AI Segment Reporting Automation ASC 280: 2026 Practitioner Walkthrough

ASU 2023-07 didn't just add a few new footnote lines. It fundamentally changed the data burden of ASC 280 segment reporting, and for most finance teams, the manual processes that worked before won't survive the quarterly close timeline that the new interim requirements demand. AI automation is the practical answer, but almost no published guidance explains the mechanics. This walkthrough does.

Key takeaway: ASU 2023-07 is effective for calendar-year 2024 annual reports (filed Q1 2025) and interim periods beginning Q1 2025. If your team is still building segment expense schedules by hand, you're already behind.

What ASU 2023-07 Actually Changed in ASC 280

ASU 2023-07, issued by FASB in November 2023, is the most significant amendment to ASC 280 in nearly three decades. The standard already required segment revenue, profit or loss, and assets. The update adds four new disclosure obligations that create real data infrastructure problems:

1. Significant segment expenses regularly provided to the CODM and included in reported segment profit or loss
2. An "other segment items" amount (the residual between segment revenue minus significant expenses and segment profit/loss, with a description of its composition)
3. The title and position of the CODM, plus an explanation of how the CODM uses the reported profit or loss measure
4. All of the above on an interim basis, not just annually

The effective dates matter: annual periods beginning after December 15, 2023 (calendar-year 2024 Form 10-K, filed Q1 2025), and interim periods beginning after December 15, 2024 (Q1 2025 Form 10-Q for calendar-year filers). All prior-period segment information must be retrospectively restated.

One surprise that catches companies off guard: Deloitte's 2024 Roadmap to Segment Reporting confirms that single-segment entities are fully subject to ASU 2023-07. If your company reports as one segment, you still must disclose significant segment expenses, the CODM's title, and how the CODM uses the profit measure. KPMG estimates this affects 30 to 40% of S&P 500 companies that report as a single segment.

For a deeper comparison of how ASU 2023-07 interacts with the separate DISE requirements, see DISE vs Segment Reporting: What Changes for US Registrants by Filer Type.

The Five Data Problems ASU 2023-07 Creates

Before mapping AI to solutions, it's worth naming the specific pain points precisely. These are the operational gaps that drive comment letter risk.

1. The "significant segment expenses" determination. This is a facts-and-circumstances judgment, not a bright-line rule. As PwC's segment reporting guide puts it: "Companies will need to carefully evaluate which expenses are 'regularly provided to the CODM', this is a facts-and-circumstances determination that will require a systematic inventory of internal management reporting packages. The documentation burden is significant." Most companies don't have a clean, automated inventory of what their CODM actually reviews.

2. CODM documentation is informal and scattered. Board decks, email summaries, Tableau dashboards, and ad hoc management reports all potentially constitute CODM-reviewed materials. Systematically capturing and archiving these is a manual nightmare without tooling.

3. ERP misalignment. SAP, Oracle, and Workday cost centers rarely map cleanly to reportable segments. The gap between ERP output and disclosure-ready segment data requires manual transformation that is both time-consuming and error-prone.

4. Retrospective restatement is a data archaeology problem. EY's 2024 Financial Reporting Developments identifies this as "the sleeper issue": companies must reconstruct which expenses were regularly provided to the CODM in prior periods, and for many, that historical documentation doesn't exist in structured form.

5. Interim compression. Producing segment expense disclosures quarterly, not just annually, compresses an already tight close calendar. The Q1 2025 10-Q was the first filing where calendar-year filers had to meet this bar.

Where AI Automation Actually Applies: A Step-by-Step Map

Not every step in the ASC 280 workflow is automatable. Here's an honest map of where AI adds genuine value and where human judgment is non-negotiable.

Step 1: CODM Report Ingestion and Expense Identification

The first question under ASU 2023-07 is: which expenses are "regularly provided to the CODM"? Answering it systematically requires parsing the actual management reports the CODM receives.

AI tools apply natural language processing (NLP) and document classification to board packages, management decks, and internal dashboards to extract expense line items and map them to segment structures. Named entity recognition can identify cost categories ("R&D," "cost of revenue," "selling expenses") and flag which appear consistently across reporting periods, supporting the "regularly provided" determination.

This is emerging but not yet standardized. Gartner projects that by 2027, over 50% of large public companies will use AI-assisted tools for at least one step in their financial close and disclosure process, with segment reporting and footnote drafting among the top three use cases.

Human judgment required: The final determination of which expenses are "significant" and meet the CODM-reviewed threshold is an accounting judgment that must be signed off by the controller or CFO. AI surfaces candidates; humans decide.

