Analyze large compliance documents
Upload policies and evidence packs in chat. Modes, plan access, and monthly upload fair use match current product limits.
ISMS Copilot can analyze multi-page policies, evidence packs, and control matrices in a single conversation. Use Think for deeper single-shot reasoning, Beyond for multi-step plan/draft/verify work, and keep Fast for shorter Q&A. Provider routing depends on plan and Advanced Data Protection (see Chat modes); do not assume one vendor brand name for every request.
Modes for large analysis
| Mode | When it helps |
|---|---|
| Fast | Quick questions over a few documents |
| Think | Harder single responses over long packs |
| Beyond | Multi-step deliverables (gap report, plan, then draft) |
- Free: Fast only. Think / Beyond need Plus+ or an active Plus trial.
- Essential (grandfathered): Fast + Think. Beyond / web research are Plus+ features.
- Plus and above: Fast, Think, Beyond, and web research (subject to ADP and usage windows).
Toggle the mode in the chat composer before you send. If Think or Beyond is unavailable, the app shows an upgrade or trial path for your plan.
Long Fast and Think threads may compact older context so the conversation can continue. That is a product reliability feature, not a separate vendor-only mode. Details: Think mode context compaction.
Document upload limits
You can upload multiple files in a single conversation (typical batch cap: 10 files). Supported formats and per-file size limits are enforced in the app (PDF/Office vs text-style files).
Monthly completed uploads (fair use):
| Plan | Uploads / month |
|---|---|
| Free | 10 |
| Every paid plan (Essential, Plus, Standard, Pro, Business) | 500 fair use |
Allowance resets on the product monthly schedule (UTC). Failed or stuck processing jobs are not meant to permanently burn quota the same way completed uploads do. Full matrix: Subscription plans and pricing.
If a PDF exceeds the app size limit, split it before uploading. Workspace-pinned reference files follow workspace rules and are not the same bucket as casual chat uploads.
Context and conversation size
Document text, prior messages, and skills all share the model context window. Practical tips:
- Prefer a fresh conversation when analyzing a large pack so history does not consume the window first.
- Be specific in prompts (framework, clause range, output shape) rather than "analyze these documents."
- For multi-client work, use workspaces so context stays isolated.
- If the product reports the conversation is too long, start a new thread or drop unneeded files; see Conversation too long.
Exact token ceilings vary by route and mode. Treat in-app errors as authoritative for a given request.
Best practices
Upload in a focused thread. Maximize tokens for the documents themselves rather than prior chat.
Use workspaces. Separate clients and programs so evidence packs do not mix.
Export when done. Markdown or DOCX export is useful for audit packs and gap reports.
Verify against the standard. AI accelerates drafting and review; it does not replace professional judgment or the official framework text.
Common use cases
- Gap analysis: Map policies to ISO 27001, SOC 2, or NIST CSF and list missing controls
- Audit preparation: Summarize evidence coverage by control area
- Policy review: Compare versions for inconsistencies or stale language
- Risk assessment: Work across registers and supporting docs in one thread