Ingest
Bring together the available images, video, audio, documents, scans, records, and spatial sources.
FIELD EVIDENCE + LEGACY DRAWINGS
Turn source material into validated, ArcGIS-ready data.
ContextFormers extracts assets, text, geometry, and relationships from 360° captures, LiDAR, PDFs, and as-builts—then keeps every result connected to its evidence and review state.
Facility
360° + LiDAR → spatial asset record
ONE OPERATING MODEL
Build the operational record without losing the source.
Extract
Detect, read, and locate the people, assets, text, geometry, events, and signals that matter.
Form context
Connect extracted signals to the customer's people, assets, places, terminology, relationships, and source evidence.
Validate
Retain confidence, compare evidence, apply privacy treatment, and route important results through human review.
Deliver
Return accepted output to search, GIS, asset systems, APIs, and the workflows already in use.
Trust runs through every step: privacy, governance, provenance, confidence, and human review.
TWO WORKFLOWS. ONE STANDARD OF EVIDENCE.
Build the operational record without losing the source.
AI accelerates the extraction. Confidence, provenance, and human review make the output usable.
01 · FACILITY INTELLIGENCE
Turn a physical facility into a usable digital asset record.
Capture indoor spaces with 360° imagery and available LiDAR. Detect assets and signage, read labels, locate accepted records on the floor plan, and deliver attributed output for ArcGIS Indoors and facility systems.
- Multi-view detection with confidence scores
- OCR for room, equipment, and safety labels
- Source-linked review before delivery
- GeoJSON and ArcGIS Indoors-ready feature layers
DEPLOYED IN PRODUCTION
Licensed to geoConvergence under a signed teaming, joint-development, and licensing agreement, and running within its Scan2Twin workflow for ArcGIS Indoors.
Source panorama
Detected asset + confidence + review state
02 · UTILITY DATA PREPARATION
Turn legacy PDFs into reviewed network data.
Interpret each utility’s symbols and drawing conventions. Extract candidate assets, attributes, geometry, and connectivity; align them to known geography; and route exceptions for GIS and engineering review.
- Utility-specific symbol and annotation learning
- Candidate point, line, and connectivity extraction
- Geospatial alignment using known control evidence
- Reviewed data packages for Utility Network migration
JOINT DEVELOPMENT IN PROGRESS
A signed teaming, joint development, and licensing agreement with geoConvergence — an Esri Gold Business Partner and 20-year Cornerstone Partner. Facility intelligence — indoor asset detection from 360° scans and LiDAR — is licensed into their Scan2Twin pipeline and running in ArcGIS Indoors, with joint work underway on infrastructure intelligence for utilities.
Legacy drawing
Source construction record before extraction
Candidate features → alignment → review
TRUST RUNS THROUGH EVERY STEP
Every result carries the context needed to review it.
SourceReturn to the image, drawing, or record behind the result.
ConfidenceUse model confidence to prioritize acceptance and exception handling.
Review stateKeep proposed, accepted, and rejected records explicit.
Delivery schemaMap accepted output to the fields and systems the team already uses.
PARTNER DELIVERY
ContextFormers intelligence. geoConvergence ArcGIS delivery.
A signed teaming, joint development, and licensing agreement with geoConvergence — an Esri Gold Business Partner and 20-year Cornerstone Partner. Facility intelligence — indoor asset detection from 360° scans and LiDAR — is licensed into their Scan2Twin pipeline and running in ArcGIS Indoors, with joint work underway on infrastructure intelligence for utilities.
START WITH EVIDENCE
Bring one representative facility capture or drawing set.
We’ll identify what can be extracted, what needs review, and what an ArcGIS-ready output package should contain.
hello@contextformers.com