content reliably achieves 97–99%+ overall accuracy, and proper-noun error rates drop from 15–25% to below 1%. The glossary is the mechanism that makes the difference between training documentation that can be defended on examination and training documentation that cannot.

LMS landscape in financial services

The LMS landscape in financial services is more fragmented than in other corporate L&D verticals, and the platforms that dominate the sector have weaker built-in caption support than general-purpose enterprise LMS platforms. Understanding which platform a firm uses determines which caption delivery model is available and what the examination documentation workflow looks like.

Cornerstone OnDemand (large banks and insurance companies)

Cornerstone OnDemand is the dominant LMS at large banks, regional banks, and insurance companies. Firms such as JPMorgan Chase, Wells Fargo, and many regional bank holding companies run Cornerstone for their enterprise learning management. Cornerstone provides built-in auto-captioning via an internal Whisper-family model, but without custom vocabulary support, Cornerstone’s auto-captions on financial compliance content produce the substitution errors documented in the vocabulary section above. The Cornerstone captioning page covers the platform-level workflow in detail.

The sidecar SRT delivery workflow for Cornerstone involves uploading an SRT file alongside the video during content ingestion. Cornerstone supports SRT sidecar files for all video content types. For examination documentation purposes, the SRT file uploaded to Cornerstone should be produced from the reviewed and approved caption track — not from Cornerstone’s auto-caption feature — and should be versioned in the firm’s caption records. The caption file stored in Cornerstone is the training record; it should match the version log in the captioning programme’s documentation.

iSpring (broker-dealers and insurance distributors)

iSpring Learn and iSpring Suite are widely used at broker-dealers, independent financial advisory firms, and insurance distributors. iSpring’s positioning as a rapid e-learning development and LMS platform makes it attractive for firms that produce their own compliance training content in-house using Storyline or PowerPoint-based workflows. The iSpring captioning page covers the platform-specific workflow; the key characteristic for financial services is that iSpring has no built-in auto-captioning feature for uploaded video, which means the sidecar SRT workflow is the only caption delivery path and the caption quality depends entirely on the external captioning service used.

This is structurally advantageous for documentation purposes: iSpring deployments always use a separately produced caption file, so the caption file is always available as a discrete record. The disadvantage is that there is no auto-caption fallback, which means firms that have not established a caption production workflow for iSpring-hosted content simply have uncaptioned training — an ADA Title I gap that is particularly common among independent broker-dealers that have not prioritised accessibility.

Custom LMS implementations at bulge-bracket firms

Bulge-bracket investment banks (Goldman Sachs, Morgan Stanley, Citigroup, Bank of America Merrill Lynch) and large asset managers typically operate custom LMS platforms built on internal video infrastructure, often using Kaltura, Panopto, or Brightcove as the video delivery layer with a custom learning management wrapper. These custom implementations vary widely in their caption support. Some have built robust sidecar caption workflows; others have auto-captions without accuracy controls. The examination documentation requirements are the same regardless of platform, but the workflow to satisfy them varies significantly by firm.

For custom LMS implementations, the caption programme design must address: which team owns caption production (often L&D but sometimes the video production team in large firms), how captions are uploaded to the video platform, whether the caption file is stored separately from the video in the firm’s recordkeeping system, and how caption files are retained under the SEC 17a-4 or Investment Advisers Act Rule 204-2 retention requirements. These are operational design questions that must be resolved before caption production begins, not after.

ComplianceWire and specialist compliance training platforms

ComplianceWire is used by pharmaceutical firms with financial services crossover training, but more relevant for broker-dealers are specialist compliance training platforms: Intuition (used by major banks for structured finance and derivatives training), Kaplan Financial (Series exam prep), and Navigator (AML/BSA training). These platforms supply both the content and the delivery mechanism, which creates the same third-party compliance training gap documented in the third-party compliance post: the firm licenses training content that already has captions, but the platform-generated captions have not been produced with a financial vocabulary glossary.

