Automic Workload Automation licensing. Cut the counted base before Broadcom quotes the increase.

Automic is licensed on agents, connected systems or successful job executions, so the counted base grows every time a team automates something. Reconcile it and retire what is idle before Broadcom sends the renewal quote, because quantity is easier to cut than price.

How is Automic Workload Automation licensed?

Automic is licensed on what it automates, never on named users. Depending on the paper you signed, Broadcom counts deployed agents and connected systems (nodes), successful job executions, or a blend of the two. Cost follows the number of agents, the execution volume across environments and the breadth of connected systems.

The binding definition sits in your order form, and the product description on Broadcom's website does not override it. Before anyone builds a reduction plan, pull the order form and the Specific Program Documentation (SPD) it references, and write down which metric applies to which line.

What does each Automic metric count?

Broadcom publishes its metric rules in SPDs. Two are in circulation for the core product, and each rewards a different kind of clean up work.

A few items sit outside the main unit. Know which of them you use before you compare quotes.

• Separately priced under Job Execution terms. The BS2000 agent (per server), the Avaloq agent (per Avaloq instance), Automic Omniview (per concurrent user) and Automated System Copy for SAP (per 100 copies a year).
• Analytics on outside data. Automation Analytics and Intelligence processes Automic data at no extra charge, but data from other schedulers costs 1 Job Execution license for every 10 AAI job executions.
• Included under node terms. Unlimited Automation Engines and high availability engines, at no node charge.
• Cloud Automation under node terms. Counted separately, as average daily Cloud Automation task executions divided by 100, so a busy cloud integration adds nodes without adding an agent.

Which metric is on your paper?

If your order form names nodes or agents, the node SPD of that era governs. If it names Job Executions, the Hybrid Cloud terms do. A renewal quote that changes the unit is a metric conversion, and it needs its own sizing exercise before you compare prices.

The two models penalize different habits. Node terms charge for agents and connections left behind on retired systems. Execution terms ignore agent count and charge for busy schedules, including the ones running in test.

Why does the Automic metric base keep growing?

It grows because every new automation adds to it. A team that schedules a new file transfer, connects another database or copies a workflow into a test client adds agents, connections or executions, and no purchase order is raised. By renewal, the counted base is almost always larger than the workload justifies.

The same mechanic makes consolidation the most direct way to cut cost. Fewer agents, environments and schedules mean a smaller number to renew against, and the reduction comes off quantity rather than out of a discount Broadcom has to grant.

How much does a clean agent inventory remove?

In our reviews, reconciling the agent inventory against active workloads cut the metric base by 15 to 30 percent. Every point came off the quantity Broadcom prices, before any price conversation started.

Decommission before the renewal opens, not during it. The lower count is then already in the system, and in the usage data, when Broadcom sizes the quote.

What does a test schedule cost under the execution metric?

Under Job Execution terms, schedule frequency matters more than agent count. A job scheduled every 5 minutes runs 288 times a day, or 8,928 times in a 31 day month. Say a test client carries 20 such jobs that exist only to keep an interface warm: that is 178,560 counted executions a month.

Now say your peak month, for example the one with year end close, reaches 1,150,000 successful executions. Removing those 20 test jobs takes the peak to 971,440, a cut of about 15.5 percent in the licensed quantity, without touching a production job.

• Failed runs. Jobs that end outside states 1900 and 1904 are not counted.
• Reruns. The SPD counts by end state, so a failed run is not counted and the successful rerun is.
• Environments. A job in a test client counts the same as one in production.

How do you check your own Automic usage?

• Telemetry usage data. On current releases, log on to Client 0 and open Administration, then Telemetry, then Usage Data. By default it shows the last full month of successful job executions or node counts, depending on the product you license, keeps 18 months of history, and exports to CSV. The product sends this data to Broadcom, so reconcile it before the account team reads it.
• License Reporting Tool. From versions 12.2.2 and 12.3, license usage no longer appears in the Administration panel. Broadcom points customers to the License Reporting Tool command line utility, version 6.0 or later, on the Automic download site.
• Agent list against the CMDB. Export the agents defined in each Automation Engine system and match every host against your configuration management database and decommission records. An agent on a host that no longer exists is a removal candidate.
• Connection objects. List database and REST or SOAP connection objects and check which ones a job used in the last quarter. Under node terms each one can be a counted node.
• Executions by environment. Split the monthly execution count by system and client so test and development volume sits on its own line, and find the month that sets your peak.

How do you build the usage evidence before Broadcom sends the quote?

Finish the inventory work before the renewal conversation opens. Evidence gathered while the negotiation is running arrives after the first number is on the table, and that number frames every discussion that follows.

Work through the counted items in this order:

• Agents. Find which ran a job in the last quarter, which sit on hosts decommissioned two projects ago, and which are duplicates carried into test environments that no longer need their own runner.
• Environments. An idle environment carries the cost of every agent inside it while delivering nothing, so merging or retiring one removes a block of the base in one step.
• Connectors. Unused ones widen the counted scope without adding value.

This work needs no concession from Broadcom. It takes quantity out of what Broadcom prices, and the account team can check the result in the same usage data it already receives.

Automic is one line in a larger Broadcom relationship. The portfolio view is in our VCF pillar and the Broadcom practice library.

What happened to Automic renewals after Broadcom bought CA?

