Explore selected excerpts from SEC filings, their topic labels, and the rules used to flag them. Model results are not available.
Evidence chain: source → deterministic baseline → model state → review and release
AuthoritySourceVerbatim bounded excerpt or public identifier
ComparatorDeterministic baselineVersioned rule IDs, terms, and exclusions
InterpretationStructured model outputInactive in this prototype
ControlReview and releaseStatus, errors, metrics, version
Interactive controls run locally in this browser. If they are unavailable, the static evidence remains available through this page and the public JSON snapshot.
Evidence explorer
Filter the filing record
The prototype contains a small source-checked sample, not an exhaustive company review. Filters operate only on records already present in this release; no arbitrary company lookup or URL retrieval occurs.
Showing 5 of 5 evidence items. Deterministic baseline is visible.
Prototype sample: 5 selected 10-K evidence items from 4 U.S. public companies, filed 2024-07-30 through 2025-07-30. The sample exercises the inspection workflow; it is not representative or exhaustive.
Microsoft CorporationMSFT · CIK 0000789019
AI positioning and adoption
Verbatim filing excerpt
We offer supercomputing power for AI at scale to run large workloads, complemented by our rapidly expanding portfolio of AI cloud services (including the latest models) and hardware, which includes custom-built silicon and strong partnerships with chip manufacturers.
Deterministic baseline and accounting coding
Rule candidate
Research use note (release-authored, not filing text): Distinguishes broad AI positioning from a disclosure that identifies operating infrastructure and product capacity.
The excerpt identifies supercomputing, cloud services, hardware, silicon, and supplier relationships as operating inputs, but does not quantify spending or performance.
AI at scale · AI cloud services · custom-built silicon
Excluded terms
None
Rules version
topic-rules@1-f75f317f416f
Structured model output
Not run
No fixed model release or frozen benchmark exists for this prototype; the source excerpt and deterministic baseline remain the only active evidence layers.
Classification
Unavailable
Model ID
Not assigned
Provider
Not assigned
Prompt version
Not assigned
Output schema
Not assigned
Input passage digest
Not bound
Confidence
Not measured
Walmart Inc.WMT · CIK 0000104169
Human capitalTraining and reskilling
Verbatim filing excerpt
An important part of our focus is to provide opportunities for associates to grow and learn.
Deterministic baseline and accounting coding
Rule candidate
Research use note (release-authored, not filing text): Records broad employee-learning language for review; it does not establish a named program, investment, participation, or outcome.
The excerpt mentions opportunities to grow and learn, but names no program, investment, participation measure, or outcome.
Matched rule IDs
HCM_EMPLOYEE_LEARNING_001
Matched terms
associates · grow · learn
Excluded terms
None
Rules version
topic-rules@1-f75f317f416f
Structured model output
Not run
No fixed model release or frozen benchmark exists for this prototype; the source excerpt and deterministic baseline remain the only active evidence layers.
Classification
Unavailable
Model ID
Not assigned
Provider
Not assigned
Prompt version
Not assigned
Output schema
Not assigned
Input passage digest
Not bound
Confidence
Not measured
Enphase Energy, Inc.ENPH · CIK 0001463101
RestructuringWorkforce reductions
Verbatim filing excerpt
In November 2024, we announced the 2024 Restructuring Plan, which includes a reduction of our workforce of full-time employees and contractors by approximately 17 percent, which will continue to be implemented until the first half of 2025.
Deterministic baseline and accounting coding
Rule candidate
Research use note (release-authored, not filing text): Connects a named restructuring program to workforce scope and implementation timing without treating the disclosure as a causal or assurance conclusion.
The excerpt names a restructuring plan, identifies employees and contractors as the affected population, quantifies approximate scope, and states an implementation horizon.
Restructuring Plan · reduction of our workforce · 17 percent
Excluded terms
None
Rules version
topic-rules@1-f75f317f416f
Structured model output
Not run
No fixed model release or frozen benchmark exists for this prototype; the source excerpt and deterministic baseline remain the only active evidence layers.