For broker-dealers using these platforms, the captioning approach depends on whether the platform supports caption replacement (uploading a firm-produced SRT file to replace the platform-generated captions) or caption overlay (displaying firm-produced captions on top of platform-generated captions, or toggling between them). Most specialist compliance training platforms were not designed with WCAG 2.1 AA compliance in mind and support neither option cleanly. The procurement solution — requiring SRT file delivery in the contract, with accuracy specifications tied to the firm’s standard — is covered in the vendor SLA checklist post.

Caption audit trail as regulatory examination evidence

When a FINRA examination, SEC examination, FFIEC examination, or NY DFS examination includes a request for training records, the response must demonstrate that the training was provided, that it covered the required topics, and that it met the required standards. The caption audit trail is the mechanism that supports this demonstration for video-based training programs.

What the examination document request looks like

A FINRA document request for training records typically asks for: the firm’s written continuing education program (the Firm Element document), the annual training curriculum including the titles and descriptions of modules completed, the completion records for registered persons, and the training content itself (the materials reviewed). For video-based training, the “training content itself” request means the video files and their associated caption files.

An SEC examination of an investment adviser’s compliance program typically includes a request for the compliance training schedule, the training materials, and evidence that the training was completed by appropriate staff. For video-based training, this request has the same scope: video and caption files are part of the training material record.

A NY DFS examination under 23 NYCRR 500 may request the covered entity’s Section 500.14 annual training materials and completion records. The examination is assessing whether the training was adequate to address the risks identified in the covered entity’s risk assessment. A training video whose caption track does not accurately represent the cybersecurity threats discussed in the training has a content adequacy gap that the examination may flag.

Elements of a caption audit trail

A caption audit trail for financial services regulatory training should include the following elements, maintained in the firm’s records system in a retention-compliant format:

Not every firm needs all six elements in their caption audit trail from day one. The minimum viable audit trail for a broker-dealer under FINRA examination is: caption file version, principal review record, and delivery log. The minimum viable trail for a bank under FFIEC examination is: caption file version, QA accuracy score, and delivery log. The full six-element trail is the target state for a programme operating at scale with multi-year examination exposure and SEC 17a-4 WORM requirements.

Connecting the caption record to the LMS completion record

The LMS completion record (the data that shows registered person X completed module Y on date Z) is the primary examination evidence for training completion. The caption record needs to be linked to the LMS completion record to demonstrate that the captioned version of the training — which is the version accessible to employees with hearing disabilities — is the same training that satisfies the regulatory completion requirement.

This linkage is typically established by: using a consistent module ID across the LMS completion record and the caption audit trail; storing the caption file in the same content management location as the video file with a naming convention that links them; and including the LMS module ID in the caption file metadata or in the caption programme’s records system. Firms that have not established this linkage have a documentation chain that can demonstrate training was completed and that a caption file exists, but cannot demonstrate that the training was completed using the captioned version by the employees who need it. An ADA accommodation request or OCR investigation that probes this linkage will surface the gap.

The compliance reporting post covers the leadership dashboard metrics that surface caption programme status; for financial services, those metrics should include a FINRA/SEC documentation readiness indicator that tracks whether the caption audit trail is complete for all training modules in scope for the current examination cycle. The annual review post covers the annual audit process that maintains the documentation record over time.

Production workflow for financial services training at scale

A financial services training captioning programme operating across FINRA, SEC, FFIEC, MiFID II, and NY DFS obligations needs a production workflow that integrates caption accuracy, documentation, and delivery into a repeatable process. The following workflow addresses the specific requirements of financial services training at a firm with 50–1,000 registered representatives and a 100–500 module annual training curriculum.

Step 1: Content classification and vocabulary scoping

Before producing captions on a new training module, classify the content by regulatory type: Series licensing prep, AML/BSA, ERISA/fiduciary, cybersecurity awareness, product/firm training, or general compliance. The classification determines which vocabulary domain applies, which glossary version is used, and which principal review is required.

