AI construction scheduling in 2026: what ALICE, nPlan, Karmen, SmartPM, Outbuild, and Planera each claim to do

Four different jobs under one label

Our framing, not a vendor claim: before comparing tools, decide which job you are buying for. ALICE describes scenario simulation and optimization. Karmen describes drafting and editing schedules from drawings and specs. nPlan describes risk forecasting from a large archive of past schedules. SmartPM describes automated analytics on the CPM schedules you already maintain. Outbuild and Planera are scheduling platforms in their own right that describe built-in AI. Several of these overlap at the edges, so read each vendor's own description below.

Simulate and optimize: ALICE

ALICE describes itself as AI-powered generative construction scheduling software that automates ‘what-if’ scenario exploration; its homepage says it rapidly simulates millions of possible construction scenarios to find the most efficient plan (vendor claim). It lists ALICE Optimize (works with an existing Primavera P6 or Microsoft Project schedule), ALICE Analyze (schedule, project, and portfolio analytics), ALICE Model (a baseline from a 3D model), ALICE FieldPlan (in beta, field plans from drawings), and a Schedule Insights Agent for chatting with a schedule. ALICE says you can import schedules from Microsoft Project, Oracle Primavera P6, or Oracle Primavera Cloud.

ALICE says it has been deployed on more than $410 billion of construction projects and cites 17% shorter project duration, 14% labor cost savings, and 12% equipment cost savings (vendor claims, not verified; the page does not describe how they were measured). It lists project types from data centers and energy plants to railways, highways, bridges, and tunnels, and says the platform is SOC 2 Type II compliant and Cyber Essentials Plus certified.

Pricing: not publicly listed. ALICE's pricing page offers a custom quote based on the size and scope of your projects, lists unlimited users and unlimited data, and says ALICE is most often used by GCs and owners on projects with a construction value of at least $75M.

Draft from documents: Karmen

Karmen calls itself an AI scheduling co-pilot. It says it reads project documents (drawings and specs as PDF, historic schedules and templates from P6, MSP, or XER), and can then be steered in natural language to build a WBS, activities, and logic links from scratch, update existing schedules including time-impact analysis with structured fragnets, and track float and critical-path movement between monthly updates. It exports to Primavera P6 and Microsoft Project.

Karmen says it reads thousands of pages of specs and drawings in minutes and learns from your historic schedules to sequence work the way you would (vendor claims, not verified). On security it says SOC 2 reports are shared on request, data is encrypted in transit and at rest, and project documents are never used to train shared models. Pricing: not publicly listed; the site routes to a demo booking.

Forecast risk from history: nPlan

nPlan describes AI for project controls and delivery trained on a dataset of 750,000 historical schedules representing over $2Tn of construction spend, and says its AI has $500bn of projects under active management (vendor claims). It says it works with your existing schedules and describes four uses: generating and editing schedules from scope documents, forecasting the uncertainty of every activity, flagging risky activities and checking them against an uploaded risk register, and project and portfolio summaries. A blog post featured on its homepage says Schedule Studio now connects to Claude and ChatGPT for questions, edits, health checks, and exporting an updated XER.

nPlan's homepage quotes customers on portfolio oversight and schedule risk analysis; one quote is credited to the Transpennine Route Upgrade. Pricing: not publicly listed; we found no pricing page (nplan.io/pricing returned a 404 on Oct 15).

Audit the schedule you have: SmartPM

SmartPM describes a project controls and schedule analytics platform that integrates with your existing CPM scheduling software and provides what it calls Automated Project Controls: schedule quality checks, progress against baselines, delay and compression metrics, a web-based Gantt, and company and executive dashboards, aimed at GCs, owners, and developers in commercial construction. Neither its homepage nor its pricing page describes the product as AI, which is why our directory tags it Not AI-first; it is included here because buyers often evaluate it alongside AI scheduling tools.

SmartPM is the one vendor in this group that publishes full prices. Its pricing page lists Essentials at $12K per year for up to 50 projects and Controls at $25K per year with 5 Controls slots, both with unlimited users; Controls adds what-if scenarios and time impact analysis, portfolio insights, and dedicated onboarding. The page also offers a 60-day risk-free trial.

Scheduling platforms with built-in AI: Outbuild and Planera

Outbuild calls itself the first fully integrated, AI-powered, collaborative scheduling and field coordination platform built for construction (vendor claim) and markets “AI agents for scheduling, built on construction logic.” It connects the master schedule to lookaheads and weekly work plans, and its pricing FAQ says it can replace P6, MS Project, and Excel, exports to PDF, XML, or XER, and integrates two-way with Procore and with Autodesk Construction Cloud. Outbuild says it starts at $999 per month billed annually, priced by annual construction volume rather than by user or project, with unlimited users and projects; its Pro tier is custom-priced. It cites 450+ customers and $50B+ in active projects (vendor claims).

Planera describes AI-native CPM scheduling: critical path and float management, DCMA-14 quality checks, resource loading, field schedules tied to the master, and an AI assistant it calls manny that turns plain-English requests into schedules and updates, explores what-if scenarios, flags missing logic and shrinking float, and drafts owner updates and delay narratives. Its pricing page describes customized pricing with unlimited users; prices are not publicly listed.

What to ask before a pilot

These are our suggestions, not findings. Run every tool on the same closed project you already know well, so you can judge its output against what actually happened. For tools that draft schedules (Karmen, nPlan's generative features, Planera, ALICE Model), ask how many hours a scheduler needed to turn the draft into a submittable baseline, and check logic density and open ends. For tools that forecast or score (nPlan, SmartPM, ALICE Analyze), ask what the forecast or score was on past projects versus actual completion. And confirm the round trip: whether the tool reads and writes your P6 or MS Project files without losing codes, calendars, or resource loading.