AI operations for private equity

Design-partner programme now open

Put AI to work
across the portfolio.

Ordinal turns proven operating playbooks into controlled AI workflows that each portfolio company can approve, run and measure—without centralising sensitive data.

Built for

PE operating teamsPortfolio-company leadersRegulated operations

From pilot to portfolio

One operating method. Locally controlled.

Package the workflow once, let each management team release it responsibly, then compare the outcomes that matter.

Ordinal / Portfolio 01
Illustrative product view

PORTFOLIO PLAYBOOK

Claims handling / v2.4

Sponsor operating team publishes a reusable workflow and minimum control intent.

Company ALocally approved
Company BLocally approved
Approved metric01 shared method

The operating advantage

01

Deploy what works

Turn one successful workflow into a reusable portfolio playbook.

Reusable control packs

02

Keep control local

Each company approves its own release and retains its operational data.

Independent runtime cells

03

Prove the value

Connect workflow activity to evidence, quality and financial outcomes.

Comparable metric contracts

Where to begin

High-friction work with a clear owner and a measurable result.

Start where operating value and responsible control can be demonstrated together.

01

Claims and complaint handling

Insurance

02

Customer-file quality assurance

Financial services

03

Clinical and case administration

Healthcare

04

Supplier and third-party reviews

Cross-sector

05

Regulatory evidence preparation

Cross-sector

06

Acquisition policy onboarding

Portfolio operations

Shared playbooks. Independent control.

A portfolio architecture that respects the company boundary.

The sponsor shares the method. Each company retains release authority, operational data and the ability to stop its workflows.

01 Sponsor control plane
CONTROL PACKClaims / v2.4
WorkflowControlsMetrics
Approved template

02A PORTFOLIO COMPANY

Northstar MGA

Local approval
Isolated data
Kill switch

02B PORTFOLIO COMPANY

Harbour Claims

Local approval
Isolated data
Kill switch
Raw cases do not cross this boundary
01

Distribute the operating method

Version the workflow, minimum controls and metric definition at sponsor level.

02

Approve and run locally

Configure systems, permissions and accountable approvers inside each company.

03

Return evidence, not raw cases

Share only pre-agreed receipts and outcome metrics with the sponsor.

Explore the platform

Policy to proof

Every run should be understandable.

An evidence receipt connects the approved design to what actually happened: versions, actions, human decisions, exceptions and outcomes.

See the trust model

EVIDENCE RECEIPT

CLM-2048 / 0142
Verified
Workflow
Claims review 2.4
Control pack
INS-UK 1.8
Local approval
2 of 2
Human checkpoints
3 passed
Exceptions
1 resolved
Evidence integrity
Matched
Run signature9d2a…c71f

Built for regulated operations

A common model, configured to the industry.

Begin with shared infrastructure and adapt the workflow, control and evidence layers to each operating context.

Design-partner programme

Start with two companies and one repeatable workflow.

A 12-week programme to prove control, operational value and second-company reuse.

Discuss a portfolio pilot

Field notes

How portfolio AI becomes operating infrastructure.

Questions, answered

The practical details.

01What does Ordinal replace?

Ordinal does not replace your portfolio monitoring, GRC or line-of-business systems. It adds the controlled execution layer between a sponsor playbook and the work performed inside each portfolio company.

02Does Ordinal centralise portfolio-company data?

No. Operational cases and sensitive records stay inside each company’s isolated runtime. The sponsor receives only approved evidence and metrics defined in advance.

03Who approves an AI workflow?

Every portfolio company retains local release authority. Named makers and checkers review the workflow, its permissions and its controls before it can run locally.

04Can Ordinal work with existing systems and models?

That is the intended model. Ordinal orchestrates approved workflows across existing data, software and AI services while preserving local boundaries and evidence of each action.

05What evidence does a workflow produce?

Each run can produce a structured receipt recording the version, approvals, actions, model and tool calls, human checkpoints, exceptions and customer-defined outcome measures.

06Which regulated industries are supported?

The design-partner programme is intended for insurance, financial services, healthcare and regulated business services. Initial control packs are configured around the customer’s own policies and obligations.

07What does a design-partner pilot involve?

A typical pilot covers one repeatable workflow across two portfolio companies over twelve weeks, with agreed control, operational and reuse measures.

08Is Ordinal compliance or legal advice?

No. Ordinal supports customer-defined controls and evidence. It does not provide legal advice, determine regulatory obligations or guarantee compliance.