AI that understands video content itself
With the May 7, 2026 update, Dive adopted a large-scale AI (VLLM) that understands both video and language. Rather than relying solely on speech transcription, it understands video content itself and generates contextually appropriate procedures. By interpreting the meaning of actions from objects in the video (parts, tools, equipment, etc.) and human movement (steps, actions, operations), it can segment and summarize what's happening even in videos without narration.
Create two types of subtitles
- Speech transcription subtitles. Transcription accounts for the full video context, significantly reducing garbled subtitles from mishearing.
- Action description subtitles. The AI automatically generates subtitles explaining work content, including segments without narration.
Automatic annotation and quiz insertion
The AI identifies focal points from the video and automatically positions text annotations on the video with display durations. Confirmation quiz questions are inserted automatically as well.
Share floor context with the AI
- Prerequisites specific to that video (equipment, purpose, tools, etc.) can be entered in the video description at analysis time and will be reflected in the analysis.
- Prerequisites that remain constant across videos, such as industry or company-specific terminology, can be registered once in the Terminology Glossary, eliminating the need to re-enter them each time and automatically reflecting them in analysis.
Work improvement suggestions
After analysis, you can generate improvement suggestions for the assembled procedure. Classifications include six types: Eliminate, Combine, Rearrange, Simplify (ECRS) plus safety and quality. Suggestions are for review only and don't automatically update the procedure.
Limitations and notes
- Dive does not retrain AI models separately for each customer company. Floor context, Terminology Glossary, and video description templates are provided before analysis to align with your organization's terminology and style.
- Customer data is not used internally for retraining or data review.
4 Fields That Make Up Every Step
Every Dive step consists of the same 4 fields. AI generates initial values for all 4 fields from your video, and creators simply add details while playing the video in the editor—tacit knowledge becomes structured text.
| Field | Content | On-screen color |
|---|---|---|
| Tasks to do (body text) | What to do. Work content, operations, the procedure itself | - |
| Work tips | How to do it well | Gray (neutral supporting info) |
| Quality caution | What causes failure | Blue |
| Important safety information | What is dangerous | Red |
Distinguishing Quality caution from Important safety information
The distinction is based on "whether personal injury risk is involved." Items involving personal injury risk are Important safety information; quality, material, and process concerns without personal injury risk are Quality caution. When in doubt, items without personal injury risk default to Quality caution.
Terminology and export
- The official on-screen names are "Work tips," "Quality caution," "Important safety information," and "Question"—4 terms used consistently across all screens.
- PowerPoint export places each step on one slide, including step number, title, task content, work tips, quality caution, important safety information, question, and thumbnail image (PowerPoint export is Pro plan and above).
AI-Powered Editing and Importing Existing Materials
AI handles automatic question generation, subtitle creation (with or without audio), automatic removal of redundant scenes, and automatic annotation insertion. Additionally, you can set playback speed per step—speed up redundant sections and slow down key parts.
AI organization (generative AI text refinement)
Based on text entered in the step's "Tasks to do" field, AI organizes the step title, work tips, quality caution, and important safety information. Organization has 4 modes:
- Change tone only
- Full text (reorganizes title, body, tips, quality caution, and important safety information based on the task description)
- Body text only (summarizes as bullet points)
- Simplify title (20 characters or fewer)
Choose the tone: command (plain form and imperative—standard for SOPs), polite language (for new-hire training and audit submissions), or persuasive (soft tone for training). Subtitle tone can be set separately from body text. From the SOP's basic info, use [Generate title and overview with AI] to create a title and summary from all step text. AI organization is a Pro plan feature and above.
Extract safety and cautions from existing manuals and apply to each step
When you attach existing manuals—work standards, drawings, checklists, reference materials—as "Related files" to your SOP, AI extracts "Tasks to do," "Work tips," "Quality caution," and "Important safety information," then assigns them to relevant steps.