Step 2: ERP Data Extraction and Segment Mapping

AI-assisted data pipelines can extract GL account balances from ERP systems and apply ML-based classification to map cost center data to reportable segments. The three-tier cost allocation framework maps directly to where AI adds the most value:

Tier 3 is where misclassification risk is highest and where AI governance matters most. Shared cost allocations that shift between periods without documented rationale are a specific SEC focus area.

Step 3: Building and Maintaining the Segment Data Dictionary

This is the governance infrastructure that makes AI outputs auditable. A segment data dictionary is a documented mapping of ERP cost centers, GL accounts, and expense categories to reportable segments. AI tools use it as a reference; auditors inspect it.

Best practice, drawn from Big-4 advisory work, is to maintain this dictionary as a version-controlled artifact with change logs. Every time a cost center is reclassified or a new segment is added (post-M&A, for example), the dictionary is updated and the change is documented with an effective date and approver.

Without this, AI-generated segment classifications are unauditable, regardless of how accurate they are.

Step 4: Retrospective Restatement Using AI Document Parsing

For companies that hadn't previously documented CODM-reviewed expenses systematically, the retrospective restatement requirement is a data archaeology project. AI can help by:

• Parsing historical board packages and management reports to identify expense categories that appeared regularly
• Analyzing historical GL data to reconstruct segment expense structures for prior periods
• Flagging inconsistencies between what was disclosed historically and what the reconstructed CODM reports suggest was reviewed

This is time-sensitive. Companies that deferred this work to the 2024 annual filing cycle found themselves under time pressure. Those planning for potential further FASB amendments (the FASB post-implementation review of ASU 2023-07 is ongoing) should build this infrastructure now rather than retrofitting it again.

Step 5: Segment-to-Consolidation Reconciliation

ASU 2023-07 expands reconciliation requirements: companies must reconcile the total of reportable segments' significant expenses to the corresponding consolidated line items. This has historically been a persistent manual pain point.

AI-driven reconciliation tools can automate the matching of segment totals to consolidated financial statements, flag variances, and generate audit-ready reconciliation schedules. This is one of the cleaner AI automation wins because the task is rule-based once the segment data dictionary is in place. For a broader look at how AI handles intercompany and consolidation reconciliation, see AI Intercompany Reconciliation Automation: 2026 CFO Evaluation Guide.

Step 6: Disclosure Drafting with Generative AI

Generative AI (large language models) is being piloted by large public companies and their advisors to draft segment footnote disclosures from structured data inputs. The workflow: structured segment data feeds into a generative AI tool, which produces draft disclosure language aligned to ASC 280-10-50 requirements.

The key risk is hallucination. AI can generate plausible but incorrect disclosure language, particularly around qualitative descriptions of segment performance or CODM methodology. Governance frameworks require human review of all AI-drafted disclosure text before filing, with documented sign-off by the controller or CFO.

The SEC's position is unambiguous: "The use of artificial intelligence does not alter a company's obligations under the federal securities laws. Companies remain responsible for the accuracy and completeness of their disclosures, regardless of whether those disclosures were prepared with the assistance of AI tools." For a detailed walkthrough of AI hallucination risks in financial reporting, see AI Hallucination in Financial Reporting: A 2026 Practitioner Walkthrough.

Step 7: SEC Comment Letter Pattern Detection

Segment reporting is consistently among the top 10 most frequent topics in SEC comment letters, according to Audit Analytics data covering fiscal years 2020 to 2024. The SEC's Division of Corporation Finance reviewed segment reporting disclosures in approximately 40% of its selective review letters in fiscal year 2024.

The four most common comment letter triggers post-ASU 2023-07:

1. Whether the company has correctly identified its CODM (KPMG calls this "the highest-risk area" because the new title/position disclosure forces a formal commitment the SEC can test)
2. Whether operating segments have been improperly aggregated
3. Whether non-GAAP measures have been included in segment disclosures
4. Whether "significant segment expenses" have been identified with sufficient rigor

AI tools trained on historical SEC comment letter patterns can scan draft disclosures and flag language that has historically attracted comments. This is a pre-filing quality control step, not a substitute for technical accounting judgment. You can search live comment letter patterns using SEC EDGAR full-text search to see how the SEC has challenged specific disclosure language.

Governing AI Outputs: What Auditors Will Ask

External auditors need to verify that AI-generated expense classifications and segment assignments are accurate and consistent. The audit trail requirement is not optional, and there is currently no established audit standard specifically for AI-assisted disclosure preparation. That means finance teams must build their own governance framework.

The minimum viable governance structure for AI-assisted ASC 280 compliance:

• Segment data dictionary with version control, change log, and documented approvals
• Human-in-the-loop review at every AI output stage: expense classification, reconciliation, and disclosure drafting
• Controller or CFO sign-off on all AI-drafted disclosure text before filing
• Documented rationale for significant segment expense determinations, including which CODM reports were reviewed and when
• Consistency testing across periods: AI-generated allocations must be checked for period-over-period consistency, particularly for Tier 3 shared costs

For a broader AI governance framework applicable across financial reporting workflows, see AI Governance Framework for Finance: The CFO's 2026 Practitioner Walkthrough.

