Our work
The work, in detail
Our work on Palantir Foundry and the systems around it: the problem each engagement started from, and the number it moved. Choose your industry below to see what each one looks like in yours.
The outcomes
THE TOPICS
Eight kinds of work, one pattern
Governed data, a shared ontology, and outcomes the business can measure: the same pattern at every size.
Find your programme
Choose a kind of work to see the relevant engagements, and your industry to see what each one looks like there.
One platform
Replacing spreadsheet planning for 250 users
250+ people and numerous systems onto one planning platform, replacing spreadsheets circulated by email.
Read the engagement →Contract-level
Planning at contract level where the system showed only totals
Contract-level granularity and full history where the legacy system summed everything into aggregates.
Read the engagement →Replatform
Replatforming an investment plan off a retiring modelling tool
An investment plan moved off a retiring modelling tool onto one governed, scalable model, with scenario modelling built in.
Read the engagement →Steering
One steering view where every unit reported separately
The first unified view of pricing, performance and client intelligence for the executive team, built on the business and finance data models together.
Read the engagement →Exposure
One exposure figure per counterparty where dashboards disagreed
Counterparty and obligor records resolved at source, so credit exposure is reported once, correctly, in one application.
Read the engagement →Operating model
Giving a data function a structure that scales
A diagnostic and a one-to-two-year roadmap, right-sized to the organisation: lean for small firms, layered for large ones.
Read the engagement →Control framework
Replacing manual sign-offs with controlled workflows
Manual confirmations replaced by structured workflows and real-time consistency checks across every submission.
Read the engagement →One definition
One definition for every financial KPI across a group
One authoritative definition per financial KPI, with a named owner for every source, figure and calculation.
Read the engagement →Surveillance
Federating 100+ sources across seven jurisdictions
100+ sources and 20+ processes across seven countries on one platform, with location-based access to the row, and legacy systems switched off.
Read the engagement →ESG hub
Consolidating ESG reporting from feeds and manual uploads
Automated external feeds and manual business-unit uploads consolidated into one ESG source of truth, with lineage and versioning.
Read the engagement →AI-built
A thousand filings a year, off undocumented spreadsheets
AI reads the legacy workbooks and writes down their logic; what ships is governed, auditable code. 80%+ less manual effort in the proof of concept.
Read the engagement →AI analyst
Answering executive questions from a re-engineered ontology
The product-centric ontology re-engineered into a semantic layer an AI analyst can read, then tuned for the leadership team.
Read the engagement →AI assistant
Letting non-technical users query steering metrics, with lineage
Natural-language questions over the validated ontology, with the definition and origin of every KPI on demand.
Read the engagement →Operational pipeline
Eliminating silent failures in high-volume document processing
Foundry as a live operational system: inbound statements from 8 counterparties and ~50 contracts processed end to end with zero silent failures and a full audit trail.
Read the engagement →Automated reconciliation
Cutting discrepancy identification by 90% in statement reconciliation
A language model reads the statements out of emails and PDFs and matches them to the processed files; discrepancies surface at once.
Read the engagement →Entity resolution
Resolving which records belong to one client, across every location
A real client 360 across every location, which became the foundation for the compliance and risk applications that followed.
Read the engagement →Consolidation
Three systems into one, and a run cost cut by 90%
A 26-year-old claims application and two other systems consolidated onto Foundry: one record per policy and life, and the run cost down from $10M to $1M a year.
Read the engagement →Your programme is on this list somewhere
Tell us which of these problems looks like yours, and we’ll tell you honestly how close the parallel runs.
contact@ainurinsights.com