- Extraction results appear as step-by-step candidates before application, so you can change target steps or categories, then approve or reject each item.
- Related files can be set to "Hide from viewers," keeping internal materials closed to management only.
Create SOPs from PDF and Office files
AI can auto-generate title, overview, and steps (including questions and pass/fail judgment) from Excel (.xlsx/.xlsm/.xls), Word (.docx/.doc), PowerPoint (.pptx/.ppt), OpenDocument, or PDF to create a complete SOP. Selectable options during creation include:
| Setting | Options |
|---|---|
| Body text mode | Format / Transcribe |
| Step split method | By content (individual rows or bullet points within a page become separate steps) / Per page (1 sheet = 1 page) |
| Text style | Command / Polite language / Persuasive / Follow policy |
| Questions | Create / Don't create |
| Pass/fail judgment | Add / Don't add |
| Excel PDF conversion method | 1 sheet = 1 page / Fit width to 1 page / Follow print settings |
Procedure Authoring Policy (Enterprise plan)
Define group-level "Procedure Authoring Policy" (phrasing/writing style, prohibited practices, common phrasing). When you select [Follow policy] in AI organization's text style, AI refines steps according to policy phrasing and granularity.
- Register sample SOPs or manual files, and AI auto-generates a policy draft (phrasing, prohibited practices, common phrasing).
- Use [Policy check] to have AI verify that your SOP follows organizational Procedure Authoring Policy.
Limitations and notes
- AI organization (text refinement) is Pro plan and above; Procedure Authoring Policy is Enterprise plan only.
- Precise video-to-procedure alignment is currently in beta.
- We do not use customer data internally for retraining or review.
- Manual step-splitting operations are patented by Episodic (Patent No. 7546989). Automatic video splitting is Patent No. 7716145.
Terminology Glossary and Dictionaries
Tools to align AI-generated text with your company's terminology and notation. [Dictionaries / Templates] has 5 tabs: Pronunciation Dictionary, Terminology Glossary, Notation Unification Dictionary, Templates, and Tags.
3 dictionaries
| Dictionary | What it does | Plan |
|---|---|---|
| Terminology Glossary | Register in-house terms, official names, and translations. This notation is prioritized during video and PDF AI analysis, and registered translations by language are prioritized in auto-translation during viewing | Core plan and above |
| Notation Unification Dictionary | Register replacement rules (e.g., "motor" → "motor"), which automatically apply to final AI analysis text. In the SOP editing screen, apply to all steps with a button | Pro plan and above |
| Pronunciation Dictionary | When voice narration mispronounces kanji, units, or proper nouns, specify the correct reading with furigana | - |
Registration and management
- Start the Terminology Glossary by importing from industry-specific templates: manufacturing, care, construction, food, logistics, retail, or office.
- Dictionaries can be bulk-imported via Excel or CSV, with a diff preview showing additions, updates, skips, and errors before import.
- Once registered in the Terminology Glossary, you won't need to re-enter in-house terms for each video.
Permissions and approval
Direct dictionary editing, bulk saves, and proposal approval are handled by the owner and team administrator by default. Content registered by group administrators and below is treated as a "proposal" and takes effect when the owner or team administrator approves it.
The owner can change this scope. For the Pronunciation Dictionary and Terminology Glossary separately, set user permissions for submitting proposals and for approving/directly editing—ranging from "Owner only" to "View-only user and above." Permitting direct editing down to created users allows implementation without approval (Notation Unification Dictionary inherits Pronunciation Dictionary permissions).
Structured video work instructions (video, text, checkpoints)
Dive work instructions are structured to combine text and checkpoints with video segmented by step. Text supplements judgment criteria and tips that video alone cannot convey, while checkpoints verify "correctness." This differs from traditional video manuals that present continuous footage as-is.
Our internal research shows that even in manufacturing facilities using video manuals, 87% still "end up asking a senior colleague" despite viewing. Structuring with video, text, and checkpoints aims to reduce this "asking anyway" behavior.
