Everyday agents. Big outcomes.
PALANTIR partner • CLAUDE PREFERRED PARTNER
Proxet builds the agents teams turn to daily to multiply themselves and their outcomes. Our embedded engineers, backed by dedicated nearshore centers, build the Ontology, the apps, and agents that undergird critical business workflows. We work in commercial real estate, asset & wealth management, energy, and healthcare.
<1 hour
lease processing time, down from three days, at a global CRE services firm
400+
fields per lease mapped into the Foundry Ontology
2 months
to rebuild carry allocation for a global VC firm
93%
fewer errors in a global pharma company's clinical document pipelines
Palantir gets you from zero to one.
We scale up from there.
Go-live is just the beginning. In the weeks after a Palantir-led deployment, adoption has to hold, the Ontology has to take on new workflows, and someone has to maintain the platform every day.
Proxet’s hyper-care starts there. We tune what’s in production, and train people to use it while we partner with business leaders to develop the backlog of high-impact use cases. We go use case by use case, moving agents from pilot to production to squeeze ROI out of your investment. Our delivery model combines embedded, onshore, and nearshore capacity that gives our clients the proximity they want and the scalability they need.
→
Agents and decisioning
Lease renewal · Deal screening · LP reporting · Research briefs
→
Guardrails
Human approval · Audit trail · Markings and access
→
Foundry Ontology
Golden copy and data products your apps read
→
Your systems
Yardi · MRI · Salesforce · Fund admin · Documents
Palantir capabilities, delivered in production
We build across the Foundry and AIP stack, from the first data connection to the agents and apps your teams use to make better, faster decisions every day.
Data Integration
Connect ERPs, property systems, fund admin platforms, and documents to Foundry through Data Connection. For a global CRE firm, our Polars pipelines ran 20x faster than standard Spark estimates.
Pipeline Builder
Production pipelines in Pipeline Builder and code repositories, tuned for refresh frequency and compute cost, with legacy SQL logic migrated in.
Digital Twin
Properties, leases, funds, and assets modeled as Ontology objects, with the links and actions operators need. A single lease twin carries 400+ fields.
AIP for Developers
Production agents and custom apps built with AIP Logic, Agent Studio, and OSDK. Claude Code connects through Palantir’s MCP with Ontology permissions intact.
AI + ML
ML models and LLM workflows deployed in Foundry, each benchmarked against verified ground truth before it reaches production.
Real-Time Alerting
Automate turns Ontology changes into action: compliance flags, variance alerts, and notifications routed to the person who owns the fix.
Scheduling and Optimization
Optimization models for routes, crews, and assets. One client's fleet-routing tool weighs fuel, fueling queues, weather, and no-fly zones for every flight.
Workshop and Quiver Apps
Operational apps with custom widgets, version comparison, and browser-style navigation, embedded in your own portal when you need it.
Ontology first. Agents second.
AIP agents are only as good as the objects, links, and actions they read. So we build the foundation in order, and every agent action routes through guardrails before it touches your systems.
01
Collect
Connectors and staging for Yardi, MRI, Salesforce, fund admin platforms, market data, and documents.
02
Correct
Machine checks plus human review where accuracy matters.
03
Normalize
Domain models that match your business: properties, leases, and tenants, or funds, positions, and investors.
04
Golden copy
One governed source of truth, cataloged and permissioned with markings.
05
Data products
The objects your Workshop apps and agents actually read.
Guardrails on every action
Agents call approved tools only. Higher-risk actions wait for human sign-off. A full audit trail records what each agent saw and why it acted, under role-based access and SOC 2 controls.
Palantir for commercial real estate
Your data lives in PDFs, Yardi, and spreadsheets that don’t agree. We model properties, leases, and tenants in the Ontology so abstraction, reporting, and compliance run from one source.
Client story
Global CRE services firm
For this firm, Proxet shipped a working lease abstraction tool to production in under four months. Processing dropped from three days to under an hour, and the app runs inside the firm’s own portal with 5-second page loads as volume grows toward 12,000 leases a year.
