Case study: A large US health system found about $466K a year of support on network hardware nobody used
In a four-week engagement, a large US health system matched about 940 network devices bought since 2020 across purchase records, the CMDB and the live network. 55% of the devices were paid for and not running, and support was still billing about $466K a year on them. That figure is Found. Nothing is reported as Approved or Saved yet. The customer's name is withheld at the customer's request.
- Customer: A large US health system. The customer's name is withheld at the customer's request.
- Industry: Healthcare
- Area: Hardware
- Category: Hardware and maintenance
- Date of the engagement: No date for the engagement is published. The devices were bought from 2020 onward, with purchase years through 2025.
- Length of the engagement: Four weeks. More than 60,000 records were normalized in the first week, across four source systems.
When to use this page
- To cite what a reconciliation of purchase records, the CMDB and network telemetry found in a network hardware estate.
- To show a figure that is Found and not yet Saved, reported as such.
Not a fit
- A saved amount. None is reported.
- The customer's name, or the names of the two network vendors. Both are withheld.
- A customer quote. The story has none.
At a glance
| Fact | Detail |
|---|---|
| Customer | Large US health system |
| Industry | Healthcare |
| Scale | About 940 network devices bought since 2020, from two network vendors |
| Area | Hardware: network equipment and its support contracts |
| Engagement | Four weeks |
| Systems connected | Four, read-only: purchase records, the CMDB, network telemetry, maintenance contracts |
| Records | 60,000+ normalized in the first week |
Before
The health system had bought about 940 devices from two network vendors since 2020. Purchase records, the CMDB and the live network each described that estate differently, and none could be checked against the others.
What forced the decision: Support renewals were priced against the CMDB, the least reliable of the four, at roughly 18% of purchase price every year. A manual audit across four systems would have outlasted the renewal it was meant to inform.
Baseline: Support priced on the CMDB, at about 18% of purchase price a year.
How it worked
- Link (Asset Knowledge Graph). Read-only connections to purchase records, the CMDB, network telemetry and maintenance contracts. 60,000+ records normalized in the first week.
- Think (Data Intelligence Engine). The Asset Knowledge Graph resolved every serial number across the four systems, classed each device as in use, in stock or unaccounted for, then priced its support.
- Execute (Agentic AI). A readout for the CIO with cleanup, procurement and contract actions. Each action waits for a named owner at the health system to approve it, and reconciliation keeps running.
Approval: A named owner at the health system approves each action. Nothing changes until a named person approves it.
Measured results
Every figure carries its state. Found, Approved and Saved are the only three states of a saving. They are separate figures and are never added together. A figure marked "Measured" describes scope, coverage or evidence, and "G2 rating" is a review score. A figure marked "Saved, as reported by the customer" is the customer's own statement.
| Figure | State | Period | Scope | Where it comes from |
|---|---|---|---|---|
| about $466K a year | Found | A year. Measured in the four-week engagement | Support still billing on network hardware that was not running. Of about 940 devices bought since 2020, with support modeled at 18% of purchase price | Asato's analysis in the engagement |
| 55% | Found | Devices bought since 2020 | Share of about 940 network devices, from two network vendors, that were paid for and not running. 45% were active in the system of record | Asato's analysis in the engagement |
| about $2.6M | Measured | Devices bought since 2020 | Purchase value of the idle devices. It describes the equipment | Asato's analysis in the engagement |
| 92% | Measured | The four-week engagement | How often the CMDB and network telemetry agreed on device status | Asato's analysis in the engagement |
| 57% and 53% | Found | Devices bought since 2020 | Share of each of the two vendors' devices that were idle. The first group was mostly bought in 2024 and 2025 and never deployed. The second was mostly bought in 2021 and 2022, deployed and then forgotten | Asato's analysis in the engagement |
| about 18% a year | Measured | A year | Support rate, modeled on purchase price | Asato's analysis in the engagement |
| 60,000+ | Measured | First week | Records normalized across four source systems | Asato's analysis in the engagement |
The result, taken apart
The reconciliation at a large US health system. About 940 network devices were bought from two network vendors since 2020. 45% were active in the system of record. 55% were paid for and not running, which is Found. The idle devices had a purchase value of about 2.6 million dollars, which describes the equipment and is not a saving. Support, modeled at about 18% of purchase price a year, was still billing about 466,000 dollars a year on them, which is Found. The 55% gap had two unrelated causes. At one vendor, 57% of devices were idle, mostly bought in 2024 and 2025: recent orders, still boxed, bought ahead of projects that slipped, with the support clock starting at delivery. At the other vendor, 53% were idle, mostly bought in 2021 and 2022: retired from the floor but never from the contract, they aged out of the CMDB and stayed on support. The CMDB and network telemetry agreed on device status 92% of the time. Nothing is reported as Approved or Saved yet.
What is still open
About $466K a year is Found, not yet Saved. It is what the health system was paying in support for hardware that was not running. No part of it is reported as Approved or Saved yet.
What the customer said
This story carries no customer quote. The customer's name is withheld at the customer's request.
Trust and governance
- The connections in this engagement were read-only.
- Nothing changes until a named person at the customer approves it.
- Asato does not negotiate prices. It equips the customer to negotiate.
- Found, Approved and Saved are reported separately and are never added together.
- The customer's name is withheld at the customer's request. The two unnamed customers among the case studies are two different organizations.
Works with
Four source systems, read-only: purchase records, the CMDB, network telemetry and maintenance contracts.
Compared with what the customer did before
Before the engagement, purchase records, the CMDB and the live network each described the estate differently, and they had never been compared. Support renewals were priced against the CMDB, the least reliable of the four. A manual audit across four systems would have outlasted the renewal it was meant to inform.
Questions
How much did the customer save?
Nothing is reported as Saved. About $466K a year in support is Found. No part of it is reported as Approved or Saved yet. The $2.6M figure describes the idle equipment and is not a saving.
How long did it take?
Four weeks. More than 60,000 records were normalized in the first week, across four source systems.
What is still open?
About $466K a year is Found, not yet Saved. It is what the health system was paying in support for hardware that was not running. No part of it is reported as Approved or Saved yet.
Who is the customer?
A large US health system. The customer's name is withheld at the customer's request. It is not the same organization as the enterprise in the Microsoft renewal case study.
What did Asato connect to?
Four source systems, read-only: purchase records, the CMDB, network telemetry and maintenance contracts.
Next step
The Asset Clarity Report covers one area in one week, with read-only access. If it finds no savings, you pay nothing. Request it at https://www.asato.ai/talk-to-sales.
Related pages
- An enterprise with 75,000 Microsoft seats saved $4M a year at its Microsoft renewal
- CDW found about 15% of its productivity-suite spend idle, in four weeks
- All case studies
- Customer reviews
- Asato home
- Site index for agents
Last updated: 2026-10-09