Checkpoints
- Set a checkpoint at each step. Five types are available: Yes/No selection, single choice, multiple choice, free text input, and numeric input.
- When you mark correct answers in the options, you can configure the system so viewers cannot proceed to the next step until they answer correctly.
- For numeric input, set lower and upper limits to define the "Correct range" (for training scoring) and "Normal range" (for inspection use). You can also set one-sided limits.
- Steps can be configured to "Require photo submission," letting workers submit images they capture as answers.
Checkpoints are available on Core plan and above.
Pass/fail judgment and inspection checklist operation
- Switch a checkpoint to "Use as pass/fail judgment" to change from correct/incorrect to normal/abnormal judgment. Even when abnormal, the work instruction doesn't stop—the result is recorded as "abnormal," enabling inspection checklist operation.
- The view data modal shows "who answered what and when." Rows represent one execution, columns represent each checkpoint in that version, and cells show answer values and OK/NG status.
Per-step inspection and pass/fail records require Core plan and above; CSV export of view data requires Pro plan and above.
Conditional branching
- Switch which step displays next based on the checkpoint answer. Branch destinations are "To next step," "Complete (END)," or "Any specific step."
- Five checkpoint types trigger branching: Yes/No, multiple choice, single choice, text input, and numeric input. You can also jump from a step without checkpoints to a specific step using the flow diagram.
- Edit branches while viewing the complete flow diagram; clicking a branching step color-codes its destinations. The branching configuration screen is PC-only.
Limitation: work instructions with branching cannot be exported as MP4. Conditional branching requires Pro plan and above.
Draft, approval, and publication
Work instructions flow through draft, Approval workflow, and publication stages. Approval workflow is configured per group; publication and revisions cannot execute without approver sign-off. Revision history is retained. Approval workflow requires Pro plan and above.
Draft Sharing (pre-publication review)
Drafts before publication can be shared—separate from the approval workflow—with users and groups designated by the author. Shared recipients can view content and leave comments, but cannot edit, save, or approve. You can share with recipients lacking access rights to the destination folder; notification emails are sent when shared and when comments are posted. Users under approval review can view drafts without being explicitly designated.
Plan conditions (features in this section)
| Feature | Available plans |
|---|---|
| Checkpoints per step | Core and above |
| Per-step inspection and pass/fail records | Core and above |
| Conditional branching | Pro and above |
| CSV export of view data | Pro and above |
| Approval workflow | Pro and above |
| Draft Sharing and comments | All plans including Free |
Cross-workspace search and AI search
As work instructions grow, anything unfound goes unused. Dive offers full-text search and AI search. Both are available on all plans including Free.
Full-text search scope
- Work instruction title and overview
- Step content (Step title, text, Work tips, quality notes, Important safety information, checkpoints)
- Text in video and Subtitles in the video
- Text objects in AR
- Tags assigned to work instructions and Folders
Search syntax
- Supports phrase search ("…"), exclusion (-word), and AND/OR operators.
- Results rank by weighted score, prioritizing title-related content.
- Tolerates spelling variation and typos—searching "サーバ" finds "サーバー."
- Filter by tags (multiple selection, OR logic).
AI search (conversational)
- Ask in natural language what you want to do or what problem you face; AI identifies matching work instructions and steps across your workspace.
- Cites the work instructions and steps backing the answer and tells you which step number to view.
- Asks clarifying questions when information is insufficient. Continue the conversation with follow-up questions; chat history is saved.
- Speak your question via the microphone icon, freeing both hands during work.
AI search recommends only from work instructions the user can access.
Offline use
With pre-caching, work instructions function offline without internet just as they do online. Translation works offline too.
What saves to the device
- Work instruction content, media (video and images), translation, and narration are saved together.
- Select a translation language before pre-caching, and that language's translation and audio save with it. Skip selection to save original only.