For each content type, identify the top 20–30 regulatory terms and acronyms that are specific to that module and not already in the base financial services glossary. These module-specific terms should be submitted to the glossary before caption production begins, not after. The timing problem documented in the glossary maintenance workflow post — specifically the announce-day problem for new product names — applies equally to regulatory vocabulary changes. When FinCEN issues new guidance that introduces a new term, that term should be in the financial services glossary before the AML/BSA training module that covers the new guidance is captioned.

Step 2: Caption production with domain glossary

Submit the training video for captioning with the current financial services domain glossary applied and the module-specific supplemental terms included. For financial services content, this step should produce a DCMP-protocol accuracy score above 97% on the first pass. If it does not, the gap is typically in module-specific vocabulary that was not captured in the pre-production vocabulary scoping step.

The first-pass accuracy target for financial compliance content is higher than for general corporate L&D content. The error rate calculator post covers the DCMP formula and the 99% threshold; for financial services content where individual term errors create regulatory documentation gaps, the practical target should be 99%+ on full-transcript WER and 100% on regulatory citation strings (specific rule numbers, specific regulatory agency acronyms used as citations).

Step 3: Principal review of caption track

Under FINRA Rule 3110, a qualified principal must review training materials before they are distributed to registered representatives. For video-based training with caption tracks, the principal review should include the caption track. The principal review of the caption track does not require reviewing the entire transcript word-by-word; it requires reviewing the regulatory vocabulary sections specifically — the sections where the training is making substantive compliance points that registered representatives are expected to understand and apply.

The DCMP spot-check protocol described in the QA methodology post can be adapted to serve as the principal review methodology: a 10% random sample plus all sections covering regulatory citations and rule-specific vocabulary. The principal review record should document: who conducted the review, when, which sections were spot-checked, and the finding (acceptable accuracy / corrections required). This record is the supervisory documentation that satisfies FINRA Rule 3110 for the captioned version of the training material.

Step 4: Version logging and delivery to LMS

After principal review and any corrections, log the final caption file version: version number, glossary version used, review date, reviewer name, and QA accuracy score. Upload the caption file to the LMS using the sidecar SRT workflow and link the caption record to the LMS module ID. Store the caption file in the firm’s WORM-compliant records retention system with the glossary version log and principal review record attached.

For firms operating under SEC 17a-4 WORM requirements, the caption file storage system must be the same system used for the rest of the training record, or must be linked by module ID to that system so that examination document production can retrieve the caption file alongside the completion records. Storing caption files in a separate, non-WORM-compliant location and relying on version control in a code repository (a common IT workflow) does not satisfy SEC 17a-4(f) requirements for training records.

Step 5: Annual review and examination preparedness

The annual training programme review under FINRA Rule 1240 should include an assessment of the caption programme’s documentation completeness. The review should confirm: all training modules in scope for the current Firm Element curriculum have current caption files in the records system; all caption files have associated principal review records; the glossary version log covers all modules produced in the review period; and the caption audit trail is complete and linkable to LMS completion records for all registered persons who completed training in the review period. The budget planning post covers how to build the annual programme budget; for financial services, the budget should include a line item for examination documentation review that accounts for the time required to confirm audit trail completeness before each examination cycle.

Eight failure modes in financial services training captioning

These are the most common ways financial services caption programmes fail to satisfy their regulatory and accessibility obligations.

1. Relying on LMS-native auto-captions for examination documentation

   LMS-native auto-captions — Cornerstone’s internal Whisper, Workday Learning’s Google STT, iSpring’s absence of auto-captions — are not produced with a financial domain glossary and do not meet the 99% WCAG 2.1 AA accuracy threshold on compliance content. More critically, LMS-native auto-captions are typically generated on demand when a learner requests them, not stored as a versioned file in the firm’s records system. This means there is no version-controlled caption record that can be produced during a FINRA or SEC examination to demonstrate what captions were delivered. A firm that relies on LMS-native auto-captions has no caption documentation, not just inadequate caption documentation. The LMS auto-caption accuracy comparison post covers the platform-specific accuracy benchmarks; for financial services, the conclusion is the same across all seven major platforms: LMS-native captions are not an acceptable solution for regulatory training documentation.

