Resource search

Search DataDrivenAEC

Open the full search page →

Pelles AI

AI platform for trade contractors that turns drawings, specs, and contracts into RFIs, submittals, and QA/QC checklists

Document-AI platform built for construction trade contractors (MEP, mechanical, electrical, plumbing) that reads drawings, specifications, contracts, and schedules to generate project-ready deliverables -- RFIs, submittal logs, QA/QC checklists, and proposal sections -- each cited back to its source document. A companion feature ("Compare") automates revision review by matching and aligning drawing sheets to flag changes between versions, and a custom-agent offering ("Workshop") builds workflow-specific AI agents around a client's existing processes.

Sources disagree and no primary-source funding announcement was found to reconcile them: Crunchbase records a single disclosed round -- a $75K pre-seed in February 2023 led by Brick & Mortar Ventures (matching this queue's seed source); StartupHub.ai reports $2M total raised; CB Insights lists the company at "Seed VC (Alive)" stage with 8 investors (Centre Street Partners, Brick & Mortar Ventures, Ginossar Ventures, The Garage, CEMEX Ventures LEAPLAB, and 3 unnamed others) but no total amount. raisedEarly Stage
Founded 2023 · No total customer count publicly disclosed. Company site names 7 trade contractors as customers or references (F.G. Haggerty, Kidwell, DeKalb Mechanical, Press Mechanical, Vertical Mechanical Group, L.A. Fuller & Sons, Peck Hannaford + Briggs) plus "10+ additional" unnamed. Marketing claims "3,000+ construction projects" and "65,000+ issues caught early" processed on the platform (self-reported, methodology not disclosed). ~15 employees per StartupHub.ai (unverified aggregator estimate). Bronze Associate Member of SMACNA (Feb 2024) and an Autodesk AECO Technology Partner.
Website →← All Tools

Why This Tool Exists

The Problem

Trade contractors manually read hundreds of pages of drawings, specifications, contracts, and addenda to prepare bids, draft RFIs, and track scope changes across document revisions -- the vendor cites contractor estimates of 100+ hours spent per project on this review, with issues often missed until later phases and causing rework.

The Solution

Pelles ingests project documents (drawings, specs, contracts, schedules) and generates draft deliverables -- RFIs, submittal logs, QA/QC checklists, proposal sections -- with citations back to the source page, rather than returning a list of search results. "Pelles Compare" automates revision review by matching and aligning drawing sheets to instantly highlight changes between versions. "Pelles Workshop" builds custom AI agents around a client's specific workflows (e.g. automated transmittals, punchlists, change-order recommendations), while "Pelles Core" offers a pre-built, plug-and-play version of the same capabilities.

How You Use It

Delivery Method
SaaS (Web Application)Custom AI Agents (Workshop tier, built and run by vendor)
Integrations
ProcoreAutodesk Construction CloudAutodesk BIM 360Primavera P6BuildingConnectedSmartsheetSharePointOneDriveBoxDropboxGoogle DriveMicrosoft ExcelMicrosoft TeamsOutlookPlanGridNewforma (Beta)BuildOps (Beta)Gmail (Beta)DocuSign (Beta)Bluebeam (Beta)Trimble Connect (Beta)Trimble Project Sight (Beta)Fieldwire (Beta)OpenSpace (Beta)Trimble Viewpoint Vista (Beta)Sage 300 CRE (Beta)CMiC (Beta)Accubid (Beta)
Project Phases
Bidding/Procurement, Construction Administration, Closeout
Project Types
Commercial, Institutional

Data Transparency

Exactly what this tool uses and how

Input
What it needs
Required:Project documents (drawings, specifications, contracts, addenda)
Optional:Project schedules (e.g. Primavera P6), Historical bid/estimate data, Connected platform data via integrations (Procore, ACC, etc.)
Formats:PDF, DWG/BIM files (via Autodesk Construction Cloud/BIM 360 integration), Microsoft Excel, Native files synced from integrated platforms -- exact supported format list not publicly enumerated
Output
What you get
Format:Web-based dashboard plus generated documents; two-way sync back to connected project-management platforms
Fields:RFIs, Submittal logs, QA/QC checklists, Proposal sections, Drawing revision comparison/redline reports, Transmittals, Change-order recommendations, Source citations linking each output back to its originating document/page
Algorithm
How it works
Model:Not publicly specified -- vendor does not disclose the underlying LLM(s) or whether models are proprietary or built on third-party foundation models. Public GitHub repository (aws-serverless-docling) indicates use of the open-source Docling document-parsing library deployed on AWS Lambda for at least part of the document-ingestion pipeline.
Accuracy:Not publicly specified -- no independently audited accuracy figures published. Vendor cites customer-reported "70-80% time savings" in plan/spec review (per one estimator) and "100+ hours saved per project" -- self-reported claims, methodology not published.
Privacy
How your data is protected
Retention:Per the privacy notice, personal data is retained "as necessary for the purposes set forth" and may be kept longer for legal/regulatory compliance or litigation risk; no fixed retention period is disclosed for project/document data specifically.
Training:The homepage states customer data is "never shared or used for model training." The privacy notice separately confirms, specific to Google Workspace API data, that it is "not used to develop, improve, or train generalized AI and/or ML models."
Compliance:No SOC 2, ISO 27001, or other third-party security certification publicly disclosed, Enterprise encryption at rest and in transit (vendor claim), 7-year audit trail retention (vendor claim), RBAC + SSO authentication (vendor claim)
API
Integration
Endpoint:Contact vendor for API access -- no public API documentation found
Method:Not publicly documented

Use Cases

  • ·Extracting scope, requirements, and hidden details from drawings, specs, and contracts during bid preparation
  • ·Automating RFI drafting and submittal log creation from project documents, with source citations
  • ·Comparing drawing/document revisions to instantly flag scope and specification changes across versions ("Pelles Compare")
  • ·Generating trade-aware QA/QC checklists from project documents
  • ·Building custom AI agents ("Pelles Workshop") for contractor-specific workflows like transmittals, punchlists, and change-order recommendations

Pricing

Free
Not publicly specified -- no self-serve free tier found; site offers a 30-minute walkthrough/demo rather than free signup
Pro
Not publicly specified -- no self-serve pricing tiers published
Enterprise
Custom/quote-based pricing via sales contact (demo required); "Pelles Workshop" custom-agent builds are priced separately from "Pelles Core"
Sources & Research NotesResearch date: 2026-07-17
Fields Checked (12)
problem, solution, deliveryMethod, integrations, disciplines, projectPhases, yearFounded, founders, privacy-training, privacy-retention, customerBase (named customers, SMACNA membership, Autodesk partnership), algorithm-approach (Docling usage confirmed via public GitHub repo)
Not Found (9)
exact total funding amount (sources conflict: $75K vs $2M vs undisclosed), pricing tiers, API documentation, SOC 2/ISO certification, underlying LLM/model vendor, independently audited accuracy metrics, exact employee count, data hosting region/provider, exact input/output file format list
About this research: This record summarizes the cited sources and fields checked. A research date does not establish hands-on testing. Consult the original sources for current details and limitations.