AI agents are beginning to change business operations by carrying out sequences of steps through connected tools and workflows, rather than only generating a response to a prompt. In 2026, the clearest evidence points to growing use in selected multi-stage work—not a wholesale shift to autonomous, end-to-end operations. For leaders, the practical questions are where agents are being used, what it takes to scale them, and how to tell reported promise from measured business value.
What is changing in business operations?
Operations are moving, in some organizations, from isolated AI assistance toward repeatable workflows in which an agent can gather information across steps, prepare an analysis or report, or participate in an internal process. That is a meaningful change in how work is organized, but it does not mean an agent is independently running a whole department.
As an Amazon Associate I earn from qualifying purchases.
OpenAI’s 2025 State of Enterprise AI report describes enterprise use becoming more deeply integrated into repeatable, multi-step workflows across functions and business units. The report draws on a vendor-produced survey of 9,000 workers across almost 100 enterprises, so it is evidence of reported usage patterns rather than an independent audit of business outcomes.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallMulti-stage work is more common than end-to-end deployment
Anthropic’s 2026 State of AI Agents report, based on a survey of more than 500 technical leaders in late 2025, distinguishes between agents used for multiple steps and agents deployed across teams or functions. These are survey responses, not a census of businesses.
#1 Best Overall
| Reported deployment scope | Survey finding | What the measure indicates |
|---|---|---|
| Multi-stage workflows | 57% of surveyed organizations said they use agents for multi-stage workflows, according to Anthropic’s late-2025 survey of more than 500 technical leaders, published in its 2026 report. | Agents are being used for sequences of work within a process. |
| Cross-functional or end-to-end processes | 16% of surveyed organizations said agents are used in processes spanning multiple teams or business functions, in the same Anthropic survey. | Use across organizational boundaries is less widely reported than multi-stage use. |
The distinction matters: an agent that performs several steps inside one bounded workflow is not equivalent to an agent coordinating work across multiple teams. A pilot can be useful at either level, but the second raises more questions about access, ownership, handoffs and accountability.
Where are organizations using or expecting to use agents?
Anthropic’s 2026 report identifies data analysis and report generation, as well as internal process automation, among the most impactful agent use cases beyond coding. These percentages describe respondents’ reported assessments; they are not independently measured savings or productivity improvements.
| Use case beyond coding | Reported measure |
|---|---|
| Data analysis and report generation | 60% of respondents identified this as an impactful use case beyond coding in Anthropic’s late-2025 survey of more than 500 technical leaders, reported in 2026. |
| Internal process automation | 48% identified this as an impactful use case beyond coding in the same survey. |
The task pattern is more important than the label: an agent may collect information from several steps, prepare a report, or take part in a repeatable internal workflow. The survey does not establish that every organization has deployed these patterns successfully or achieved a specific financial return.
Free tools Windows power users keep installed
One-click scans. No signup required.
Near-term expectations span several functions
Anthropic also asked about expected near-term impact. These are expectations reported in its late-2025 survey—not proof that the impact has already occurred.
| Area | Expected near-term impact reported |
|---|---|
| Software development | 57% of respondents expected near-term impact, according to Anthropic’s 2026 report based on a late-2025 survey of more than 500 technical leaders. |
| Customer service | 55% expected near-term impact in the same survey. |
| Marketing and sales | 46% expected near-term impact in the same survey. |
| Supply chain, logistics and operations | 44% expected near-term impact in the same survey. |
The spread suggests that interest is not confined to software teams. It does not, by itself, show that agents are already delivering the same level of value across these functions.
What makes scaling agents difficult?
Anthropic’s survey respondents identified technical and organizational requirements alongside the potential use cases. The reported issues show why a promising demonstration may not translate directly into a dependable production workflow.
| Reported scaling issue | Survey finding | Operational implication |
|---|---|---|
| Integration challenges | 46% of respondents identified integration challenges in Anthropic’s late-2025 survey of more than 500 technical leaders, published in 2026. | The agent needs the right tools and system access to complete the workflow. |
| Data quality requirements | 42% identified data quality requirements in the same survey. | Unreliable, incomplete or inconsistent records can undermine the task. |
| Change management needs | 39% identified change management needs in the same survey. | Employees need clear roles, training and a way to handle exceptions. |
Check workflow scope
Define whether the proposal covers one bounded task, several steps within a process, or work spanning teams. Those scopes create different operational demands. Set a clear boundary for what the agent may do and specify where a person must review, approve or take over.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Map integrations and permissions
List the systems and tools the workflow actually requires, then confirm that access is both sufficient and limited to the task. An agent connected to business systems can act on information or trigger steps; permissions should reflect the consequences of those actions, not just the convenience of a demo.
Rank #3
Validate the inputs
Check whether records, documents and other context are accurate, current and consistent enough for the intended work. Decide how the process responds when information is missing or conflicting instead of assuming the agent can resolve every data problem.
Assign human ownership and exception handling
Name the person or team accountable for the workflow, output review and escalation. Specify what happens when the agent is uncertain, encounters an unusual case, or produces an action that cannot proceed. Prepare staff for changed responsibilities as well as for using the software.
Is agent governance keeping pace?
Deloitte’s 2026 State of AI in the Enterprise release reports that 21% of surveyed companies said they had a mature model for agent governance. Deloitte’s survey was conducted in August–September 2025 and covered 3,235 business and IT leaders in 24 countries and six industries. This is a self-reported maturity measure; it does not mean the other 79% had no controls.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesFor an operational review, examine whether permissions, oversight and accountability match the agent’s actual actions. Governance should make clear who can authorize access, who monitors the workflow, who is responsible for errors or exceptions, and how the process is reviewed as it changes. A maturity percentage is a useful signal that governance deserves attention, not a substitute for checking a particular organization’s controls.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Do adoption figures prove business value?
No. The cited findings mostly describe survey respondents’ usage, expectations or assessments. They do not establish a universal productivity gain, cost reduction, revenue effect or headcount change caused by agents. A reported use case or expected impact should not be treated as a measured result.
McKinsey’s 2026 State of AI Global Survey provides broader context: organizations are scaling AI across more functions, but about 6% of respondents qualified as AI high performers under McKinsey’s stated definition. That figure concerns AI broadly, not agents specifically, and should not be used as an estimate of agent success.
Before scaling a pilot, define the operational result that would count as success and compare it with a relevant baseline. Track the quality and completion of the work, the time or effort involved, the frequency of human intervention, and the handling of exceptions. Separate those observed measures from employee or leader perceptions and from forecasts of future impact.
How should leaders decide what to scale?
Use a staged decision rather than treating “agent adoption” as a single threshold. A workflow is a stronger candidate when its scope is clear, the necessary data and integrations are dependable, and a named owner can review performance and handle exceptions.
- Start with a bounded process: choose work with a defined beginning, expected output and escalation path.
- Set action limits: determine which steps the agent can complete, which require approval, and which remain human-owned.
- Test the operating conditions: evaluate representative inputs, missing information and exceptions—not only a clean demonstration case.
- Measure outcomes: compare actual workflow results with a baseline, and distinguish measured performance from survey-based expectations.
- Expand only when ready: increase scope or connect more teams only after access, data quality, review ownership and governance are appropriate for the added risk.
The 2026 evidence supports a measured conclusion: agents are entering multi-step business workflows, while cross-functional use is less frequently reported and scaling still depends on integration, reliable data, employee change and governance. The case for wider deployment should rest on results observed in the specific workflow, not on adoption percentages alone.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
