Autonomous data modelling
Relationships across invoice, order, receipt, vendor, employee, and payment data are inferred automatically. No manual data model per system or ERP version.
PLATFORM
One platform for risk analytics, investigation, and evidence
Connect your data. Discover risks, investigate findings, and deliver audit-ready evidence—without moving between four tools to do it.
CAPABILITIES
What the platform does
Unlimited risk analytics
Run hundreds of analytics across millions of transactions, and define new ones in plain language rather than script.
Conversational data analysis
Ask questions of your data directly—find duplicate payments in AP, explain the top case—and get answers tied to the underlying transactions.
Investigation workspace
Cases, notes, documents, and reviewer feedback in one place, with AI summaries and search across the engagement.
Explainable findings
Every finding carries the transactions behind it, the logic applied, and the data lineage—so it can be traced, challenged, and used as audit evidence.
Dashboards and reporting
Executive dashboards and audit-ready reports generated from the same findings, so reported figures reconcile with what was tested.
Semantic and fuzzy matching
Near-duplicate, fuzzy, and semantic matching at scale—required for duplicate payments, vendor deduplication, and relationship analysis.
Relationship and graph analysis
Model vendors, employees, bank accounts, and addresses as a connected graph to expose collusion that pairwise rules cannot see.
Isolated client environments
A dedicated workspace per organisation on AWS, with least-privilege access and no cross-client data movement.
DIFFERENTIATORS
Why teams choose it
Move from “what looks risky” to “why is it risky”. Connect related transactions and entities and understand the underlying exposure.
Once data is modelled, the marginal cost of an additional analytic is low—so scope is decided by what matters rather than by analyst capacity.
Evidence and lineage are attached to findings from the start, rather than reconstructed when someone asks how a number was derived.
Data preparation, analytics execution, investigation, and reporting are automated, which removes most of the elapsed time in a risk assessment.
Detection, investigation, dashboards, and case management in one platform reduces dependence on separate tools and specialist teams.
Managed SaaS by default, with private cloud and on-premises options where data residency or architecture requires it.
COVERAGE
Ten risk domains on one platform
Each domain has its own page covering the scenarios tested and the data required.
Risk domains
Next
Frequently asked questions
What does the foretale.ai platform actually do?
foretale.ai connects to your ERP and transaction data, models the relationships in it automatically, runs risk analytics across the full population, and returns findings with the underlying transactions and logic attached. Findings then move into investigation workflow where they are validated, escalated, tracked, and closed.
Do we need to build a data model first?
No. Autonomous data modelling is part of the Understand step. foretale.ai infers how your invoice, order, receipt, vendor, employee, and payment records relate, so analytics can start without a manual mapping exercise for each system or ERP version.
Which ERP systems are supported?
The platform is ERP-agnostic and has been validated on real enterprise datasets from SAP, Oracle, and Microsoft Dynamics environments. Because relationships are inferred from the data rather than hard-coded per system, other sources can be onboarded the same way.
Can we create our own analytics?
Yes. Analytics are defined in plain language rather than script, so risk and audit teams can create and adjust tests themselves without writing code or waiting on engineering.
How is our data isolated from other customers?
Each organisation runs in a dedicated workspace on AWS with its own configuration, storage prefixes, and database authorisation. Access is authenticated through Amazon Cognito with role-based, least-privilege authorisation.
Is customer data used to train AI models?
No. Customer content is not used to train third-party foundation models. AI features operate only on data the authenticated user and project are authorised to access.
NEXT STEP
See it on your data.
Request a focused walkthrough—usually within one business day.