- Work records during offline use (checkpoint answers, etc.) stay on the device and sync automatically once connectivity returns.
Browser vs. app differences
- Offline playback works in web browsers. Offline, pre-cached work instructions appear at a dedicated page (https://app.divedx.com/offline). Available on PC and Windows tablets.
- The smartphone app auto-downloads video during normal viewing, stably accumulating it in app Storage—unlike browsers. We recommend the app for poor-signal sites.
Multi-resolution and adaptive delivery for low-bandwidth environments are available on all plans including Free.
Limitations
Enterprise plan includes "Device cache settings" to switch to a mode storing no View data on the device. When enabled for a team, offline playback and pre-caching disable, requiring constant network connection.
QR Codes, Printing & Portability
Issue QR codes and print in bulk for posting on site or equipment. Scan the code to access procedures right where you need them.
Issuing QR Codes
- QR codes can be issued at the procedure or folder level. Scanning a folder's QR code opens the procedure list for that folder.
- Enable "Open from the specified step" in the share modal to issue a URL and QR code that opens a specific step instead of the first one. You can also set expiring external share URLs.
- Issue time-limited QR codes that become inaccessible after expiration.
Bulk Printing QR Codes
- Select multiple procedures and subfolders within a folder, then bulk print their QR codes onto A4 paper.
- Choose from 1, 4, 9, or 24 QR codes per A4 sheet.
- Toggle accompanying information individually (name is on by default; description, step count, URL are off by default; thumbnail). QR code size automatically optimizes based on layout and displayed fields.
- Folder rows display "Folder" instead of step count and are distinguished by solid cell borders.
Exporting to Paper & Files
- Four export formats available: PDF, Excel, PowerPoint, and video (MP4).
- When exporting to PDF, Excel, or PowerPoint, enable "Add a jump URL to each step" (on by default) to embed links in each page, row, or slide that open that step directly. Print it on paper and still jump back to the original video procedure.
- PowerPoint export uses one slide per step, including a cover slide (procedure name, description, revision number) and revision history slide (revision, changes, proposer, proposal date, approval date, approver).
- MP4 export outputs the entire procedure as video at up to 30 fps and 720p resolution, with the same look and feel as Dive's interface.
- Excel export supports: procedures, procedure lists within folders, Skill Map, individual user procedure view history, and member/group lists.
Plan Requirements (Features in this section)
| Feature | Available Plans |
|---|---|
| QR code issuance, external sharing, MP4 export | Lite and above (Free excluded) |
| PDF export, Excel export | Core and above |
| PowerPoint export, SOPs list export | Pro and above |
Desired Procedures (Idea Board)
"Desired procedures" is a board where everyone—including view-only users—posts "I need this procedure." Others vote with "+1," and vote count directly drives creation priority. Each person votes only once per request. Available on all plans.
The same board serves two roles. For viewers, it's "Desired procedures." For creators, it's an "Idea board" for deciding what to build next.
From Request to Completion
- Requests are tracked by status: "Desired," "In Progress," or "Complete."
- Marking a request "Complete" links the fulfilled procedure to the card and notifies the requester and all "+1" voters with an unread badge and email. The requester sees "Responded to your request."
- Each card has a comment thread; commenting sends email notifications to the requester and all "+1" voters.
Visibility & Promotion to Creation Plan
- Desired procedure visibility is filtered by group affiliation. Requesters and "+1" voters outside your viewing range appear as anonymous avatars.
- Floor-level requests can be promoted to the "Creation Plan" board—a top-down planning tool for managing assignments, deadlines, and progress. Notifications are sent to the original requester and "+1" voters when complete.
Posting desired procedures is available on all plans, but managing the creation plan requires Pro or above.
Training (Courses)
Dive separates "Training" and "Development" as distinct features. Training (Courses) is a system for bundling a set of procedures that must be reviewed and distributing them to target users. Simply leaving procedures available results in "view and done," but courses enable continuous management of viewing order, Assessment tests, deadlines, and progress tracking. Development (Skill Map), the other feature, is a system for evaluating, planning, and recording individual capability, described in the section below.