2. Caption files not retained in a WORM-compliant system

   SEC Rule 17a-4(f) requires that broker-dealer records maintained on electronic media be stored in a non-rewriteable, non-erasable format. A caption file stored in a shared drive, a cloud storage folder, a version control repository, or an LMS content library is not WORM-compliant unless the storage system has been specifically configured to satisfy Rule 17a-4(f). Firms that store training records in a compliant system but caption files outside that system have a records retention gap. The resolution requires either migrating caption file storage to the compliant system or establishing a documented policy that designates a specific WORM-compliant location for caption file retention and integrates caption file production into the records management workflow.

3. Third-party compliance training accepted without caption review

   Broker-dealers that license AML/BSA training, reg-BI training, or Series prep content from third-party providers frequently accept whatever captions the vendor provides without reviewing them for accuracy on financial vocabulary. The assumption is that a specialist compliance training vendor has produced accurate captions. This assumption is wrong in most cases: specialist compliance content vendors produce content with general-purpose ASR captions or with human-edited captions that have not been reviewed with a financial vocabulary standard. The firm’s ADA Title I caption obligation applies to licensed training content. The firm’s FINRA Rule 3110 supervisory documentation obligation applies to the training content it uses with registered representatives, including licensed content. Accepting third-party captions without review means accepting liability for accuracy that the vendor has not guaranteed.

4. No glossary maintenance for regulatory vocabulary updates

   Financial regulatory vocabulary changes annually. FinCEN issues new guidance. FINRA amends rules. SEC adopts new regulations. Each regulatory change introduces new vocabulary into compliance training content before the glossary has been updated to include it. A firm whose AML/BSA glossary was last updated in 2023 and has not been updated since will produce caption errors on every training module that covers regulatory updates issued since 2023. The glossary maintenance programme described in the glossary maintenance workflow post must include a specific step for regulatory vocabulary updates: when FINRA, SEC, FinCEN, or DFS issues new guidance or amendments that will be covered in training content, the new regulatory vocabulary should be submitted to the glossary before the first training module on that topic is captioned.

5. Principal review of video training excludes caption track

   Under FINRA Rule 3110, a qualified principal reviews training materials before distribution. This review typically covers the video content: the script, the presenter’s statements, the graphic content, and the substantive compliance points made. Caption tracks are rarely included in the supervisory review because compliance officers reviewing training content are not thinking about accessibility. This creates a Rule 3110 gap: the caption version of the training, which may contain material inaccuracies in regulatory terminology, has been distributed to registered representatives without supervisory review. Firms should add caption track review to their Rule 3110 training material review checklist, either as a full transcript review or as a DCMP spot-check on regulatory vocabulary sections.

6. Caption documentation not linked to LMS completion records

   A firm can have both a compliant caption file and compliant LMS completion records and still have an examination documentation gap if the two records are not linked. When an examiner requests training records for a specific registered person, the examiner wants to see: (a) that the person completed the training, (b) what training was completed, and (c) that the training was accessible. These three elements come from three places: the LMS completion record, the training content record, and the caption record. If the caption record is stored in a separate system with no cross-reference to the LMS completion record, the documentation chain is broken. Building the cross-reference — typically by including the LMS module ID in the caption file name or in the caption documentation record — costs one field in the captioning workflow and eliminates the examination documentation gap.