Classification
Unavailable
Model ID
Not assigned
Provider
Not assigned
Prompt version
Not assigned
Output schema
Not assigned
Input passage digest
Not bound
Confidence
Not measured
Alphabet Inc.GOOGL · CIK 0001652044
Human capitalTraining and reskilling
Verbatim filing excerpt
We provide a variety of high quality training and support to managers to build and strengthen their capabilities-–ranging from courses for new managers, to learning resources that help them provide feedback and manage performance, to coaching and individual support.
Deterministic baseline and accounting coding
Rule candidate
Research use note (release-authored, not filing text): Identifies the target of training and the capability-building objective while leaving expenditure and effectiveness unclaimed.
The excerpt identifies managers as the training population and capability development as the objective, but does not quantify program inputs or outcomes.
training · support to managers · strengthen their capabilities
Excluded terms
None
Rules version
topic-rules@1-f75f317f416f
Structured model output
Not run
No fixed model release or frozen benchmark exists for this prototype; the source excerpt and deterministic baseline remain the only active evidence layers.
Classification
Unavailable
Model ID
Not assigned
Provider
Not assigned
Prompt version
Not assigned
Output schema
Not assigned
Input passage digest
Not bound
Confidence
Not measured
Microsoft CorporationMSFT · CIK 0000789019
AI positioning and adoption
Verbatim filing excerpt
Microsoft is a technology company committed to making digital technology and artificial intelligence (“AI”) available broadly and doing so responsibly, with a mission to empower every person and every organization on the planet to achieve more.
Deterministic baseline and accounting coding
Rule candidate
Research use note (release-authored, not filing text): Records broad AI positioning language for review; it does not establish an operating strategy, investment, adoption, or outcome.
The excerpt mentions AI in mission and responsibility language, but does not identify an operating activity, resource, expenditure, adoption, or outcome.
Matched rule IDs
AI_STRATEGY_001 · AI_RESPONSIBLE_USE_001
Matched terms
artificial intelligence · available broadly · responsibly
Excluded terms
None
Rules version
topic-rules@1-f75f317f416f
Structured model output
Not run
No fixed model release or frozen benchmark exists for this prototype; the source excerpt and deterministic baseline remain the only active evidence layers.
Classification
Unavailable
Model ID
Not assigned
Provider
Not assigned
Prompt version
Not assigned
Output schema
Not assigned
Input passage digest
Not bound
Confidence
Not measured
No evidence item matches these filters
Reset the filters or choose another company, form, or topic. This state does not imply that the filing contains no relevant disclosure.
Illustrative cross-year excerpts
Selected Microsoft AI disclosure language
These excerpts come from different filing locators and describe different disclosure objects. They demonstrate traceable wording inspection only; the prototype does not code them as an added, removed, or changed disclosure.
10-K · 2024-06-30
Microsoft Corporation
Microsoft is a technology company committed to making digital technology and artificial intelligence (“AI”) available broadly and doing so responsibly, with a mission to empower every person and every organization on the planet to achieve more.
Comparison note (release-authored, not filing text): This 2024 overview excerpt and the selected 2025 infrastructure excerpt come from different filing locators; they illustrate wording specificity and are not coded as a disclosure change.
We offer supercomputing power for AI at scale to run large workloads, complemented by our rapidly expanding portfolio of AI cloud services (including the latest models) and hardware, which includes custom-built silicon and strong partnerships with chip manufacturers.
Comparison note (release-authored, not filing text): This 2025 infrastructure excerpt and the selected 2024 overview excerpt come from different filing locators; they illustrate wording specificity and are not coded as a disclosure change.
The rule layer identifies candidates for inspection through versioned rule IDs, matched terms, and exclusion evidence. It reports no probability or quality score and makes no conclusion about completeness or materiality.
Topic definitions
Human capital: Disclosure about employees as an operating resource, including workforce composition, development, retention, and organizational capability.