What you can do
- Bundle multiple procedures into one course and assign and distribute it by Group or individual.
- Course name and description can be AI-generated; procedures added to the course can come from both manual selection and AI suggestions.
- Set Sequential access (next procedure cannot be opened until the previous one is completed).
- Set strict completion criteria (completion is not granted if there are incorrect answers).
- Attach Assessment tests. Questions support both manual creation and AI-generated content; questions can be generated by AI from the procedures added to the course. Pass threshold is specified by correct answer rate (default 80%).
- Set course deadlines; learners receive automatic reminder emails at three timings—3 days before, 1 day before, and after the deadline (distributed daily at 8:00 AM Japan time).
- View progress in a grid of "target users × course procedures." Completion is green, incorrect answers are red, incomplete is yellow, and unread is gray.
- Course analytics show test takers, pass rate, Average correct rate, and Average attempts, plus per-question correct rates and answer distribution (revealing where learners Stumble).
- Solo completion Check determines whether someone completed a procedure without asking others—as one-shot completion, Stumble then completion, midway exit, no Checkpoint, or unviewed. Click to drill into Stumble details.
- Personal page shows one user in three tabs: "Training (Course enrollment)," "Development (Capability and Skill Map)," and "Usage (viewing and usage patterns)."
How to use it: Prerequisites
- Course completion is determined by "completion of all procedures in the course" and "passing the Assessment test." The Assessment test becomes available after completing all procedures and can be retaken even if failed.
- Solo completion judgment is based on answer records. Procedures without questions are marked "no Checkpoint" and cannot be assessed for Solo completion. Per-procedure question settings are a Core plan and above feature.
- Reminders go to the learner. Each of the three stages sends only once; already-completed users do not receive reminders. If the deadline is changed, it resends with the new deadline.
Limitations and notes
- Training (Courses) is available on Core plan and above; not available on Free or Lite.
| Feature | Available plan |
|---|---|
| Training management (Courses) | Core・Pro・Enterprise |
| Course Assignment / progress management | Core・Pro・Enterprise |
| Deadline reminders | Core・Pro・Enterprise |
| Per-procedure question settings | Core・Pro・Enterprise |
Development (Skill Map)
Development (Skill Map) is a feature for evaluating individual capability, creating development plans, and maintaining records. It is designed with alignment to ISO 9001 Section 7.2 "Competence" requirements.
What you can do
- Manage skills in a three-tier structure: Category, Subcategory, and skill. Over 30 Industry templates are available as a starting point. Available fields include manufacturing, care, construction, food, logistics, retail, and office.
- Define evaluation criteria for each skill across four levels (Lv1 to Lv4). Each cell in the table is "skill × Members"; a circle is divided into four sections with achievement level shaded; when Lv4 is reached, the circle is completely filled. Skills under Development appear in a different color.
- Set a target level (Level threshold) for each skill.
- Link Related procedures to a skill; when users view that procedure, the history is tallied as a Development record for the Target skills. Use "Bulk link with AI" to get suggestions and apply them in bulk after confirming.
- Two evaluation streams exist: self-assessment by the user and confirmed assessment by administrators. When self-assessment increases, administrators see "Consider Change level."
- Set staffing plan for each skill (number of people with desired skill level); the Gap column shows shortfalls when compared to Members' actual Skill Status.
- Set Development period and Development owner for each skill level. Development owners can check their "apprentices" list to see which skills need improvement and progress.
- Assessments can be saved and approved as snapshots, maintaining history. Export to Excel the Skill Map Status at that point for archiving and audit approval records. Skill Map also supports import from Excel.
- Skill Catalog enables centralized management of Team-common skill definitions. When the catalog updates, it automatically syncs to all Groups using it.