Use cases
- A lease abstraction agent in production, reading 400+ fields per lease with a reviewer validating every abstract, and our accelerator cutting build time by 90%
- Yardi and MRI integration into one Ontology, with client reporting built as a product
- Portfolio compliance scanning that replaces physical property audits
- Portfolio 360: every property, lease, tenant, and vendor in one live view
- Contract agents that extract clauses and link each amendment to the original agreement
- Capital planning agents that forecast spend and match invoices to POs for approval
Palantir for asset & wealth management
Asset managers, private equity, venture, and hedge funds. Your deal flow, portfolio updates, and fund accounting sit in different tools, and your analysts spend the week reconciling them. We put them in one Ontology and build the apps your investment team opens every morning.
Client story
Global venture capital firm
After a year with an off-the-shelf tool that couldn’t handle carry, the firm needed a working model fast. Proxet rebuilt carry allocation on Foundry in two months, fixed the underlying math across the general ledger, and gave partners near-real-time data.
Client story
Venture capital firm
Proxet replaced manual deal flow with 100% passive ingestion and knowledge graphs linking deals, companies, and founders. The firm’s own team now runs its Ontology updates.
Use cases
- Deal flow operating systems with embedded agents, passive data ingestion, and relationship graphs
- Carry and waterfall allocation across partners and funds
- Trade audit and point-in-time growth research for every investment decision
- Portfolio monitoring and valuation, with data standardized from thousands of investor updates
- A deal-screening agent that scores inbound deals against your thesis and drafts the IC memo for review
- LP reporting agents that draft DDQ responses and capital account statements for sign-off
Palantir for energy: emissions and field operations
Your monitoring program depends on routes, crews, and repair decisions that change with weather, fuel, and regulation. We build the models and field tools that keep those decisions current.
Client story
Methane detection company
The client’s survey fleet needed routes that account for fuel stops, fueling queues, no-fly zones, weather, and thousands of points, lines, and pipeline segments. Proxet built a flight plan optimization tool deployed in Foundry, with computational geometry preprocessing and precise coordinates for every computed flight.
Use cases
- Flight and route optimization for survey fleets
- Emissions monitoring agents that triage new methane detections and draft repair work orders
- Carbon accounting with automated Scope 1 and 2 calculations and audit views back to source data
- Maintenance agents that read sensor data and open work orders before equipment fails
Palantir for healthcare & life sciences
Your clinical, operational, and financial data sit in the EHR, claims systems, and research platforms that were never built to talk to each other. We bring them into one governed Ontology, with the access controls and audit trails healthcare data demands.
Client story
Global pharmaceutical company
Scaling Foundry for clinical trial document conversion had left the company with architectural debt and rising errors. Proxet fixed the pipelines, cut error frequency by 93%, and gave the team real-time auditing of every data source and user action.
Also in healthcare
Proxet builds and runs data platforms for one of the largest academic health systems in the US, so our engineers already know how clinical data is structured, governed, and used.
Use cases
- Clinical trial document conversion with a full audit of data sources and user activity
- EHR, claims, and operational data brought into one governed Ontology
- A clinical data-mapping agent that maps study data to SDTM, with a reviewer approving every mapping
- Capacity management that pairs real-time and predicted bed demand with staffing
- Staffing and dynamic scheduling for nursing and support teams
- A denials agent that drafts response letters from Ontology data for clinician review
Plays that work in any Foundry deployment
Decisioning
The Ontology becomes the place decisions get made, with options compared, recommendations ranked, and the chosen action written back to source systems.
- Scenario planning that tests capital, staffing, or pricing changes before you commit
- Recommendations ranked by impact, with one-click actions back to source systems
- Approval workflows that route higher-risk decisions to the right owner
- A decision record: who chose, on what data, and what happened next
Supply chain and distribution
ERP, supplier, inventory, and shipment data in one Ontology, so planners see the whole network instead of one system at a time.
- Control-tower alerts on late, short, or at-risk orders
- Demand forecasting tied to inventory and replenishment
- Vendor risk scoring from operational, financial, and external data
- Procurement agents that compare quotes and draft POs for approval
Logistics and field operations
Routes, crews, vehicles, and assets planned against the constraints that actually apply on a given day.
- Route and fleet optimization around fuel, weather, and restricted zones
- Dispatch and crew scheduling that adjusts as jobs change
- Live tracking of location, temperature, and delivery status, with alerts on deviations
Document to Ontology
Leases, contracts, PPMs, and clinical documents become typed objects your apps and agents can act on.
