From trusted sourceto cited decision artifact
Six stages. One pipeline. Source map → claim engine → workflow engine → draft artifact → reviewer loop → decision pack. Every claim traceable to its source.
Track the workflow from source map to decision pack
Six engines. Every claim cited. Every reviewer action logged.
Source map
Map sources and universe
Claim engine
Map claims to evidence
Workflow engine
Apply your methodology
Draft output
Generate cited workpaper
Reviewer loop
Human validates
Strategy pack
Export shortlist · brief · board pack
Source map
Trusted inputs go in: filings, transcripts, broker research, internal models, regulatory data. Every source is registered, hashed, and verified before it can be cited.
- Start with the source types your team already trusts
- Provenance and version pinning
- Private workspace ingest
- Cross-source deduplication
Claim engine
Each assertion is mapped to the source span it came from. Unsupported claims surface for the reviewer before the draft ships.
- Claim-level evidence linking
- Page and timestamp citation
- Contradiction detection
- Confidence + assumption markers
Workflow engine
Your IC framework, scoring rubric, and red-flag library run as a workflow — not a prompt. Same questions, same depth, every diligence.
- Configurable methodology
- Sector-specific templates
- Red-flag library
- Reusable scoring rubrics
Draft artifact
Memo, brief, sector note, or board pack — drafted with every claim citation-linked. No fabrication, no untraceable paragraphs.
- IC memo + risk register
- Executive summary + appendix
- Inline citation rendering
- Editable in your tools
Reviewer loop
Your analysts review claim-by-claim. Approve, flag, or rewrite — every action recorded against the evidence span.
- Claim-by-claim sign-off
- Reviewer state machine
- Comment threads on evidence
- Role-based approval gates
Decision pack
Decision-ready output exported. Every paragraph traceable to its source. The audit trail ships with the file.
- PDF, DOCX, structured JSON
- Tamper-evident audit log
- Reviewer attribution
- Compliance-ready export
What ships in every decision pack
Memo, evidence, audit trail, exports — every paragraph traceable to its source.
EBITDA margin is sustainable through FY27, supported by pricing discipline.
Conf · medium- FY24 10-K[p.86]“Pricing actions offset 240bps of input-cost pressure.”
- Q3 transcript[00:18:42]“Inventory normalization is a 2H dependency.”
| Claim | Source | Value | Conf. |
|---|---|---|---|
| Margin durability | Broker consensus | Mixed | Conf · medium |
| Regulatory tail | EU CBAM | Elevated | Conf · medium |
| Capex trajectory | 10-K filings | +18% YoY | Conf · high |
- Approved
J. Singh
Senior Analyst
- Flagged
M. Patel
Sector Lead
“Needs second broker source on margin durability.”
- Pending
R. Chen
Investment Committee
Decision artifact
- IC memo · brief · board pack
- Cited executive summary
- Inline evidence rendering
- Editable in your tools
Evidence appendix
- Source documents (PDF, DOCX)
- Page-level citation index
- Transcript timestamps
- Internal-note references
Workflow output
- Risk register + red flags
- Peer benchmark set
- Scoring rubric output
- Scenario / sensitivity
Audit trail
- Reviewer attribution log
- Tamper-evident actions
- Source-diff over time
- Compliance-ready export
Private Equity Firm Evaluates $500M Acquisition in 48 Hours
Competitive bid deadline in 3 days
Full DD completed in 48 hours
Won bid, 3.5x return in 2 years
“15Rock’s 48-hour turnaround was the difference between winning and losing this deal. The analysis depth matched what we’d expect from a 6-week engagement.”
From trusted sources to cited strategy workpapers
Six stages. Every claim mapped to evidence. Every reviewer action audit-trailed.
Ingest sources
Public data, private overlays, internal models, transcripts, regulatory data — every input is registered, hashed, and verified before it can be cited.
- BloombergNEF — Steel 2030Q3 2025
- IEA WEO 2025Oct 2025
- Form 10-K (12 filings)FY24
- Internal model — capex base casev3.2
- Earnings transcripts (8)Q1–Q3
5 of 5 sources verified · ready for claim mapping
Map claims to evidence
Every assertion in a draft is linked back to the source span it came from. Unsupported claims are flagged before they reach a reviewer.
