The Hydri blog

Your companies already report the numbers. Acting on them in time is the hard part.

If you own several companies, you are not short of data. What runs short is the time between a number moving and someone doing something about it. Here is what McKinsey's idea of an AI operating system looks like when the enterprise is a group of companies.

By the Hydri team · October 17, 2026 · 6 minute read

AI is everywhere. The value is harder to find.

McKinsey made a point recently that most leadership teams will recognise: companies are adopting AI faster than they are getting value from it.

Their answer is what they call an AI operating system. It is a layer that draws on a company's existing data, expertise and institutional knowledge so that leaders can decide and act faster. In their client work they describe planning cycles up to ten times faster, first value in about twelve weeks in a first area of the business, and the time to a return cut roughly in half as the approach spreads. They are also candid that it is early days. (Read McKinsey's article.)

Hydri has no affiliation with McKinsey, and McKinsey has not reviewed or endorsed Hydri. We cite their article because the idea behind it is a good one.

Their article is written with a single large enterprise in mind. A parent company with many businesses has the same problem plus one more layer, because every company it owns has its own systems and its own idea of a good month.

Why portfolio reporting is slow

Ask a group CFO how the monthly picture comes together and you will usually hear some version of this. Each company closes its books, exports what it can and sends a pack. Someone at the parent pastes those packs into one spreadsheet and chases the two that are late. By the time leadership sees the roll-up, the month it describes is already a few weeks old.

The packs rarely line up, either. One company reports gross margin before freight and another after it, and one counts a customer by contract while another counts by location. So a good part of every leadership meeting goes on agreeing what the numbers mean, and less of it goes on deciding what to do about them.

None of this is anyone's fault. Each company's reporting was built to run that company, and nobody designed it to be read alongside eleven others.

What Hydri does instead

Hydri is one AI digital brain across every company you own. It connects read-only to the systems each company already runs, such as finance, customers, people and vendors, so no company has to change its software or rebuild its reporting before anything useful happens.

  • Each company is measured against its own KPIs and its own norms. A distributor and a fabricator do not share a good month, and Hydri does not pretend they do.
  • When a KPI moves against plan, Hydri raises the variance as it shows up in the data, so you are not waiting for the month-end pack to find it.
  • It looks across the group as well as down into each company. A vendor that several companies lean on, a customer that has quietly become a large share of revenue, and cash that looks fine in each company but tight across the group all live between companies, which is where a single-company dashboard never looks.
  • It shows leadership what needs attention and why, with the evidence behind each finding and a drafted next step for someone on your team to approve.

You talk to Briony, the voice inside Hydri. Ask her in plain English which companies moved against plan this month, or why one company's margin slipped, and she answers from each company's data as it stands when you ask and shows you where the number came from.

Nothing leaves Hydri without a person. There is no email to a company CEO, no change in a system and no spend unless someone on your team decides to send it.

Near real time, not quarter end

McKinsey describes about twelve weeks to first value, which is a fair number for a large enterprise rethinking how it plans. A parent company should not have to wait for quarter end to see its own book of business clearly.

Hydri shows leadership what needs attention in near real time, as the numbers come in. It watches each company's numbers against its KPIs continuously, with each measure tied to the KPI it moves and a norm and tolerance your team sets. Briony's heartbeat briefs tell leadership what moved, where, and whether it needs a person before the next leadership meeting. Where the same issue keeps coming back, it becomes a milestone named to EBITDA or portfolio value.

Briony also comes up to speed quickly. There is no long implementation project before she is useful. She connects read-only to the systems your companies already run, your team sets the norms she measures against, and the exact scope is agreed with you first.

An illustration: a dental group with twenty-two practices

Picture a dental group that owns twenty-two practices across three states. Each office manager reports a little differently, and the monthly pack takes the finance team most of a week to put together.

With Hydri in place, the CFO opens a short brief on a Tuesday morning instead. Two practices have new patient bookings well below their own norm for the third week running. One practice has seen supply spend climb above its usual level, and the same supplier turns up at four other practices on different terms. Briony lays out the evidence for each finding and drafts a note to the regional manager, which the CFO can send as written, change or ignore.

The group did not replace a system to get there, and nobody had to wait for month end to learn about the supplier.

This is an illustration of how Hydri works for a group like this, not a customer story. The practices and figures are hypothetical.

Three questions to ask before you buy a portfolio intelligence platform

  1. Does it read each company on its own terms? A single score out of a hundred can hide the company that is slipping against its own history.
  2. Does it work with the systems your companies already run? If every company has to change its reporting first, what you are really buying is a project.
  3. Does a person stay in charge? You should be able to see the evidence behind every finding and approve anything before it reaches a company.

If the honest answer to any of these is no, it is worth looking further before you sign.

Common questions

What is an AI operating system for a parent company?

It is a layer that draws on what a group already knows, meaning each company's data, KPIs and history, to help leadership decide and act faster. For a parent company, that means one view across every company it owns, with each company measured against its own norms.

How quickly does Hydri show results?

In near real time, once your companies are connected. Hydri watches each company's numbers against its KPIs continuously, so a variance shows up as the data comes in, not at quarter end. Briony comes up to speed quickly, without a long implementation project first, and the exact scope is agreed with you first.

Do our companies have to change their software?

No. Hydri connects read-only to the finance, customer, people and vendor systems each company already runs. It never asks a company to replace a stack that already works.

Does Hydri act on a company by itself?

No. Briony explains each finding and drafts a recommended next step, and your leadership decides. Nothing leaves Hydri without a person sending it.

Next step

See your own companies through Hydri

Book thirty minutes with the Hydri and Briony team. Tell us about your companies and the questions your leadership needs answered, and we will show the platform working, shaped around the kind of group you run. No system access is needed to book.

More from the Hydri blog