Document & data intelligence
Large, unstructured, badly organized source material turned into something answerable. Parsing, retrieval, citation, and the evaluation work that proves the answers hold up.
Products & capability
The Tools at Hand
Here’s a look at what we’re working on that’s available to the public, an idea of some of the things we specialize in, and a look ahead at what’s to come.
Available now
BidBrief
AI document intelligence for engineering and construction document sets — and any other field where the governing sentence is buried in hundreds of pages. Ingests the full set, preserves its structure, and answers questions with citations back to the source page. Live on the web; native iOS in development.
Document parsing
Retrieval & citation
LLM orchestration
Multi-expert analysis
Structured extraction
Export & reporting
In development
WhisperTicket
Order-taking by voice, for rooms where nobody has a spare hand. A server speaks the order the way they’d say it out loud; it arrives as a structured ticket against the right table and seat, priced from the real menu. Built around the messy reality of a noisy floor rather than a quiet demo.
Speech recognition
On-device processing
Menu parsing
Structured capture
Floor & seat mapping
Offline tolerance
Private beta
CareShar
One shared hub for families managing someone else’s care. Medications, appointments, care observations and tasks in a single timeline every guardian can see — so the person who took the 3 a.m. call and the person driving to the next appointment are working from the same information.
Shared care timeline
Medication tracking
Role-based permissions
Guardian invitations
Push notifications
Export & reporting
The rest of the portfolio
What else is in the workshop
Named once they’re at beta or beyond. Until then, here’s the problem each one solves and what it can do.
Remote field inspection viewing
In build
An inspection happens in a pipe, a plant or a crawlspace, and the
person who needs to see it is three hours away. This puts the live
feed from the field machine on any browser, so the reviewer watches
it happen instead of waiting for the drive to arrive.
Live streaming · Real-time sync · Desktop agent · Browser client · Session recording
Field inspection & defect coding
In build
Inspection software for crews who work where the signal doesn’t.
Records assets, observations and standards-based defect codes against
a local database, captures video alongside the data, and syncs when
there’s a connection to sync to.
Offline-first · Local data store · Standards-based coding · Video capture · Hardware integration
Deliverable reconciliation
In build
At closeout, somebody has to prove the work matches the record. This
reconciles what was scheduled against what was actually inspected and
delivered, finds the segments with no matching record, and produces
the discrepancy report as a finished workbook.
Data reconciliation · Discrepancy detection · Spreadsheet pipelines · Report generation
Project dashboards from raw schedules
In build
Project data lives in a spreadsheet nobody can read at a glance. This
turns that export into a stage-by-stage picture of where the work
actually stands, generated on upload rather than rebuilt by hand
every reporting cycle.
Automated analysis · Progress modelling · Interactive dashboards · Scheduled reporting
Relationship & referral intelligence
In build
Membership organizations are full of people who should be talking and
never meet. This reads what members actually do, matches them on
substance rather than category, and grounds every suggested
introduction in evidence instead of guesswork.
Semantic matching · Vector embeddings · Evidence grounding · Multi-agent analysis · Match scoring
Adaptive study & assessment
In build
The document-intelligence engine behind BidBrief, repointed at
coursework. Generates study material from a student’s own
source documents, mines every page for answers with citations, and
adapts to what they keep getting wrong.
Question generation · Deep document mining · Citation tracking · Practice modes · Study exports
Delivery & project intelligence
In build
Internal tooling that reads the state of ongoing work and reports on
it — a business-intelligence layer over delivery, rather than
another place to type status updates by hand.
Automated status · Scheduled jobs · Reporting pipelines · Dashboards
Disciplines
For the problems you face today, and tomorrow
Different products, four recurring shapes. If your problem looks like one of these, we’ve built it before.
Business intelligence
Operational data assembled into something a decision can be made from — reporting pipelines, scheduled aggregation, and dashboards that answer the question people were actually asking.
Technology integration
Making systems that were never designed for each other work together. APIs, auth, data migration, and the sequencing that keeps a cutover from taking the business down with it.
Applications, mobile and web
Native mobile and typed web applications built to ship — including the deployment, monitoring and release plumbing that decides whether version two ever happens.
Capability
What we work in
Listed because it’s running in production somewhere, not because it’s fashionable.
AI & document intelligence
LLM orchestrationRetrieval & RAG
Document parsingCitation tracking
Multi-expert analysisStructured extraction
Evaluation harnesses
Voice & media
Speech recognitionOn-device processing
Audio pipelinesLive video streaming
Session recording
Data & reporting
Reporting pipelinesSpreadsheet automation
Data reconciliationInteractive dashboards
Scheduled jobsVector search
Applications
Native mobileProgressive web apps
Offline-first syncReal-time updates
Push notificationsDesktop applications
Platform & integration
API designAuthentication
Role-based permissionsData migration
Hardware integrationCloud deployment
CI/CD
Practice
Domain-driven designAutomated testing
Threat modellingAccessibility
Technical documentation
What are you working on?
Tell us what it looks like from where you sit. If we’re not the right people, we’ll say so.