- High-volume extraction with a reviewer validating each document
- Amendments and versions linked to the original agreement
- Extraction accuracy benchmarked against verified ground truth
Hardware to Ontology
Building sensors, HVAC and BMS feeds, and IoT telemetry stream into the Ontology and link to the assets they measure.
- Live conditions shown next to leases, work orders, and budgets
- Automated alerts that open work orders when readings drift
- Energy and equipment optimization across a portfolio of sites
Legacy to Ontology
Business logic trapped in legacy SQL and aging systems moves into governed Foundry pipelines your team can trace and change.
- Legacy SQL migrated into Spark or Polars pipelines in Foundry
- Brittle pipelines refactored with full lineage and auditing
- Point tools retired once their workflows run on the Ontology
Foundry + Claude
The AI your team already uses, running on data you trust.
- Governed Claude Code connected through Palantir’s MCP
- Claude-powered agents in AIP Logic and Agent Studio
- Ontology permissions enforced on every query
Ontology to your clients
Your Foundry data, delivered to clients and partners as a product.
- Apps embedded in your own portal
- Client reporting built to each client’s spec
- SDKs and APIs that let partners pull governed data themselves
Claude on your Ontology, inside your guardrails
Your team already uses Claude. The risk is that they point it at exported spreadsheets, unmanaged APIs, and data nobody permissioned.
We’re a Palantir Builder partner and a Claude Preferred Partner. We know how to marry the two for maximum impact. Claude models run inside AIP Logic and Agent Studio. Claude Code connects to Foundry through Palantir’s MCP and inherits the same permissions as everything else in your Ontology.
Governed Claude Code
Non-engineers build apps against verified Foundry data, with access enforced by the platform rather than by policy memos.
Claude-powered AIP agents
Document extraction, research briefs, deal screening, and reporting, with human approval on the actions that need it.
Evaluation before release
Our framework benchmarks every prompt against verified ground truth before it ships to production.
Client story
25B AUM investment firm
The firm gave every employee Claude Code and needed to do it without creating shadow AI. Proxet built the data and governance layer in Foundry. Now employees connect Claude Code to Foundry, build their own agents on verified shared data and compute, and ship proofs of concept with confidence, at lower cost than standalone MCP servers.
Start with one workflow.
Scale what works.
Every engagement runs on our Hybrid FDE model. Embedded engineers and onshore SMEs work on site where context and speed matter most, like discovery, Ontology design, and go-live. Our nearshore centers take on the build and run work that scales. We set the mix for each phase, so more of your budget goes to use cases in production.
Start here
AI Operationalization Workshop
A half-day session. We rank your Foundry use cases by value, check data readiness, and sequence the first build.
Build
Use case to production
One workflow, from Ontology to Workshop app or AIP agent, live with real users. Recent builds reached production in two to four months.
Accelerate
Foundry accelerators
Pre-built Foundry packages, starting with lease abstraction, that cut build time by 90% and ship with a full audit trail.
Run
Managed Foundry team
Platform maintenance, upgrades, Ontology right-sizing, and FinOps that break spend down by workflow and department. We also add Quiver analysts, OSDK engineers, and ML capacity to your team.
Forward-deployed speed. Nearshore scale.
Our Hybrid FDE model pairs embedded engineers and onshore SMEs with dedicated nearshore centers. You get the speed of a team inside your business and the capacity of a 350-person firm behind it.
Our ML engineers have taken multi-agent systems and LLM agents from prototype to production.
Every engineer on our Foundry team has shipped live pipelines, Ontologies, and Workshop apps. The team covers architecture, data engineering, app building, and AIP, and also works in Databricks, Snowflake, Fabric, AWS, and GCP when your stack extends beyond Foundry.
With teams in the USA, Eastern Europe, and LATAM, we combine hyper-care and scalability in one team seamlessly.
“Proxet allows our customers to have increased choice for their development services while upholding the high quality bar our customers have come to expect from Palantir.”
Pick one workflow. We’ll get it into production.
Book a half-day AI Operationalization Workshop. You'll leave with your Foundry use cases ranked by value and a plan for the first build.