LearnCompareSolutionsPricingArchitectureFAQAboutSign in
SCHEDULE-RISK PLATFORM FOR R&D PROGRAMS

Give your board a date you can defend.

CritPath AI runs real schedule math over your actual plan — so you know how solid every date is, see trouble weeks early, and the plan reschedules itself when a vendor slips. Built for biotech, deep-tech, and federally funded R&D.

$10/user/month — every feature. AI employees metered at cost + margin. No credits.

Built for Biotech program leaders · Deep-tech founders / CTOs · Federally-funded PIs · Grant program managers

CritPath AI Plan view: live Gantt with the critical path highlighted, a Decision Gates card with a GO and a PENDING gate, and the Monte Carlo completion distribution with P50–P95 percentiles

A vendor slips. Your plan re-plans itself.

The re-plan that used to cost your PMO a day happens before the meeting ends.

65 seconds, no cuts: a tox-study risk is marked Occurred → the schedule re-plans retroactively → the buffer fever chart and the P90 date move.

1
Log what happened

A CMC vendor slips a month. A risk you logged just hit its trigger. Mark it occurred — that's all the input CritPath needs.

2
The whole program reschedules

Every downstream task, buffer, and milestone re-cascades in seconds. The engine re-runs the full schedule math automatically — nothing is hand-dragged.

3
See where the plan stands now

New dates on the Gantt, buffers re-checked — and on projects running Monte Carlo, a fresh read on how solid the new finish is. Ready for the meeting you're already late for.

Sound familiar?

Dates nobody can defend

Your Gantt shows one date and says nothing about how likely it is — so every board deck is a promise you can't back. CritPath shows how solid each date is and which tasks are quietly driving the risk.

One slip, a week of re-planning

A CMC vendor slips a month and every downstream milestone silently moves. Re-planning by hand costs your PMO a day each time. Here, you log the slip and the plan catches up in seconds.

Reporting is a second job

One program PM spent eight hours a day re-typing emailed status updates. Your status report already exists — it's the live schedule. Export it, or have an AI employee draft it, every number computed.

The engine R&D actually needs

Built from the first line for the uncertainty and decision density of research programs — not retrofitted from construction or IT delivery.

Know how solid every date is

Monte Carlo simulation — 10,000 runs over your actual dependency network — turns a single guessed date into a defensible range and the real chance of hitting your target. Your engineers already estimate in ranges; here the range survives all the way to the board deck.

Find what's eating your finish date

Critical path plus critical chain (CCPM): the bottleneck, the buffers protecting your milestones, and a fever chart that shows margin burning before the date moves.

Decisions that live in the plan

Go / No-Go / Pivot gates are schedule objects with named owners, co-signs, and an audit trail — and a No-Go or Pivot triggers a fresh run of the schedule math.

Staff it to finish sooner

Resource leveling plus a staffing optimizer that shows what each hire or reassignment buys you — the speed-versus-cost trade-off, priced in days and dollars.

AI employees, governed

Configurable agents read the live schedule and propose updates for your review. Nothing touches the plan until a named human accepts — every acceptance audited.

Regulatory-grade rigor

AACE RP 132R-23 Level 4 risk-driven scheduling with a tamper-evident audit trail — built for pharma IND programs and federally funded R&D. 21 CFR Part 11 on the Enterprise roadmap.

AI employees. Proposes. You approve.

Build AI employees — an Estimator, a Risk Analyst, a sub-PM — with skills you define. They read your live schedule, run grounded analysis, and return typed proposals — you review the proposed change, not pages of prose, and accepting takes seconds. Nothing lands until a named human accepts; every acceptance is audited, and when you enable the adversarial reviewer on a persona, a separate, isolated review pass — blind to the generation context — grades every run against your own quality rubrics, and blocking failures prevent acceptance. Autonomous AI-employee runs activate at general availability.

Billed per completed job at its estimated token cost plus a stated margin — no credits, no black-box meter. An AI employee holds no seat and no subscription of its own: you pay for the work it actually finishes, and nothing when it is idle. See how AI usage is billed.