Steel-sector capex is set to rise 18% YoY through 2027, driven by green-premium product lines.
Conf · high- BNEF Steel 2030[p.42]“Greenfield DRI-EAF capex commitments up 18.4% YoY, concentrated in EU and India.”
- ArcelorMittal 10-K[FY24]“Capex guidance raised to $4.5–5.0B, weighted to decarb.”
- IEA WEO 2025[ch.6]“Green-premium product demand grows at 11% CAGR through 2030.”
Apply your methodology
Your scoring rubric, peer logic, value-creation lens, and deliverable standard run as a workflow — not a prompt. Same questions, same depth, every engagement.
| Claim | Source | Value | Conf. |
|---|---|---|---|
| Market growth | BNEF + IEA | +11% CAGR | Conf · high |
| Capex trajectory | 10-K filings | +18% YoY | Conf · high |
| Margin durability | Broker consensus | Mixed | Conf · medium |
| Regulatory tail risk | EU CBAM | Elevated | Conf · medium |
Generate the draft
Shortlist, Value Scan, benchmark, sector note, or board pack — drafted with every claim citation-linked. No fabrication, no untraceable paragraphs.
Value Scan — Project Aurora
Exportable · Cited · Audit-trailed
- Executive summarycomplete
- Shortlist rationale & key claims (12)complete
- Evidence appendix (47 citations)complete
- Risk registerin-review
- Reviewer sign-offpending
Human validates
Your analysts review claim-by-claim. Approve, flag, or rewrite — every action is recorded against the underlying evidence span.
- Approved
J. Singh
Senior Analyst
“Capex thesis well-supported — green-premium evidence is the strongest.”
- Flagged
M. Patel
Sector Lead
“Margin durability claim needs a second broker source.”
- Pending
R. Chen
Investment Committee
Export the artifact
Decision-ready output: shortlist, brief, benchmark, or board pack — every paragraph traceable to its source. The audit trail ships with the file.
Claim engine linked 47 citations to 12 claims
08:14 · automated
J. Singh approved sections 1–3
10:42 · reviewer
M. Patel flagged margin durability for re-evidence
11:08 · reviewer
Workflow engine regenerated section 4 with new source
11:31 · automated
R. Chen exported decision pack (PDF + DOCX)
14:20 · reviewer
How we talk about the workflow
Six terms you'll see across the product. Each maps to a concrete artifact, not a marketing concept.
Source map
Every trusted input — public data, private overlays, internal models, regulatory data — registered and verified before it can be cited.
Claim engine
Maps each assertion in a draft to the evidence span it came from. Unsupported claims surface before review.
Workflow engine
Your methodology — scoring rubric, peer logic, value-creation lens, and deliverable standard — running as a repeatable workflow, not a prompt.
Reviewer loop
Analyst review at the claim level. Approve, flag, or rewrite — every action recorded against the evidence.
Strategy pack
The output: shortlist, Value Scan, benchmark, brief, or board pack — every paragraph traceable to its source.
Audit trail
Every source ingestion, claim mapping, reviewer action, and export — captured and exportable with the artifact.
Process FAQs
What a 2-week prototype includes
When the 48-hour pilot ships and you choose to go deeper, the next two weeks turn one workflow into a working artifact.
Workflow scoping
Which workflow, why, and what success looks like.
Source map
Which inputs feed it, the gaps, and the ingestion plan.
Draft artifact structure
The shape of what gets shipped — memo, register, or scorecard.
Review loop
Analyst, sector lead, and IC sign-off built into the artifact.
Pilot readout
What worked, what didn't — every claim with its citation.
Go/no-go recommendation
An honest scope for production, or the case to stop here.
Start your 48-hour analysis today
Join leading institutions getting answers in hours, not weeks.
No generic demo. We'll scope one real workflow.
Start small
Most teams begin with one workflow, one source set, and one output.