FOLIO Ai brings authoring and document control to the flight operations manuals you keep in Word: revise them, release each revision to crews and staff by base, role and fleet, and see who has read and signed it. Word chapters import with their structure and export back with change bars, AI review reads each section beside the regulations it cites, and an approver other than the author signs off each proposed change.
Email me the brochure, FOLIO Ai
FOLIO Ai is FlightAtom's aviation manual authoring and document control software for flight operations manuals and other controlled manuals. Document controllers import Word chapters with their structure recovered, run AI review against the text of each regulation a section cites, send proposed changes to an approver other than the author, and export revisions back to Word with change bars. Releases are distributed by base, role, aircraft type and route, and each reader's read-and-sign acknowledgment is bound to the exact release.
FOLIO Ai's AI review sets a manual section beside the bundled OTAR Part 121 text of each rule it cites and is set to raise only findings worth a ticket. Every fix waits for a person to accept it.
FOLIO Ai screens with synthetic demonstration manuals, from the Word import dialog to the auditor's evidence room. The AI review findings shown are illustrative. Open any screenshot for a closer look.
The import dialog shows the structure FOLIO Ai recovered from a Word chapter, with each block typed as text, table, list, caution, warning or note, and the reconciliation check that seals a signed receipt over the exact source file on import. Synthetic demonstration data.
Section 8.1.3 in the editor with its AI review panel: findings rated by severity and category, each quoting the passage, citing OTAR 121.280 where relevant and proposing replacement wording for a person to accept, reject or defer. The findings shown are illustrative examples written for this demonstration and returned through FOLIO Ai's review pipeline, not live model output. Synthetic demonstration data.
The technical reviewer opens the diff of a submitted proposal. FOLIO Ai froze both chapters of the author's draft on submission, and the live-workspace view lists each block the proposal modifies or adds, with its section path, the struck-through live wording and the proposed wording. Synthetic demonstration data.
A proposed change in its second approval stage: the technical review is complete, the quality sign-off needs two approvals under a four-eyes rule, and each signed decision records the stage, reviewer, decision, rationale and evidence hash. Synthetic demonstration data.
At the end of a controlled release, the assigned reader types their workspace name, confirms they read this exact release and re-confirms their identity before the acknowledgement is recorded, with the active page time kept as supporting evidence. Synthetic demonstration data.
Verified Knowledge answers a question with an exact extract from section 8.3 of the active release, using deterministic extractive search with no model-generated confidence score, and verifies a signed query receipt. Synthetic demonstration data.
An external auditor opens one sealed release through an expiring link without a FOLIO Ai account, and sees the release seal, artifact hash, acknowledgement totals and the frozen chapter content the administrator chose to include. Synthetic demonstration data.
FOLIO Ai keeps the Word source, every committed change, each release and every reader's acknowledgment traceable to a numbered revision.
FOLIO Ai rebuilds each chapter's sections from heading styles or numbered paragraphs, including WARNING, CAUTION and NOTE callouts, lists, merged-cell tables, and MEL-style chapter and ATA item banners such as 21-31-11. Every import is checked against an inventory of the Word file: unresolved tracked changes, macros and embedded objects stop it, other exceptions are flagged for review, and an accepted import gets a signed receipt over the exact source file. Text-based PDFs import too, split into chapters from their bookmarks with headers, footers and contents pages left out; a scanned PDF without a text layer is refused until it has been through OCR.
Authors work in private drafts, and reviewers see a section-by-section diff of exactly what was frozen for approval. Approval workflows can run as serial stages with parallel reviewer pools, minimum approval counts, a four-eyes rule and due dates, with each reviewer confirming their identity, by password or a single sign-on check, before the decision is signed. Self-approval and changes built on an outdated revision are refused.
Each release freezes signed content tied to its revision and is assigned by department, base, role, aircraft type, tail, route and jurisdiction, with a deadline to acknowledge. Readers open a what-changed view at chapter, section and block level, then acknowledge by typing their name and re-confirming their identity, and the receipt is bound to the exact release they read. An administrator report shows every assignment as acknowledged, pending, overdue or revoked, and verifies each receipt's signature.