7. MiFID II training records maintained separately from caption records in EU entities

   EU-based or dual-regulated financial services firms (US-registered with EU entities, common in global investment banks) maintain separate records systems for US regulatory compliance and EU regulatory compliance. Caption records for MiFID II training purposes may be maintained in the EU entity’s system, while the caption production workflow and glossary management are controlled by the US L&D team. When the MiFID II training records are reviewed by an NCA, the caption records need to be producible from the EU entity’s system. If the caption files are stored only in the US records system, the EU entity has a MiFID II documentation gap even if the US entity has complete documentation. Dual-regulated firms should establish a clear policy on where caption records are stored for each regulatory purpose, and ensure that EU entity training records include caption files for all video-based training in scope for MiFID II competence demonstrations.

8. Caption programme budget not including examination documentation overhead

   Financial services firms that build a caption programme often budget for caption production cost (per-minute vendor fees or subscription cost) and labour cost (the time to upload and manage captions) but not for examination documentation overhead: the time to build and maintain the caption audit trail, the annual review of documentation completeness, the principal review of caption tracks, and the records management integration work. For a broker-dealer with 100 training modules per year and a 3-year examination cycle, the examination documentation overhead is a one-time audit of the full 300-module caption record at each examination. That audit is not trivial. Firms that have not budgeted for it discover the gap when the examination document request arrives. The hidden FTE cost post documents the general labour cost of caption programme operations; in financial services, the examination documentation labour is an additional cost layer that non-financial-services L&D teams do not face.

FAQ

Does FINRA require caption tracks on training materials, or is that just an ADA requirement?

FINRA does not have a rule that specifically requires caption tracks on training materials. ADA Title I requires accessible training, which for video-based training means WCAG 2.1 AA synchronised captions. However, the interaction between ADA Title I and FINRA Rule 3110 creates a de facto captioning requirement: Rule 3110 requires that training materials be reviewed and that they be accessible to all registered representatives, including those with hearing disabilities. A firm whose training program is not accessible to registered representatives with hearing disabilities has a Rule 3110 supervision gap — the training program is not “reasonably designed” to achieve compliance if it excludes a category of registered representatives from receiving the training. The ADA Title I caption obligation and the FINRA Rule 3110 training accessibility obligation point to the same operational requirement: accurate captions on all video-based training content provided to registered persons.

How does SEC Rule 17a-4(f) WORM compliance apply to caption files specifically?

Rule 17a-4(f) requires that records maintained on electronic storage media be stored in a non-rewriteable, non-erasable format for the retention period applicable to the record type. Caption files that are part of the training record are subject to this requirement. The typical retention period for training records under Rule 17a-4 is three years. A caption file stored in a writable cloud storage folder (Google Drive, SharePoint without WORM configuration, or similar) does not satisfy 17a-4(f) unless the storage system has been configured to enforce non-overwriting for the retention period. The same SEC-registered compliance records storage systems that firms use for trade records, customer records, and supervisory procedure documentation should be used for training records including caption files. Many broker-dealers already have a third-party records retention vendor (Iron Mountain Digital, Smarsh, or similar) that provides WORM-compliant storage; adding caption files to the documents archived in that system is the standard approach.

What is the minimum caption accuracy requirement for FINRA-supervised training content specifically?

FINRA does not publish a specific numerical accuracy standard for training content captions. The applicable accuracy standard comes from ADA Title I and WCAG 2.1 AA, which most practitioners implement at the DCMP 99%+ threshold. For FINRA examination purposes, the question is not whether the firm used the 99% standard; the question is whether the training documentation accurately represents what was trained. A caption track with a 95% overall accuracy rate on a Series 7 prep module may still have material errors on the examination vocabulary that was the module’s content. The 99% WCAG 2.1 AA threshold is not the only relevant standard — for regulatory citation accuracy within financial compliance content, the practical standard is 100% on the specific rule numbers and regulatory agency names that constitute the examination-required knowledge. A module that misrenders FINRA Rule 1240 as “FINRA rule 1 240” or “Finnra Rule one-two-forty” has failed on the most important element of its content, even if the overall WER is below 1%.

Are there special considerations for captioning AML/BSA training produced by a third-party vendor like ComplianceWire or Navex One?

