“The next era of enterprise AI is all about delivering software that increasingly does the work, while people define the objectives and set the policies.”
Redeployed is a weekly newsletter that breaks down one important AI story at a time for product builders and engineering leaders. Every issue explores what the shift means for technology companies and how leaders can respond.
Oracle is introducing a new way for employees to work with its Fusion Cloud applications. With Fusion Companion and Fusion Agentic Home, users can describe a business objective while the software coordinates the transactions, approvals, and specialized agents needed to carry it out.
The October 26 announcement builds on Oracle's Fusion Claw runtime and 25 agentic applications introduced in September. Together, they represent an effort to move enterprise software beyond providing tools that employees operate and toward systems that execute business workflows.
An employee could eventually complete a process without knowing which application to open, which screens to navigate, or which individual transactions need to happen along the way.
For software companies, that raises a consequential question: what happens to your product's competitive advantage when an agent becomes its primary operator?
Your Moat Moves Below the Screen
For years, software companies have competed through better navigation, cleaner dashboards, faster setup, and more intuitive workflows. Those qualities remain important, but their role may change when an agent becomes the primary operator of the underlying system.
If an AI system can update an account, compare suppliers, or move a workflow forward through APIs and authorized actions, visual usability alone may no longer be enough to differentiate the product.
Durable value could move deeper into the stack. Proprietary business data gives agents useful context. Transaction systems allow them to change important records. Reliable APIs let them act without depending on fragile browser automation, while permissions and audit records establish what they can do and how their actions can be inspected.
Oracle's architecture makes this direction explicit. Fusion Claw places agent execution inside what the company calls an Enterprise Operating Envelope, which defines policies, permissions, risk thresholds, approval requirements, and escalation boundaries. It also generates an Outcome Receipt documenting the evidence, decisions, and transactions involved.
That governance layer may become as strategically important as the interface sitting above it.
The Agent Could Own the Customer Relationship
Oracle's Fusion Companion is designed to work through external AI tools such as ChatGPT, not only through Oracle's own interface. It can also coordinate authorized actions across other enterprise systems.
This creates a potential strategic risk for every application underneath the agent layer.
If employees begin their work inside ChatGPT, an enterprise assistant, or another general-purpose interface, that assistant could become the product they interact with most. The applications executing the transactions would remain essential, but they could become less visible to the people using them.
For SaaS companies, this raises questions about who controls the customer relationship. A product may continue delivering critical business functionality while another company owns the interface through which users discover, request, and manage that functionality.
Products with deep data, execution rights, and embedded business logic may remain difficult to replace. Products differentiated primarily by packaging common workflows into a cleaner user experience could face greater pressure.
None of this means traditional SaaS interfaces are about to disappear. Oracle has announced capabilities and a product direction, but it has not provided independent customer results demonstrating that employees are abandoning its applications in favor of agent-driven experiences.
The strategic question is what happens if that behavior becomes more common.
Seat Pricing Meets Agent Usage
If agents begin performing work that previously required employees to log into software, seat-based pricing could come under pressure.
An agent might execute more transactions than a human user while requiring fewer named accounts. That could weaken seat growth even as the operational value delivered by the product increases.
Software companies may respond by charging for agent access, transaction volume, completed workflows, or usage of proprietary data and business logic. Some vendors could generate additional revenue from agent activity while continuing to charge for human seats.
Outcome-based pricing is another possibility, although it introduces measurement and accountability challenges. A system can complete a workflow without producing the business result a customer expected.
Vendors will need clear definitions of what the software controls, what the customer controls, and how value is measured. The transition will likely produce hybrid pricing models before it produces widespread outcome-based contracts.
For SaaS leaders, the immediate question is whether their pricing model reflects the value their product creates when software performs more of the work.
The Interface Changes Jobs
Enterprise software has historically been organized around human operators. An employee opens an ERP system to reconcile transactions, enters an HCM application to update a workforce plan, or navigates a CRM to prepare an account briefing. The software stores the data and provides the tools, while the employee carries the process from one screen to the next.
Oracle is proposing a different relationship. Its Ledger application can examine accounting entries, investigate anomalies, and prepare authorized actions. Workforce Orchestration can compare hiring, development, redeployment, and external talent options before coordinating an approved staffing plan. Procurement Spend can identify contract leakage, calculate potential savings, and recommend corrective action.
The interface does not disappear, but its purpose changes. Instead of guiding users through every step, it becomes the place where people define objectives, establish authority, review exceptions, approve consequential decisions, and inspect what the system did.
Oracle describes this as a move from a “system of record” to a “system of outcomes.” The phrase is convenient marketing, but the product direction behind it is meaningful. Enterprise applications are beginning to compete on whether they can complete work, rather than simply organize it.
For product teams, that changes what a good user experience looks like. The important interactions may increasingly involve setting a goal, understanding an agent's proposed plan, intervening when necessary, and confirming that the right outcome was achieved.
The Product Roadmap Changes
Technology companies should examine how their products behave when the operator is software.
APIs built as secondary integrations may need to become first-class product surfaces. Permissions must account for agents acting on behalf of people, including limits on data access and transaction authority. Systems also need reliable records of what an agent intended to do, which evidence it used, what approvals it received, and which actions it completed.
These capabilities become especially important when an agent coordinates work across multiple applications. A workflow may involve retrieving information from one system, making a decision based on another, and updating records in a third. Engineering teams need to ensure those actions remain authorized, observable, and recoverable when something goes wrong.
Product design must evolve alongside the architecture. Teams will need interfaces for defining objectives, reviewing proposed actions, approving high-impact decisions, investigating exceptions, and interrupting execution when necessary.
A useful starting point is to examine one important workflow and ask whether an agent could complete it through supported APIs and permissions without relying on a person to navigate the product.
If the answer is no, the gap may reveal which parts of the product need attention first.
This issue of Redeployed is brought to you by Tecla Labs: The difficult part is rarely deciding that an agent would be useful. It is identifying the workflow, data, approval boundaries, and operating owner required to make it dependable. Tecla Labs maps how work moves through a company, identifies where AI can produce measurable value, and defines the first system worth building. Start with a free AI Opportunity Assessment to explore a prioritized roadmap grounded in your workflows and data.
If an Agent Used Your Product Tomorrow
The question for software leaders is what would remain valuable if an agent became the primary operator of their product.
Could that agent retrieve the right context, execute an authorized action, handle an exception, and produce an auditable record without depending on a person to navigate the interface?
Those capabilities are becoming important product decisions, even for companies whose customers still work primarily through traditional screens. The interface will continue to matter, particularly when people need to supervise decisions and understand what the software has done. But the execution layer underneath it may increasingly determine whether the product remains essential.
If your strongest advantage is that your screens are easier to navigate, Oracle just offered a reason to reconsider the roadmap.
Connect With Other Technology Leaders
If you want to exchange practical ideas with senior technology and product leaders navigating the AI era, check out the upcoming GILD Forums. They bring together experienced operators for peer discussions around the technology and business decisions they are working through right now.
More to come…
– Gino Ferrand, Founder @ Tecla Labs