Restructuring: A defined reorganization, exit, disposal, rightsizing, or cost-reduction program with an identified operational or accounting consequence.
Workforce reductions: Disclosure of an announced, initiated, or completed reduction in employees, positions, or workforce capacity.
Training and reskilling: Disclosure of employee or manager training, learning, reskilling, or capability-building language, including bounded descriptive commitments that do not establish program scale.
Recruiting capacity: Disclosure about the ability, constraint, or infrastructure involved in attracting and hiring employees.
AI positioning and adoption: Disclosure that connects artificial intelligence to products, operations, infrastructure, investment, strategic priorities, or clearly stated positioning; broad mission language remains descriptive.
AI-related risk: Disclosure of operational, governance, legal, financial-reporting, cybersecurity, or workforce risk arising from AI systems or adoption.
Matching rules
Each matched rule ID resolves to this bounded, literal phrase definition. Matching is case-insensitive phrase inclusion with the stated any/all mode and exclusions; it is reproducible on the excerpt and is not a probability.
AI_STRATEGY_001 — AI positioning (ai-strategy; any of: artificial intelligence, available broadly). Exclude: third-party ai. Flags a literal AI positioning reference for human inspection; it does not infer adoption, spending, or outcomes.
AI_RESPONSIBLE_USE_001 — Responsible AI positioning (ai-strategy; any of: responsibly). No exclusions. Flags responsibility language adjacent to an AI reference without treating it as governance evidence.
AI_INFRASTRUCTURE_001 — AI infrastructure capacity (ai-strategy; any of: AI at scale). No exclusions. Flags explicit infrastructure-capacity language for inspection; no capacity, spend, or performance is estimated.
AI_CLOUD_SERVICES_001 — AI cloud services (ai-strategy; any of: AI cloud services). No exclusions. Flags a literal AI cloud-services reference as an operating description, not as a quantified result.
AI_HARDWARE_001 — AI hardware or silicon (ai-strategy; any of: custom-built silicon). No exclusions. Flags explicit hardware or silicon language for inspection without inferring supplier dependence or investment scale.
HCM_EMPLOYEE_LEARNING_001 — Employee learning language (training-reskilling; all of: associates, grow, learn). Exclude: customer. Flags a bounded employee-learning statement while preserving the limitation that no named program, cost, participation, or outcome is disclosed.
HCM_MANAGER_TRAINING_001 — Manager training (training-reskilling; all of: training, support to managers). Exclude: customer training. Flags training directed to managers and leaves program scale and effectiveness unclaimed.
HCM_CAPABILITY_BUILDING_001 — Manager capability building (training-reskilling; any of: strengthen their capabilities). No exclusions. Flags explicit capability-building language without converting it into a measured human-capital outcome.
RESTRUCTURING_NAMED_PLAN_001 — Named restructuring plan (restructuring; any of: Restructuring Plan). No exclusions. Flags a named restructuring plan and preserves the filing's stated scope and timing.
WORKFORCE_REDUCTION_001 — Workforce reduction (workforce-reductions; any of: reduction of our workforce, workforce reduction). Exclude: risk of workforce reduction. Flags a stated workforce reduction without inferring causes or downstream effects.
QUANTIFIED_SCOPE_001 — Quantified workforce scope (workforce-reductions; any of: 17 percent). No exclusions. Flags an explicit percentage in the bounded excerpt; the displayed amount remains filing-authored and approximate when stated that way.
Model output remains unavailable
The model view is intentionally rendered as “Not run.” It has no provider, model ID, prompt version, confidence, latency, cost, or summary because no fixed release and benchmark exist.
Scope boundary
Evidence inspection, not automated judgment
This prototype does not provide investment recommendations, assurance opinions, causal conclusions, unrestricted prompting, arbitrary URL retrieval, confidential uploads, or user accounts. Missing extracted evidence does not establish that a disclosure is absent.