TSM.AI Pre-auditAI pre-audit for Towards Sustainable Mining
Submit your operation's policies, plans, procedures and reports — in any language. TSM.AI evaluates the submission against all 31 indicators across the 9 TSM protocols of the Mining Association of Canada, producing verdicts on the C/B/A/AA/AAA scale (or Yes/No for the two binary protocols), backed by verbatim evidence, gaps and concrete recommendations.
From documents to a TSM-readiness verdict
A three-step pipeline that mirrors TSM's level-by-level methodology faithfully.
Upload to the document library
Drop in any number of documents — PDF, DOCX, XLSX, PPTX (up to 50 MB each). On upload, full text is extracted, a 1–2 sentence summary is generated, 3–8 topic tags are assigned, and the document is classified against the 31 indicators (returning indicator codes ranked by relevance).
Run an assessment
Name the operation (defaults to "North Pine Mining"), choose which documents to include (default: all ready), choose the output language for AI-generated text, and click Begin Assessment. The system processes 31 indicators in batches of 6 concurrent evaluations, streaming verdicts as they resolve.
Read the formal report
Editorial-style printable report with hero photograph on the cover, protocol-by-protocol deep dive, verbatim evidence, gaps and recommendations. Export by Print to PDF in the browser. Server-side PDF generation is on the roadmap.
Two categories. Nine protocols. Thirty-one indicators.
The TSM framework spans the full operational footprint of a Canadian-style mining operation, from Indigenous and Community Relationships to Tailings, Water, Biodiversity and Climate Change.
Five protocols on social licence
- ICR Indigenous & Community Relationships 4 ind.
- CMD Crisis Management & Communications 3 ind.
- SHS Safety & Health 5 ind.
- PRD Preventing Child & Forced Labour 2 ind.
- EDI Equitable, Diverse & Inclusive Workplaces 3 ind.
Four protocols on stewardship
- TMG Tailings Management 5 ind.
- WAT Water Stewardship 3 ind.
- BIO Biodiversity Conservation Management 3 ind.
- CCH Climate Change 3 ind. · binary
Energy & GHG Emissions plus Climate Change are scored Yes/No for the binary protocol indicators.
The TSM C / B / A / AA / AAA ladder
For each indicator, the evaluator is given the verbatim criteria text at every level and decides which level the operator most plausibly meets given the evidence. Level A is TSM's stated goal.
No system in place
Below the entry threshold for the indicator's expected practice.
Procedures established
Procedures exist but are not formalised at the management-system level.
System in place
Documented management system aligned to TSM expectations.
TSM goalIntegrated
System is integrated into operations and reviewed periodically.
Excellence & leadership
Demonstrated excellence; leadership in the protocol's domain.
Yes / No verdicts for binary protocols
Two protocols use a binary verdict instead of a level: Energy Use & GHG Emissions and Climate Change. For these, TSM.AI returns YES or NO against the criterion text, with the same evidence-attribution rigour.
Insufficient Evidence — visually distinct from C
If no documents in the assessment's set are tagged as relevant to an indicator, TSM.AI records INSUFFICIENT_EVIDENCE automatically without an LLM call. The rationale explains what kinds of documents would be needed; the gaps array lists the expected-but-absent requirements. It is not a failure — it is an absence.
There is no single pass / fail. There is a posture.
TSM does not yield a binary outcome. The roll-up captures the operation's TSM-readiness posture across all 31 indicators.
TSM-readiness roll-up
Each indicator at level A or above (or YES) counts as at-goal. Indicators below A (or NO) count as below-goal. Insufficient Evidence is reported separately and does not count as below-goal.
Illustrative numbers from the demo. Computed once when the assessment completes and stored on the assessment row.
Every indicator produces four structured outputs
No paraphrased fluff. Direct quotes from your documents, attributed to the source, with verifiable gaps cited to specific criteria language and one-to-one recommendations.
Level letter (or Yes/No) + rationale
One of C / B / A / AA / AAA (or YES / NO for binary protocols), plus 2–3 sentences citing the specific criteria language that supports or fails to support each higher level.
Verbatim quotes, attributed
Verbatim quotes from your documents that support the verdict — no paraphrase. Each is tagged with the source document name.
Specific missing requirements
Specific missing requirements drawn from the criteria text — not generic observations. Cite the exact language from the indicator that is not met.
Concrete actions, mapped to gaps
Concrete actions that map one-to-one to the gaps. Owner-friendly, scoped to a sprint or quarter, ready to brief into your next TSM cycle.