Yes, and they compound the general third-party training captioning problem in two ways. First, AML/BSA training vocabulary (SAR, CTR, FinCEN, 314b, OFAC, FATF, structuring, smurfing) has among the highest proper-noun ASR failure rates of any compliance content category, and third-party providers rarely produce captions with a financial crime vocabulary glossary. Second, the firm’s AML programme compliance documentation — which is subject to review by FINRA, OCC, FDIC, Fed, or FinCEN depending on the firm type — includes evidence of annual AML training. That documentation is the training record, including captions. The approach most firms take: include a caption SRT file delivery requirement in the third-party training vendor agreement (with a specified accuracy standard for financial vocabulary), conduct a DCMP spot-check on each module received, and replace the vendor-provided SRT file with a corrected version if accuracy does not meet the standard. The cost of that correction is an operational cost that should be included in the AML training budget. The alternative — using the vendor-provided captions uncorrected — accepts AML documentation risk that is disproportionate to the correction cost.

How should a financial services firm handle the caption documentation requirements for EU and US registered persons simultaneously?

For a dual-regulated firm with both FINRA-supervised US operations and MiFID II-regulated EU operations, the caption documentation requirements from each jurisdiction should be mapped separately. The US requirements (FINRA Rule 3110 supervisory review, SEC 17a-4 WORM retention) and the EU requirements (MiFID II competence demonstration, EAA/EN 301 549 accessibility, DORA training documentation) have different documentation formats, retention periods, and examination presentation requirements. The practical approach is to produce captions once — with a domain glossary that covers both US and EU financial vocabulary, and with a glossary that includes terms from both jurisdictions where they differ — and then route the caption records to the appropriate documentation system for each regulatory purpose. The principal review record should note which regulatory frameworks the review is satisfying. Dual-regulated firms that have not mapped their caption documentation obligations across jurisdictions are at risk of satisfying one jurisdiction’s requirements while leaving the other’s incomplete.

Does the NY DFS Cybersecurity Rule require captions on cybersecurity awareness training videos specifically?

23 NYCRR 500.14 requires annual cybersecurity awareness training but does not specifically require video format or captioned video format. The accessibility obligation comes from ADA Title I (for covered entities with 15+ employees) and the New York Human Rights Law (for covered entities with 4+ employees). A covered entity that delivers its annual cybersecurity awareness training via video must provide WCAG 2.1 AA captions to meet those accessibility obligations. For DFS examination purposes, the relevant question is whether the training was provided to all required personnel — not just personnel without disabilities. An annual cybersecurity training programme that is not accessible to employees with hearing disabilities has not been provided to all required personnel, which is a Section 500.14 compliance gap. DFS examinations of cybersecurity programmes are increasingly scrutinising training delivery adequacy, not just training completion rates; accessibility is part of delivery adequacy.

What is the most cost-effective caption programme structure for a small broker-dealer with ten to twenty registered representatives?

For a small broker-dealer with 10–20 registered representatives, the annual training curriculum under FINRA Rule 1240 typically involves 20–40 video modules per year, most of which are licensed from a third-party compliance training provider. The most cost-effective structure is: (1) include a caption SRT file delivery requirement in the third-party training vendor agreement, with a specified accuracy standard for financial vocabulary; (2) conduct an annual DCMP spot-check on a sample of received modules to verify vendor caption accuracy; (3) store caption files in the firm’s compliant records retention system (the same system used for other Rule 17a-4 records) with module IDs that link to LMS completion records; and (4) include a caption review step in the principal’s annual training material review under Rule 3110. For small broker-dealers producing in-house video content (which is less common), a per-minute professional captioning service with a financial domain glossary is the appropriate production model. The budget planning post covers how to model the cost at different production volumes; for a 10–20 representative firm, the total annual caption programme cost including documentation overhead is typically $2,000–$5,000 per year, primarily in vendor SRT review and principal review time rather than caption production cost. The live demo shows the captioning workflow and glossary application in practice.