Private equity
AI for Private Equity Portfolio Companies
Last updated
The short answer
New Odyssey is an AI implementation partner for private equity firms and their portfolio companies in the US and UK. We build AI agents that do back-office work inside a company's existing systems, such as the month-end close, new-hire onboarding, contract review and privacy requests. The approach is to prove one workflow at one portfolio company, measure it against a baseline, and then repeat the same build at the next company. Packaged Pegasus agents cover the common workflows; custom agents cover the ones that are specific to a company.
Why portfolio companies are a good fit for AI agents
Mid-market and lower mid-market portfolio companies tend to have the same back-office shape: skilled finance, HR and operations people moving work between systems that were never built to talk to each other. Buy-and-build makes it sharper. Each acquisition adds another legal entity, another accounting system and another way of doing the same job.
That is repeatable work with a clear owner, which is what an AI agent is good at. It is also the same problem at the next company in the portfolio, so a build that works once does not have to be invented again.
What we build
In every case the agent works inside the systems the company already runs, a person approves at defined points, and every action is logged.
| Workflow | What the agent does | How it is delivered |
|---|---|---|
| Month-end close | Transaction matching, expense categorization, approval routing and the close task list | Pegasus Finance Agent |
| New-hire onboarding | Document collection, system provisioning, benefits enrollment, training and manager notifications | Pegasus People Agent |
| Contract review | Extracts key clauses, compares them with your playbook, flags deviations and writes the review summary | Pegasus Contract Agent |
| Privacy requests | Intake, identity verification, data discovery across connected systems, redaction and the response | Pegasus Compliance Agent |
| A workflow specific to one company | Designed around that company's process and systems | Custom agent |
How a portfolio roll-out runs
Each stage is a decision point. The firm and the portfolio company only continue if the previous stage justified it.
- Pick one company and one workflow. Choose the process that costs the most hours and has a named owner.
- Assess it. We map the process as it actually runs, record a baseline and say plainly whether it is worth building.
- Build and prove it on that company's real systems and data, with the people who do the work.
- Measure against the baseline. If the numbers do not hold up, we recommend against scaling it.
- Repeat at the next company. The second build reuses most of the first, so it is faster.
What a deal team or operating partner should ask any AI partner
- Do they build, or only advise? A roadmap does not change how the close runs.
- Can the first step be scoped for a fixed fee on one process?
- What does the agent do on its own, and where does a person approve?
- Is there a log of every action that an auditor or a buyer's diligence team could follow?
- What carries over from the first portfolio company to the second?
Proof
We have not published a private equity engagement. The closest public example of how we build is the AI help desk we built for the Talking News Federation, which is in production: each inbound email becomes a ticket, is classified and routed, and has a reply drafted, and a person approves every reply. Every AI decision is written to a tamper-evident audit trail.
Past delivery work also includes the Royal Navy, the Cabinet Office, Sopra Steria, TPX Impact and the British Wireless for the Blind Fund.
Frequently Asked Questions
- Do you work with the private equity firm or with the portfolio company?
- Either. The work happens inside a portfolio company, with the people who run the process. The firm is usually involved in choosing where to start and in deciding whether to repeat it elsewhere.
- Where should a portfolio start?
- With one company and one process that has a clear owner and a measurable cost, often in finance. Starting with a portfolio-wide program before anything is proven is the most common way these efforts stall.
- Do you only install Pegasus, or do you build custom agents?
- Both. Where a packaged Pegasus agent already covers the workflow we use it, because it is faster. Where the workflow is specific to the company we build a custom agent.
- Which countries do you work in?
- The United States and the United Kingdom. New Odyssey is headquartered in London with US operations in Dallas, Texas.
- How much does it cost?
- We do not publish prices. Every engagement starts with a fixed fee for a defined scope on one process, and the next stage is priced from what that shows.
Keep exploring
Related reading
- AI implementation partners for private equity: US firms comparedWho does what for US portfolio companies.
- AI implementation partners for private equity: UK firms comparedWho does what for UK portfolio companies.
- AI consultancyHow we find where AI belongs and build it.
- Custom AI agentsAgents built for a workflow of your own.
- PegasusPackaged AI agents for common operations workflows.
- Finance automationThe month-end close and finance operations.
Next step
Start with one portfolio company
Tell us which company and which process. We will say plainly whether it is worth building.