Draft all annual grant Form 4s simultaneously by uploading a spreadsheet with the grant information. Section 16 Agent uses AI to parse the spreadsheet, so you do not need to reformat anything.
Draft forms simultaneously for reporting persons at different issuers. Each issuer keeps its own isolated workspace, settings, roster, and footnote history.
The application uses AI only for the steps where deterministic rules are underfit, otherwise strict rules control. This combination harnesses the power of AI while providing similar predictability and reliability of traditional computing. We also test the application's accuracy to empirically measure its performance.
Footnotes are suggested utilizing the reporting person and issuer’s house-style. If provided in the prompt, live values such as dates, prices, etc. are updated automatically subject to the user’s review.
Insider transaction details are extraordinarily sensitive, and may constitute MNPI before the filing is made. The application treats them that way by default and is architected to protect confidentiality.
Runs on Amazon Bedrock in our own AWS environment — never stored, never trained on, never seen by a human.
Encrypted in transit and at rest, with authenticated sessions on every request.
Each organization's data is scoped to its own workspace, visible only to its own team.
Nothing exports without your sign-off, and every export is acknowledged and logged.
Deletion is complete — the document, its prior versions, and the underlying record.
Section 16 Agent is the world's first and only automated Form 3 and Form 4 drafting tool. Section 16 Agent drafts Forms 3 and Forms 4 (Form 5 support is planned for Q4 2026) from a simple description provided by the user either via email to Sam, an AI drafting agent, or in the application's user-friendly interface. It replaces the old, pre-AI era method of manually filling out Word or web templates and massively simplifies the workflow--saving time, reducing deadline risk, maintaining accuracy, while freeing practitioners to work on more valuable matters. It handles multiple transactions simultaneously, understands derivative securities and holdings rows, and automatically configures itself to the filing style and preferences of each company it serves. It even automatically drafts specific footnotes that are identical to the user and their company's house style, automatically updating e.g. price and dates and vesting terms, leaving only a quick review and approval for the user. It does all this in service of one goal: to make filings effortless.
A combination of deterministic rules and AI. AI interprets the user's prompt and processes it to draft a form that satisfies EDGAR's exact specifications. Users can export a PDF draft and/or an .xml EDGAR-submission package.
Both. In-house counsel and corporate secretaries use it to draft filings for their own issuer's directors and officers. Outside counsel use it to draft for every issuer they represent: each issuer gets its own workspace holding that issuer's insider roster, reporting-person details, and footnote history, and one login switches between workspaces. Every workspace is isolated, so a firm's clients never see one another's information.
Yes. In-house counsel can draft multiple forms simultaneously by either describing each reporting persons' transaction(s) or uploading a spreadsheet, such as one provided by the equity plan administrator for an annual officer or director grant. Outside counsel can select any issuer and reporting person from their client roster and describe each applicable transaction(s). From there, just one click and all forms will generate simultaneously. The result: dozens of Forms 4 can be drafted in seconds.
All data is encrypted in transit and at rest. Every session is authenticated, and each organization's workspace is isolated so its information is visible only to its own team.
No. AI processing runs on Amazon Bedrock within an AWS environment controlled solely by Section 16 Agent, under an enforced zero-retention configuration: inputs and outputs are never stored, never used to train any model, never shared with the model provider, and never accessible to a human operator. Every AI call in the application routes through a single enforcement point, so no stage of the pipeline can bypass that policy.
Within your organization's workspace, whether that is the issuer or its outside counsel, and nowhere else. Details of an insider transaction are some of a reporting person and issuers' most sensitive information, so the application treats them accordingly by default and keeps the retained footprint deliberately small. The submitted description is carried into the draft so that the user can verify the generated form against the original wording. That draft is the requested deliverable and remains under the organization's control. Everything surrounding it is short-lived: footnote working data clears within 24 hours, unsaved drafts within 30 days, and pipeline records retain only job status and CIKs, never the description itself. Forms saved to History persist until the organization deletes them, and deletion is complete, removing the document, its prior versions, and the underlying record. All data is encrypted in transit and at rest, and scoped so that it is visible only to the organization that created it.
Not yet. While in Beta, Section 16 Agent produces the EDGAR-ready XML submission package which may be filed via submission on EDGAR by an account administrator for the reporting person (someone who owns their CIK & CCC codes). Section 16 Agent provides instructions on how to self-file with every exported form. Alternatively, the same EDGAR-ready submission package can be provided to a traditional filing service. Eventually, direct filing support is planned.
Speed, accuracy, and less repeated work. Drafting that takes an hour of careful manual entry takes seconds, boilerplate like footnotes and reporting-person details are reused instead of retyped, and formatting rules are checked automatically.
Full access, all completely free during the public beta.