The CIO’s AI strategy is no longer only a question of which application to buy. It is also a question of who or what should perform the work—and what authority that worker should have. Rajjie Sarmey, writing for CIO, proposes Enterprise Work Architecture (EWA) as a way to redesign workflows across people, AI agents, applications, data, and delegated authority. It is a proposed approach, not an established industry standard.
Why application modernization is not enough
Employees often bridge gaps between applications and organizational teams. Replacing or upgrading one system may improve that system without fixing the work that crosses several of them. Sarmey illustrates this with a billing dispute: an invoice does not match a delivery, and resolving it may involve CRM, ERP, fulfillment, contracts, finance, and operations. People can end up gathering evidence, reconciling records, and routing approvals across those boundaries.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
MINISFORUM MS-02 Ultra Workstation Mini PC, Intel Core Ultra 9 285HX (24C/24T, up to 5.5GHz), PCIe... | $1,659.00 | Buy on Amazon |
| 2 |
|
GMKtec EVO-X2 AI Mini PC Ryzen Al Max+ 395 Superchip 128GB LPDDR5X 2TB SSD | $3,649.99 | Buy on Amazon |
Adding automation to the existing sequence can make an inefficient process run faster without making it better. As Sarmey puts it, “Automation without work redesign can turn process debt into machine-speed process debt.” His proposal is to start with the business outcome, then redesign the work and its controls before deciding where agents belong. CIO’s article on Enterprise Work Architecture
How should CIOs redesign work for AI agents?
Sarmey’s EWA sequence is Outcome → Work → Authority → Execution → Evidence → Economics. Each step constrains the next: first define what success means, then design the tasks and permissions needed to reach it.
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1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitches#1 Best Overall
- High-Performance AI Processor：The MS-02 Ultra features an Intel Core Ultra 9 285HX (24C/24T, up to 5.5 GHz, 13 TOPS NPU), delivering fast and efficient performance for AI inference, algorithm development, and media workloads. A PCIe x16 expansion slot supports desktop-class GPU upgrades for advanced model training and accelerated computing tasks. It's ideal for creators, engineers, and teams handling intensive parallel workloads.
- 4 × M.2 PCIe 4.0 + 4 × DDR5 SODIMM slots:Four DDR5 SODIMM slots support up to 256 GB of memory, while ECC helps maintain data integrity in mission-critical environments. Four PCIe 4.0 M.2 slots support up to 24 TB of storage, supporting RAID 0/1/5/10, combining high-speed performance with data protection. It allows for the creation of independent scratch disks, media libraries, and project drives, providing high-throughput for production workflows.
- PCIe & USB 4.0 v2: Up to three PCIe slots can be equipped, including a dual-slot x16 GPU. The main slot supports PCIe 5.0, meeting the needs of high-bandwidth creative and computing workloads. USB 4.0 v2 (80Gbps) supports high-bandwidth external storage and displays.
- Ultra-fast Networking: Wi-Fi 7 further enhances wireless performance with next-generation speeds and low-latency stability. Intelligent bandwidth switching optimizes throughput in different network environments, ensuring optimal performance for enterprise or local networks. Dual 25GbE ports (providing up to approximately 3.125 GB/s bandwidth, about 25 times faster than traditional 1GbE), enabling seamless large-scale file transfers and parallel computing. 10GbE and 2.5GbE ports, with support for Intel vPro technology, ensure enterprise-grade remote management and deployment flexibility.
- Server-grade thermal architecture: Utilizing a dedicated CPU/GPU airflow design, equipped with a 6-pipe dual-fan cooler, it maintains stable performance even under sustained loads, delivering up to 140W Turbo power while maintaining a 100W TDP, and operating with noise levels as low as 36 dB. An integrated 350W power supply ensures stable and reliable output for demanding computing tasks and fully loaded extended configurations.
- Outcome: Name the measurable business result, such as resolving a billing dispute accurately and promptly.
- Work: Break the process into tasks. Decide which are necessary, which can be removed, and where human judgment is important.
