Your AI Now Understands CPM
LLMs can't calculate a critical path. They hallucinate schedule analysis. FPM gives any AI deterministic forensic math through the Model Context Protocol.
AI Is Powerful but Unreliable for Schedules
Ask an LLM to analyze a 10,000 activity P6 schedule and you'll get confident, articulate, and completely wrong answers.
Can't Calculate CPM
Forward/backward pass, float analysis, longest path require exact computation, not prediction.
Hallucinates Causation
Concurrent delay requires measuring each change through the dependency network. LLMs guess at causation.
Not Auditable
Generated answers can't be traced to source data. Try defending that in arbitration.
The Forensic Math Engine Behind Your AI
FPM doesn't replace your AI. It gives your AI a source of truth. Ask questions in plain English, get answers backed by real calculations.
Your AI
Claude, GPT, Gemini, or your enterprise LLM
Prompts stay here
Forensic Math
Schedule analysis, delay attribution, concurrent measurement
Deterministic results
Upload your XER
Import P6 schedules through the FPM web portal. The engine recalculates every snapshot as source of truth.
Connect your AI
Point Claude Desktop, VS Code Copilot, or any MCP-compatible client at your FPM server.
Ask questions in plain English
"Why is my project late?" Your AI calls the right tools, backed by deterministic math.
18+ Forensic Tools, One Protocol
Your AI gets structured access to FPM's forensic engine, from high-level discovery down to individual relationship changes.
Discovery
Browse schedules, search activities by code or name, filter by status and critical path. Get project-level metadata and calendar configurations.
CPM Analysis
Dates, float, taint status, constraints — with base vs. revised variance calculations. Trace driving and near-critical paths through the network.
Delay Attribution
Milestone slip magnitude and delay drivers by category: Progress, Duration, Logic, Structural. Window-by-window forensic analysis.
Change Tracking
Field-level changes, logic modifications, and full activity evolution across every schedule update. The audit trail for litigation.
As-Planned vs As-Built
WOET execution analysis: day-by-day Progress/Gain/Extended Duration/Void classification, activity-level history, timeline detail, data quality checks, and reconciliation audit.
From Upload to Litigation Strategy in Minutes
A forensic analysis session: four questions that uncover a $14M delay story.
This is a 3-year schedule history for a highway interchange reconstruction:
12 updates — March 2019 through June 2022
342 → 491 activities — 44% scope growth over the project life
Original contract: NTP Mar 2019, completion Apr 2021 (760 calendar days)
Current projected: Oct 2022 — 18 months past original completion
The schedule grew substantially and the project is well past its original completion. Let me run the delay analysis across all 11 windows.
The project was never on track. Here's the milestone trajectory for Substantial Completion:
| Update | Data Date | Projected | Slip |
|---|---|---|---|
| Baseline | Mar 2019 | Apr 2021 | — |
| Update 3 | Sep 2019 | Jul 2021 | +98d |
| Update 6 | Aug 2020 | Nov 2021 | +214d |
| Update 9 | Apr 2021 | Mar 2022 | +338d |
| Update 12 | Jun 2022 | Oct 2022 | +542d |
542 days of slip — 18 months beyond the original completion. The delay breaks down into three major categories:
Environmental permits (~180d): Wetland mitigation and stormwater basin permits held the critical path for over a year
ITS fiber optic submittal (~140d): A single submittal-review-resubmit cycle paralyzed the ITS pathway for 10 months
DOT scope additions (~130d): Auxiliary lane, ramp reconfiguration, and retaining wall additions came in waves across Updates 4–9
Categories overlap: concurrent delays on parallel paths mean the sum exceeds total slip. This is expected in forensic analysis.
This is the strongest finding in the schedule. Two environmental permits — SR18-ENV-210 (Wetland Mitigation) and SR18-ENV-215 (Stormwater Basin Permit) — tell a remarkable story:
SR18-ENV-215 follows the same pattern: added at baseline, duration doubled, constraint pushed twice, then deleted at Update 10.
104 combined impact-days of owner-caused delay on permits that shouldn't have existed
Both activities were on the driving path to Substantial Completion for their entire lifespan
The owner's environmental consultant added them "as a precaution" — they were never required by the Army Corps permit
Here's the responsibility breakdown across 542 days of total delay:
Three key takeaways for litigation:
The phantom permits are your strongest finding. 104 impact-days of critical-path delay on activities that were deleted — never needed. This is a clean owner-liability story with no concurrent contractor delay to muddy it.
The project was behind from the start. By Update 3 (Sep 2019), the schedule had already slipped 98 days. Every subsequent update added more delay. There's no inflection point where the contractor could have recovered.
The ITS submittal is a shared-liability landmine. The 10-month submittal cycle has fingerprints on both sides: slow owner review, but the contractor's initial submittal was rejected twice. Expect this to be the most contested item in arbitration.
All figures are from FPM's analysis engine. Deterministic, auditable, and traceable to the XER source data across all 12 schedule snapshots.
Every number comes from FPM's forensic engine. Calculated, not generated.
Built for Trust
When the stakes are $60M and a year of arbitration, you need guarantees, not probabilities.
Your prompts stay local
Only narrow, specific data queries reach FPM. Your reasoning, strategy, and analysis context never leave your machine.
Deterministic, not generative
Every number returned is calculated by the forensic engine. Auditable, reproducible, and traceable to source data. No generation, no hallucination.
Bring your own AI or use ours
Connect Claude Desktop, VS Code, or any MCP-compatible client. FPM also includes a built-in conversational agent with session persistence and streaming responses.
Compliance-ready
Every tool invocation is audit-logged. Results meet AACE RP 29R-03 standards. Every finding links back to the schedule data that produced it.
Give Your AI a Forensic Engine
Connect your AI to litigation-grade forensic schedule analysis. 18+ tools, deterministic results, full audit trail.
Works with Claude Desktop, VS Code, and any MCP-compatible client