They became harder. Broadcom completed the CA Technologies acquisition in November 2018, and Automic, which CA had bought earlier, came with it. Renewal terms then hardened as they did across the portfolio: steeper opening increases, more bundling, and a firmer stance on support terms and metric interpretation.

Opening increases of 20 to 40 percent were common in the renewals we reviewed, and they were negotiable downward. What moved them was usage evidence and a credible consolidation plan, presented before the number was set. An increase that looks fixed in the first meeting is being tested against whether you can describe your own usage.

Why model two renewal cycles instead of one?

A concession that expires at the next anniversary is a deferral. It lowers this year's invoice and leaves the next renewal starting from the full number. A rate held for one year against an inflated base usually costs more than a modest rate against a base you have already cut.

Broadcom publishes no Automic list prices, so the figures here are illustrative. Say you license 400 OS agent nodes at $600,000 a year, or $1,500 a node, and Broadcom opens the renewal 30 percent higher. The table compares three outcomes over a 3 year term and into the next renewal, assuming that one also opens 20 percent higher.

The third outcome retires 22.5 percent of the nodes and wins the same rate as the second, at $1,725 a node. Against the price hold it saves $555,750 over the term, and the next opening quote starts $294,300 lower.

What will the Broadcom account team say, and how should you answer?

• "The increase reflects the value of the platform." Ask for the quote broken down by metric line and quantity. Then show your reconciled count and ask for the price on that number.
• "Moving to Job Execution licensing makes agents unlimited." It does. Ask for the proposed execution quantity, the month it was measured in, and whether test volume is inside it, then price both metrics on your own data.
• "The discount depends on the portfolio agreement." Ask for Automic priced on its own and the portfolio priced without Automic. The difference is what the other products actually cost you.
• "Support terms are standard." Say you will compare offers on license plus support over the whole term, and ask for the maintenance uplift and service levels in writing.

Does a Broadcom portfolio bundle lower your Automic cost?

Only when you need the other products. Broadcom often frames Automic inside a wider portfolio deal, which can lower the Automic unit rate while raising total multi year spend. Test each bundle against these questions before you compare rates.

• Do you need each product on its own merits? If you would not buy it separately, its discount saves you nothing.
• What is the net multi year cost? Compare license plus support over the term with your current run rate. The headline rate tells you little.
• What flexibility survives if needs change? Bundled scope is harder to unwind than it is to accept, and exit terms are written at signature or not at all.
• Is support priced in the same conversation? Watch the annual maintenance uplift, any change in service level at renewal, and multi year caps on support increases, because a license concession is routinely recovered through support. Our Broadcom practice covers the enterprise agreement mechanics.

Why we advise against taking the bundle for its blended discount

The usual advice is to accept the portfolio deal because the blended discount looks strong. We disagree. In most Automic renewals we ran, bundles that included products the buyer did not need raised the total while presenting as a rate cut.

A blended discount is measured against list prices for products you were never going to buy, so a large percentage can hide a larger invoice. Price Automic alone on a reconciled count first, then add other products one at a time and keep only those that pass the need test.

Which contract terms should you ask for?

• A written metric definition. Put the SPD definition into the order form, including how test, development and disaster recovery environments are treated. SPDs are revised over time, and the Hybrid Cloud SPD already exists in a December 2024 and an October 2025 version, so name the version.
• A cap on the next renewal. A maximum increase for the following renewal, stated in percent, so the concession you win now does not reset at the next anniversary.
• Multi year caps on support increases. These limit the annual maintenance uplift and should hold service levels for the term.
• A reduction right. The right to drop nodes or execution volume at anniversary when you retire workloads, with the price falling in proportion.
• A defined peak month rule. Under Job Execution terms, agree how a one off spike, such as a migration or a year end rerun, is treated before it sets your licensed quantity.
• Unbundling at renewal. The right to renew Automic without the other portfolio products at the same unit rate.

What have we seen in recent Automic renewals?

Across roughly 15 to 25 Broadcom and Automic reviews we led in 2024 and 2025, post acquisition renewals arrived with a steep opening increase and a portfolio conversation attached. Three patterns came up repeatedly.

1. Opening increases came down. The openings described above were common, and they were often negotiated lower when usage evidence arrived early.
2. Idle and duplicated agents inflated the base. Removing them produced the reductions in the metric base described earlier, before price was discussed.
3. Bundles cut both ways. Portfolio offers helped only when the customer needed the other products on their own.

The order that worked was the same each time. Cut the metric base first by retiring idle agents, duplicate test agents, dead environments and unused connectors. Then judge any bundle on net multi year cost and independent need, and negotiate support alongside the license so a concession cannot come back through an uplift.

A better rate on software you would never have bought is scope you pay to maintain for years.

What to do next

1. This week. Pull the order form and the SPD it references, and confirm whether you are counted on agents, nodes, job executions or a blend.
2. Within a month. Export the telemetry usage data or run the License Reporting Tool, and reconcile the agent list against the CMDB and your decommission records.
3. Before the renewal opens. Retire idle agents, duplicate test agents, dead environments and unused connectors, and lengthen short interval test schedules if you are on execution terms.
4. Three months out. Model the next two renewal cycles so a concession that expires at the following anniversary is recognized as a deferral.
5. When the quote arrives. Judge any portfolio bundle on net multi year cost and independent need, and ask for Automic priced on its own.
6. Before signature. Negotiate support and maintenance with the license, covering the annual uplift, service levels at renewal and multi year caps. Our Broadcom team can run the inventory and the renewal with you.

Frequently asked questions