  • Decision gates as first-class schedule objects — owned, co-signed, audited — no PMO tool ships this
  • CPM + CCPM/TOC + Monte Carlo in one engine — incumbents have at most one
  • Every AI-employee change lands as a human-approved, audited proposal against the plan
  • AI usage metered per completed job at estimated cost + a stated margin — no credits, no seats, no black-box meter
  • WSJF + Cost of Delay wired into the simulation and the staffing optimizer
  • TOC Thinking Tools — Goal Trees, Conflict Clouds — generated from the live schedule

The gap no one else fills

Risk-math depth →Modern, self-serve UX →
The empty quadrant

Legacy risk tools

Primavera · Deltek · Pertmaster — ~$10K+/seat, Windows desktop

Spreadsheets & trackers

manual Gantts — no probabilistic math

Generic PM

Monday · Asana · Smartsheet · Linear — $7–25/mo

CritPath AI

Real risk math + modern UX + a schedule-aware copilot

$10/user/mo · all features

Schedule-risk tools have the math but not the UX. Generic PM tools have the UX but not the math. CritPath AI is the only one in both.

How we compare

CapabilityCritPath AIGeneric PMLegacy risk toolsAI-employee tools
Monte Carlo + CCPM risk math
Risk analysis & monitoring
Decision gates + retroactive reschedule
Project analysis & reporting (EVM)partialpartial
Schedule-aware AI copilotpartial
AI employees (configurable agents)partial
Human approval + audit on every AI changepartial
AI project charter → drives the schedule + copilot
Speed-vs-cost tradeoff analysispartial
Modern web UX, self-serve
R&D-native (biotech / deep-tech / federal)
Integrations (GitHub; Slack / Notion / Discord soon)
Entry price / user / mo$10$7–$25~$10K+/seat$50–$894/mo + credits

The AI-employee workspace and GitHub integration activate at general availability; Slack, Notion, and Discord integrations are on the near-term roadmap. “AI-employee tools” reflects Viktor, Motion, Lindy, and Relevance AI as verified July 2026 — genuinely autonomous, and none ships a scheduling engine.

Rigor you can check

No borrowed logos, no invented percentages — claims you can verify in the product and the literature.

10,000-run simulations, in seconds

Vectorized Monte Carlo over your full dependency network — with sensitivity ranking of which tasks actually drive the finish date.

Benchmark-validated engine

The resource-leveling engine is validated against published academic scheduling benchmarks (RCPSP/PSPLIB) — not a heuristic we eyeballed.

$800K per day

The cost of delay in late-stage pharma (Tufts CSDD). CritPath prices delay into prioritization and staffing decisions via Cost of Delay + WSJF.

Tamper-evident audit trail

Per-organization append-only hash chain with in-app integrity verification — the AACE Level 4 foundation, shipped today.

Pricing

One simple seat — $10/user/month, every feature. The market’s own alternatives are a $149 desktop add-in or a $10,000-a-seat analyst tool; this is the engine in between.

Pro

Most teams

$10 /user/mo

Everything: CPM, TOC/DBR, Monte Carlo, decision gates, WSJF, portfolio dashboard, 2FA/MFA, audit log, API. Unlimited projects.

Join the waitlist

Enterprise

Custom

Dedicated CSM, training, custom integrations; on-prem, SOC 2, SAML, HIPAA BAA, and 21 CFR Part 11 on the Enterprise roadmap.

Talk to us

Build your own organization of AI employees — configurable agents with custom skill sets that work alongside your team. Billed separately by usage, metered per completed job: the estimated token cost of that job plus a stated platform margin. No credits to buy, no hourly minimum, no surprise burn — and a run that fails bills nothing.

See full pricing, seat rules, and how AI usage is metered

Model your next milestone.

Spin up a project, import your task list from CSV or Excel, and see how solid your dates really are — in under five minutes.

CritPath AI on PeerPushCritPath AI rating on PeerPush