How to use it: Prerequisites
- Start by creating the skill framework From industry templates, adjust it to your company's Categories and skills, then deploy operationally.
- Viewing history accumulates as Development records only when procedures are linked to a skill. Skills without linked procedures are managed by assessment entry alone.
- Self-assessment does not automatically become confirmed assessment. Confirmed assessment is determined by administrator action.
Limitations and notes
- Skill Map is available on Core plan and above; not available on Free or Lite.
- Skill Catalog with Team-common skill definitions is Enterprise plan only.
- Skill Map Export to Excel is Pro plan and above.
| Feature | Available plan |
|---|---|
| Development management (Skill Map) | Core・Pro・Enterprise |
| Skill Map Export to Excel | Pro・Enterprise |
| Skill Catalog (Team-common skills) | Enterprise |
AR and Smart Glasses (supplementary information for video procedures)
Dive's core is video procedures that work on smartphones and web browsers. AR is not a replacement, but supplementary information overlaid on the actual equipment to clarify "where" and "which direction"—details that video and text alone may not convey clearly. Our help articles recommend using AR at key points as an auxiliary feature, not for all steps.
What you can do
- Overlay AR Objects positioned in a step on the actual equipment to show location and orientation. AR settings can be added to any type of procedure: video, image, or PDF.
- Enable or disable AR per step. Scan data for alignment can be configured or removed at the step level.
- Text placed within AR is included in cross-procedure search results.
- Since alignment uses actual site and equipment photos directly, there's no need to print AR markers. One smartphone is all you need to build AR content without a CG engineer.
Alignment methods
| Method | Details |
|---|---|
| Spatial recognition (VPS: Visual Positioning System) | Aligns to point cloud and map data scanned from the site. No markers needed. Can also integrate with Immersal map data. |
| Plane detection (markerless) | Detects and places on planes without scan data. Falls back to this method if spatial recognition is unavailable. |
Supported devices
- Smartphones and tablets: ARCore/ARKit-compatible devices (Android devices with Depth API support recommended).
- Smart glasses: monocular types (Vuzix M400 series / RealWear Navigator series) and AR glasses types (Xreal Air 2 Ultra / XREAL One, One Pro / MiRZA / dynaEdge XR1). The product site lists compatibility with 8 or more types.
- When the official XREAL Eye camera module is attached to XREAL One, you can use 6DoF, AR procedure alignment, and remote support that transmits camera feed from the glasses' perspective.
Remote support
- Video call while streaming real-time camera feed from the field smartphone or smart glasses.
- AR Objects can be placed in the other person's space, enabling communication with spatial references like "there" and "here."
- Up to 15 participants per session.
Limitations and prerequisites
- AR Objects are positioned and edited in the smartphone app on AR-compatible devices. AR editing cannot be completed in a browser alone.
- Creating AR procedures, using smart glasses apps, and alignment with point clouds require Pro plan or above. Immersal VPS integration and remote support are Enterprise plan only.
- In Lite plan, you can view already-created AR procedures but cannot edit AR.
Operations, Permissions, Reports & Exports
Permissions and Groups
- Five permission levels exist: Owner / Team Administrator / Group Administrator / Created user / View-only user. All five share the same account limit.
- You can restrict folder access by group. However, Owner / Team Administrator have privileged access and can view all folders regardless of access restrictions.
- For field staff without email addresses, you can issue Email-free accounts using Login ID (all plans).
| Plan | Number of groups | Hierarchy levels |
|---|---|---|
| Free / Lite | No groups | — |
| Core | Up to 5 groups | Up to 1 level |
| Pro | Unlimited | Up to 1 level |
| Enterprise | Unlimited | Unlimited |
Group management and folder access restriction are Core plan and above features.
Approval Workflow and Revisions
- SOPs follow a draft → approval workflow → publish cycle. Approval workflows are configured per group, and SOPs cannot be published or revised without approver sign-off.