Edited chapters export back to Word from the imported source file, with changed paragraphs patched in place and marked by margin change bars. Authors edit in typed blocks for procedures, warnings, checklists, tables, diagrams, abbreviations and linked cross-references. A manual-wide table of contents rebuilds as the outline changes and refuses to build from an incomplete manual, and any chapter can be printed or saved as a PDF.
An evidence room shares one release with an auditor through an expiring, revocable link, with every access logged, and an administrator chooses whether it includes the manual text and acknowledgment receipts. Signed audit packages freeze approvals, requirement mappings, release history, distribution and acknowledgment coverage into a download that is verified before it is served, and the revision log records every committed change, administrator edits included, with a SHA-256 hash of its content.
Load requirements from any authority as versioned records and pin each one to the exact manual section that meets it, with an owner, due date, review interval and full action history. Definitions, warnings and policy text are versioned as governed blocks pinned to every section that uses them, and a preview lists each affected section before a new version is adopted. An impact scan opens owned tasks when a mapped requirement or shared wording changes, an evidence review falls overdue or a cross-reference target disappears.
Verified Knowledge answers a question only with exact quotations from the released manuals the reader is entitled to see, links each quotation to its section, and says when the evidence is too weak. It is deterministic search with no generative AI. Readers also keep private bookmarks, notes and highlights, and copies saved in the browser show whether they are verified as the current revision.
Administrators get descriptive measures of requirement mapping, acknowledgment coverage, pending proposals, overdue approval stages and open assurance tasks, with no risk score. Every figure shows the formula behind it, percentages also show their numerator and denominator, and the source records open in a drill-down and export to CSV.
Grant view, edit, review, publish, distribute and access-management rights separately for each manual; an active policy denies by default, administrators included. People sign in with their Microsoft work account, under your organization's multi-factor policy, through an OpenID Connect broker that also sets their FOLIO Ai role, and administrators can revoke a member's sessions. A versioned API with scoped service accounts, webhooks and a security-event feed connects FOLIO Ai to other systems.
FOLIO Ai puts AI review inside the manual workflow, beside the regulation text and the draft it checks. Every finding quotes the passage it concerns, and an accepted fix reaches readers only in a release that a person publishes.
AI review reads each section alongside the bundled OTAR Part 121 text of each rule it cites and is set to flag only findings worth raising as a ticket. Each is filed as OTAR compliance, missing OTAR reference, spelling, grammar, consistency or clarity, with a severity, the quoted passage, ready-to-apply replacement wording and, for OTAR findings, the regulation reference.
Review a chapter or the entire manual in one batch, steered by reviewer instructions, with findings filed per chapter and exported to a spreadsheet. AI also maps which sections reference or depend on each other across the manual, and maps each OTAR Part 121 requirement to the sections of the open chapter that address it, marked covered, partial or missing with the quoted passage.
Reviewers accept, reject, defer or resolve each finding with their name and a note. Accepting writes the correction into the working copy, never straight into a release. Before a run, FOLIO Ai tells users which content goes to which external AI provider, and it re-checks the requester's review rights before a queued review sends content and again before findings are saved.
Agree a representative sample and a baseline from your current process. These are proposed evaluation measures, not published performance results.
A complex DOCX manual and a representative change request.
Review the import, draft, independent approval, release, and acknowledgment.
Import corrections; revision cycle time; acknowledgment completion.
Yes. FOLIO Ai imports Word (DOCX) chapters and rebuilds their sections from heading styles or numbered paragraphs, including WARNING, CAUTION and NOTE callouts, lists, merged-cell tables, bracketed OTAR references, and MEL-style chapter and ATA item banners. Unresolved tracked changes, macros or embedded objects stop the import until they are fixed in Word, other exceptions are flagged for review, and an accepted import gets a signed receipt over the exact source file. Text-based PDFs import too, split into chapters from their bookmarks with headers, footers and contents pages left out; a scanned PDF without a text layer is refused until it has been through OCR.