Pre-loaded with North Pine Mining
When you open the app, it is pre-populated with a fictitious Canadian mining operation called North Pine Mining — a library of plausible corporate documents and a pre-computed assessment with all 31 indicators resolved.
The operation
North Pine Mining — a fictitious Canadian boreal mining operation used to showcase real-world structure across the 9 TSM protocols and Indigenous & community context.
Plausible document library
Indigenous engagement framework, community investment policy, crisis management plan, OHSMS, EDI strategy, child & forced labour due diligence, tailings management standard, water management plan, biodiversity action plan, climate strategy & GHG inventory.
Pre-computed assessment
All 31 indicators resolved. Mix of A and AA verdicts in well-evidenced areas, transitional B in mid-implementation areas, a couple of Cs flagging immediate action, plus YES/NO on the binary climate indicators.
Inspect the existing assessment, then upload your own documents and run a fresh evaluation against the live LLM pipeline. New documents and assessments persist across sessions and are visible to anyone who opens the app.
A live assessment dashboard and a printable formal report
Watch the assessment run indicator by indicator. Then read the editorial-style report with hero photograph on the cover and protocol-by-protocol deep dive.
Assessment dashboard
Three columns: a left rail with the indicator list grouped by protocol (each row showing its tier badge, code, name).
Center detail panel showing the selected indicator's findings as an editorial article.
Right coverage matrix showing which documents support which indicators.
Live progress: the frontend polls every 3 seconds while the assessment runs, ~6 indicators flip from skeleton to resolved every minute.
Formal report
Editorial typography with hero photograph on the cover and page-break controls.
Protocol-by-protocol deep dive with verbatim evidence, gaps and recommendations.
Print to PDF via the browser. Server-side PDF generation is on the roadmap.
Disclaimer stating determinations are advisory and do not constitute a verified TSM self-assessment or external verification.
Common questions
Is TSM.AI affiliated with the Mining Association of Canada?
No. TSM.AI is an independent pre-audit instrument built by Data Riders. Not affiliated with the Mining Association of Canada (MAC) or with the TSM program. Reports include a clear disclaimer.
What does the assessment cover?
All 31 indicators across the 9 TSM protocols: Indigenous & Community Relationships, Crisis Management & Communications, Safety & Health, Preventing Child & Forced Labour, Equitable Diverse & Inclusive Workplaces, Tailings Management, Water Stewardship, Biodiversity Conservation, and Climate Change. The C/B/A/AA/AAA scale applies to seven protocols; Yes/No applies to the two binary protocols.
What languages are supported?
Documents can be submitted in any language. The interface ships in English and Portuguese. LLM-generated rationale, gaps and recommendations adapt to the user's selected output language. Other UI languages are roadmap.
How are levels determined?
For each indicator, the evaluator is given the indicator's code/name/purpose, the verbatim criteria text per level, and the full text of every document tagged as relevant during ingestion (truncated to 15,000 characters each). The model picks the level the operator most plausibly meets given the evidence, citing the specific criteria language that supports or fails to support each higher level.
What is the overall outcome?
There is no single pass/fail in TSM. The roll-up reports four numbers: at-goal count (indicators at A or higher, or Yes), below-goal count (indicators below A, or No), insufficient-evidence count, and overall TSM-readiness percentage (at-goal / 31). These are computed once when the assessment completes and stored.
How is Insufficient Evidence different from a C verdict?
A C verdict means the evidence shows the operator is below the entry threshold. Insufficient Evidence means no documents in the assessment's set were tagged as relevant to that indicator at all — it is an absence of evidence, not evidence of failure. The rationale explains what kinds of documents would be needed and recommends documents to produce.
Do I need to log in?
No. TSM.AI runs in demo-mode by design: no authentication, single shared workspace. Documents and assessments persist across sessions and are visible to anyone who opens the app. Authentication and multi-tenancy are roadmap.
Does the AI replace a real TSM verification?
No. TSM.AI is a pre-audit instrument that identifies gaps and surfaces evidence before a TSM self-assessment or external verification. Determinations are advisory and do not constitute a verified TSM self-assessment or external verification.
What file types are supported?
PDF, DOCX, XLSX and PPTX, up to 50 MB each. On upload, the document is auto-summarised, tagged with topics and classified against the 31 indicators. Full text is preserved for the per-indicator evaluation step (which truncates to 15,000 characters per document).
Other agents in the compliance family
Run your first TSM pre-audit today
No procurement. No login. Open the assessor, inspect the North Pine Mining demo case, then upload your own documents and run a fresh evaluation against the live pipeline.
TSM.AI is an independent pre-audit instrument built by Data Riders. Not affiliated with the Mining Association of Canada (MAC) or with the TSM program.