- Revision history is retained and can be restored. Approval workflows are available on Pro plan and above.
Two Ways to Share Externally
| Item | External Share URL | Guest Invitation |
|---|---|---|
| Login required | No | Yes |
| Sharing unit | Per SOP | Per folder |
| Account quota | Does not consume quota | Separate quota up to 3x account limit (e.g., Pro = 50 accounts / 150 guest invites) |
| Capabilities | View only | Information sharing only (view-only; excluded from Usage Report and training metrics) |
| Controls | Expiration, password, audit log | Audit log, IP address restriction, SSO |
| Plan availability | Lite and above | Core and above (not available on Free / Lite) |
- When issuing an External Share URL, either an expiration date or password is mandatory; permanent URLs without expiration or password cannot be issued (if no expiration is desired, password protection is required).
- Issued External Share URLs can have expiration dates changed, passwords rotated, revoked, or reactivated. All issuance, revocation, reactivation, expiration changes, and password changes are logged in the audit log.
- The External Share URL feature is disabled by default for security reasons and must be enabled by the Owner in team settings.
- With Guest Invitation, SOPs appear as a flat list rather than the folder hierarchy, keeping your internal organization structure (folder names, order) hidden from external users.
- Our help documentation recommends Guest Invitation as the security best practice for external sharing.
Usage Report and ROI
- Four-tab structure: [View & Usage] / [Stumbling Points] / [Draft Status] / [Economic Impact / ROI]. Filter and aggregate by date range and group. The [Economic Impact / ROI] tab is visible only to Owner / Team Administrator.
- View counts, user counts, procedures referenced, views per procedure, creation time, publish count and more can be reviewed by group / member / SOP. Date ranges include all-time, last week, 1 month, 3 months, 6 months, 1 year.
- We define "Solo completion" as viewing an SOP to completion and submitting a completion report, and "Solo completion rate" as the ratio who successfully passed all checkpoints without errors on first attempt.
- Stumbling Point map shows which checkpoints caused issues, how many users, and which answers they selected.
- The entire report (overview / Group / User axis / SOP axis / daily trend / draft status / Economic Impact / ROI) can be exported as a single Excel file.
- Usage Report can auto-generate a business proposal (PowerPoint) for internal approval. Supports both actual-performance versions and trial versions with estimated figures (placeholder values).
- Usage Report is available on all plans including Free.
Economic Impact is calculated by summing three components. Underlying time and cost assumptions can be adjusted on-screen to match your situation (default average labor cost: ¥3,000/hour).
| Component | Calculation |
|---|---|
| A. Initial training replacement | First-view count × default 30 minutes |
| B. Question replacement | Re-view count × default 15 minutes (self-verification instead of asking a senior) |
| C. Creation efficiency gain | Publish count × (traditional 90 min − Dive 15 min) |
Results display as total impact hours (h) / total impact amount (¥) / ROI multiple vs. current plan monthly fee / Payback period (days). A free ROI calculator using the same formula is publicly available without login.
Work Analysis (IE Time Analysis)
- Automatically decompose video into work elements and calculate standard times and ratings. Create improvement proposals from an ECRS perspective and numerically compare before and after.
- View expert model video and trainee work side-by-side in synchronized playback for comparison.
- Time analysis, improvement proposals, materialization of improved video, and comparison with target are Pro plan and above. Per-step inspection and pass/fail recording are Core plan and above.
Exports
- SOP export formats: PDF, Excel, PowerPoint, and video (MP4). MP4 can be exported per language (translated captions and audio embedded in language-specific outputs).
- PDF export supports template switching (simple default template plus custom templates created by your team).
- Exportable Excel data: SOP / SOPs list in folder / Skill Map / individual user view history / members and groups list.
- CSV export is available for view records (inspection table) and audit log. UTF-8 with BOM prevents character corruption when opened in Excel.
| Export | Available on Plans |
|---|---|
| MP4 export | Lite and above |
| PDF export / Excel export | Core and above |
| PowerPoint export, SOPs list export, Skill Map export, view records (inspection table CSV) export | Pro and above |
Other Operations Features
- Screen recording support.