Yes. Edited chapters export back to Word from the imported source file in FOLIO Ai's standard layout, with changed paragraphs patched in place and marked by change bars in the margin. A manual-wide table of contents rebuilds as the chapter outline changes, and any chapter can be printed or saved as a PDF.
FOLIO Ai's AI review reads each section alongside the bundled OTAR Part 121 text of each rule it cites. It returns findings for OTAR compliance, missing OTAR references, spelling, grammar, consistency and clarity, each with a severity, the quoted passage and replacement wording, plus the regulation reference for OTAR findings. A whole chapter or manual can be reviewed in one batch run, and findings export to a spreadsheet.
No. AI returns review findings and replacement wording, and nothing changes until a person accepts a finding. An accepted correction goes into the working copy, not into a release. Authors submit their changes as a proposal that someone other than the author must approve, and readers see a change only after an authorized person publishes a new release.
Each release is assigned by department, base, role, aircraft type, tail, route and jurisdiction. The reader confirms they have read and understood it, types their name, and re-confirms their identity with a password or single sign-on. The receipt is bound to the exact release they read, and revoked assignments cannot be acknowledged.
Yes. An evidence room shares one release with an auditor through an expiring, revocable link, with every access logged, and an administrator chooses whether it includes the manual text and acknowledgment receipts. Rooms last a week by default and up to a year. Signed audit packages freeze approvals, requirement mappings, release history, distribution and acknowledgment coverage into a verified download.
No. Verified Knowledge answers a question only with exact quotations from the released manuals the reader is entitled to see, links each quotation to its section, and says when the evidence is too weak to answer. It uses deterministic text retrieval, and each response carries a signed receipt.
FlightAtom hosts a separate FOLIO Ai deployment for each operator and connects it to your Microsoft sign-in. Support comes directly from the FlightAtom team by email and phone, and your FOLIO Ai administrator manages access and roles day to day. Support hours, response times, backups and recovery objectives are agreed in writing for each deployment rather than published as a standard service level.
Staff access and each person's administrator or author role come from your Microsoft directory and are re-applied at every sign-in. Releases reach readers through rules by department, base, role, aircraft type and route, and one report shows every assignment as acknowledged, pending, overdue or revoked. The reader profiles behind those rules are maintained in FOLIO Ai, with no HR or roster sync.
Yes, through defined interfaces. Staff sign in with Microsoft work accounts through an OpenID Connect broker. Manuals come in from Word or PDF and go out as Word with change bars, printable PDF chapters and CSV reports. A versioned API with scoped service accounts serves release records, release events, signed exports and a security-event feed, and webhooks notify other systems through a delivery worker your IT team runs.
Yes, in a web browser. Readers save the releases assigned to them and open them without a connection, and FOLIO Ai shows whether each saved copy is still verified as the current revision. Offline search runs when the full set of saved releases is present and was verified within the last 15 minutes. Readers need only a browser; there is no EFB or native app.
Your team brings the change in, from Word, a text-based PDF or the FOLIO Ai editor, and it goes through approval and a controlled release with a what-changed view for readers. FOLIO Ai does not receive OEM bulletins automatically. Requirements from any source can be loaded as versioned records pinned to manual sections, and an impact scan opens a task when a mapped requirement gets a new version.
No. Hosting region, backups and recovery objectives are set for each deployment. To measure the effect yourself, FOLIO Ai's administrator reports show acknowledgment coverage, pending proposals, overdue approval stages and open assurance tasks, each with its formula, numerator and denominator, and export the source records to CSV for a before-and-after comparison.
Yes. FOLIO Ai has its own manual workspace, sign-in, authoring, review, approval, release and distribution workflows. AURORA Ai is not required to use FOLIO Ai.
As your fleet, routes, rules or teams change, we adapt FOLIO Ai to match, and you approve each change before it goes live.
A focused walkthrough, with the people and decisions that matter to your team.