Implementation Roadmap: Sequencing Matters

The sequence below reflects what works in practice. Skipping Step 1 and jumping to disclosure drafting is the most common mistake.

1. Inventory CODM reports (weeks 1 to 4): Catalog every management report, dashboard, and board package the CODM receives. This is the factual foundation for the "regularly provided" determination. AI document parsing accelerates this but requires human review.

2. Build the segment data dictionary (weeks 3 to 8): Map ERP cost centers and GL accounts to reportable segments. Document Tier 1, 2, and 3 cost classifications. Version-control from day one.

3. Run retrospective reconstruction (weeks 6 to 12): Use AI document parsing and historical GL analysis to reconstruct prior-period significant segment expenses. Reconcile to prior filings and document variances.

4. Automate ERP extraction and segment mapping (weeks 8 to 16): Configure AI-assisted data pipelines to pull segment-level data from ERP systems and apply the segment data dictionary. Test against manual outputs.

5. Implement reconciliation automation (weeks 12 to 18): Deploy AI-driven reconciliation tools to match segment totals to consolidated financials and generate audit-ready schedules.

6. Pilot generative AI for disclosure drafting (weeks 16 to 24): Run AI-drafted disclosure language in parallel with manually drafted text. Compare, identify gaps, and establish the human review and sign-off protocol.

7. Pre-filing comment letter scan (final two weeks before filing): Run AI pattern detection against draft disclosures. Address flagged items before submission.

8. Auditor walkthrough (concurrent with steps 6 and 7): Brief external auditors on the AI-assisted workflow, the segment data dictionary, and the human review controls. Don't surprise them at fieldwork.

ASC 280 vs. IFRS 8: Does AI Tooling Work Across Both?

IFRS 8 Operating Segments uses the same management approach as ASC 280 and is substantially converged with it. The core AI automation mechanics, CODM report parsing, expense tagging, segment-to-consolidation reconciliation, and disclosure drafting, apply equally to IFRS 8. Differences in aggregation criteria and entity-wide disclosure requirements mean that the segment data dictionary must be configured for the applicable standard, but the underlying tooling is the same. Multinational finance teams reporting under both standards can generally use a single AI platform with standard-specific rule sets.

FAQ

Which step in ASC 280 automation has the highest SEC comment letter risk? CODM identification and the "significant segment expenses" determination. KPMG's 2024 Segment Reporting Handbook flags CODM disclosure as the highest-risk area because the new title/position requirement forces a formal commitment the SEC can test against actual internal reporting structures.

Can AI handle segment reporting for a company that just completed an M&A transaction? With caution. AI tools that rely on historical CODM report patterns struggle when segment structures change mid-year. The segment data dictionary must be updated to reflect the acquired entity's cost centers before AI automation can run reliably. Under ASC 805, new segments may need to be identified and disclosed from the acquisition date. See AI and ASC 805 Purchase Price Allocation Automation for the M&A data integration context.

Does the 10% quantitative threshold test change under ASU 2023-07? No. The existing thresholds under ASC 280-10-50-12 remain: a segment is reportable if its revenue, profit or loss, or assets represent 10% or more of the combined totals of all operating segments. The 75% revenue test under ASC 280-10-50-18 also remains unchanged. ASU 2023-07 adds disclosure requirements on top of the existing identification framework.

What's the risk of using AI to draft segment footnote disclosures without human review? High. Generative AI can produce plausible but inaccurate disclosure language, particularly for qualitative descriptions of CODM methodology or segment performance. The SEC has stated explicitly that companies remain fully responsible for AI-assisted disclosures. A documented human review and sign-off process is not optional.

Is FASB likely to amend ASU 2023-07 further? Possible. FASB's post-implementation review is ongoing, with staff monitoring whether the "significant segment expenses" threshold is being applied consistently. A narrow-scope amendment in 2026 or 2027 is plausible if implementation issues prove widespread. Companies making multi-year technology investments in segment reporting automation should build flexibility into their segment data dictionary to accommodate potential rule changes.

How do we handle non-GAAP measures that appear in CODM reports? Carefully. The SEC permits voluntary disclosure of additional non-GAAP segment profit measures under ASU 2023-07, but they must comply with Regulation G and Regulation S-K Item 10(e). Deloitte's analysis of Fortune 500 filers found that only 1% disclosed a voluntary non-GAAP segment measure. AI tools must be specifically trained to flag non-GAAP language in CODM reports and ensure it is treated correctly in external disclosures, not silently carried through as a GAAP measure.