- Combine video, images, and PDFs step-by-step into a single SOP.
- AI text-to-speech, voice commands (next / play / pause).
- Multi-level folder management, revision history and restore, approval workflow.
Limitations
- SOPs with conditional branching do not support MP4 export.
- Owner / Team Administrator can access all folders regardless of folder access restrictions (exempt from restrictions).
Security
Third-Party Certification and External Assessment
- Obtained ISO/IEC 27001:2022 (ISMS) international information security standard in November 2024. Certification body: G-CERTi, Certification number GIJP-1469-IC, Accreditation: IAS (MSCB-113).
- Responded to evaluation by external security assessment platform Assured and achieved a score of 92.4, placing us in the top 25% compared to the overall average of 76.5 (as of response date).
- We provide a security checklist that complies with the Ministry of Economy, Trade and Industry's "Cloud Service Level Checklist."
Data Handling
- All communications are encrypted with HTTPS. Both databases and files are encrypted, and storage is located in-region (Tokyo, Japan).
- The video analysis AI is designed not to retrain. We do not retrain or access customer data internally. You can apply AI analysis directly to highly confidential work procedures.
- Server monitoring 24/7/365. Server access is limited to system operations staff, and operations staff do not access customer data in principle (except upon customer request or for incident response).
- Uptime is 99.9% based on the past year's performance.
Authentication and Access Control
- Single sign-on supports two methods: SAML 2.0 (Okta, Azure AD, OneLogin, HENNGE One, etc.) and Microsoft Entra ID (OIDC). Prior registration with IdP is required. The option to "Disallow logins other than SSO" is also available (Enterprise plan).
- Two-factor authentication supports TOTP (Google Authenticator, Microsoft Authenticator, etc.). Mandatory enforcement for the entire team is available on the Enterprise plan, with options to specify exception groups and exception users. SSO users authenticate through their IdP and are excluded from mandatory enforcement.
- Social login (Google, Facebook, etc.) is not provided.
- IP address restriction allows you to register global IPs (IPv4/IPv6) in three formats: single, range, or CIDR. The allowlist supports Excel import and Excel export.
Control and Audit
- The audit log records logins (success, failure, MFA, SSO), user additions/deletions/permission changes, Security Settings changes, external share URL issuance/invalidation/restoration/expiration changes, and more. Logs can be downloaded as CSV. Users with Owner permissions can retrieve access logs (login history, user and group change history, etc.) within their team.
- Role-based feature usage permissions (Enterprise plan). You can set role thresholds for specific features: Procedure file export / Folder Excel export / External sharing / Media download / Download attachments / Skill map export / Usage report export.
- Device cache settings (Enterprise plan). Switch to a mode that leaves no viewed data on the device. With this setting, offline playback and prefetch are disabled.
- External share URL feature is disabled by default. It cannot be used unless the Owner enables it in team settings.
- Security Settings (IP address restriction, SSO, mandatory two-factor authentication, password policy, external sharing prohibition, feature-based role thresholds, device cache, audit logs) can only be changed by users with Owner permissions.
Security Features by Plan
| Feature | Available Plans |
|---|---|
| Two-factor authentication, IP address restriction, Audit logs | All plans including Free |
| External sharing prohibition setting | Lite and above |
| Mandatory two-factor authentication, Single sign-on, Flexible download restriction settings (role thresholds by feature), Device cache settings | Enterprise only |
Implementation Track Record
Dive is used by publicly listed companies and universities with stringent security requirements. Our clients include IHI Group, Denso, Kaneka, Nippon Steel Group, JAL Group, Unitika, SG Holdings, Actio, Sotetsu Holdings, National Center for Global Health and Medicine, Tokyo University of Technology, and Saga University.