- Authority: Define what a person, application, or agent may retrieve, interpret, recommend, prepare, approve, execute, or escalate. Do not treat technical capability as permission: “Capability cannot silently become authority,” Sarmey writes.
- Execution: Map the participating systems, APIs, data, and workflow paths. Keep systems of record responsible for authoritative data and transactional controls, even if employees interact with them less directly.
- Evidence: Make consequential actions observable and reconstructable. Preserve enough information to challenge a decision, identify an intervention, and recover from an error.
- Economics: Assess whether the redesigned work improved time, cost, quality, risk, or experience—not just whether an AI feature was deployed.
What the billing-dispute redesign could look like
In Sarmey’s illustrative scenario, machines collect and reconcile evidence about the invoice and delivery. An agent can work within defined thresholds, while ambiguous or consequential judgments go to a person. Exceptions are escalated, and governed APIs carry approved actions into the systems that hold the authoritative records. The decision, the authority under which it was made, and any human intervention remain traceable.
This is an example of a proposed design, not a reported case study or evidence of measured savings. For a real workflow, establish a baseline and compare it with the redesigned process using measures such as resolution time, human touches, exceptions, rework, cost, and customer impact.
What controls should enterprise AI agents have?
When an agent can act across application boundaries, the design needs to specify both its identity and its limits. Sarmey recommends considering controls such as:
- Identity and delegated permissions: Know which agent is acting, on whose behalf, and what access it has been granted.
- Purpose and data boundaries: Limit use of data to the task and scope for which access was authorized.
- Transaction limits: Set thresholds on actions, amounts, or records an agent may change.
- Segregation of duties: Prevent a single agent or workflow from bypassing required independent review.
- Observability and evidence: Record actions and the authority used so that consequential decisions can be examined.
- Escalation and lifecycle controls: Define when work must go to a person, and how permissions and agents are changed or withdrawn.
Authority should be proportional to consequence. Retrieving an account balance is different from changing a customer record; preparing a purchase order is different from releasing it. A system may technically support all of these actions, but that does not mean an agent should be allowed to perform them without review.
What the AI-agent adoption forecast does—and does not—say
Gartner’s newsroom release, published August 26, 2025 and updated September 5, 2025, forecast that 40% of enterprise applications would be integrated with task-specific AI agents by the end of 2026, compared with less than 5% at the time of the forecast. Gartner also forecast agent ecosystems spanning applications and business functions by 2028. These are forecasts, not measured outcomes for 2026 or proof that organizations have redesigned their work effectively. Gartner’s forecast and methodology context
The adoption projection makes workflow design more urgent, but integration alone is not a measure of value. An agent connected to an application may still be limited to recommendations or preparation; execution requires an explicit grant of authority and appropriate controls. Gartner Senior Director Analyst Anushree Verma described the expected progression this way: “AI agents will evolve rapidly, progressing from task and application specific agents to agentic ecosystems.”
Rank #2
- EVOLUTION RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 128GB pool, which is perfect for running LLMs such as Deepseek 70B Q8, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 12% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
How to measure redesigned work
Sarmey recommends tracking the outcome of the work rather than counting agents, prompts, or licenses. These are management measures to select and adapt for a workflow, not a validated universal benchmark.
- Time-to-outcome: How long it takes to reach the intended result.
- Human touches: How many handoffs or manual interventions the workflow requires.
- Exceptions and rework: How often cases leave the standard path or need correction.
- Cost and control overhead: The cost per governed outcome, including the effort needed to supervise and control the process.
- Quality, risk, and experience: Whether accuracy, exposure to harm, and customer or employee experience improved.
Use the same definitions and boundaries when comparing a baseline with a redesigned workflow. Otherwise, a shorter processing time could conceal more errors, exceptions, or downstream work.
Free tools Windows power users keep installed
One-click scans. No signup required.
Who needs to own the redesign?
Cross-application work rarely belongs to IT alone. Sarmey describes responsibility as spanning business, operations, finance, security, risk and legal, HR, audit, and enterprise architecture. A practical starting point is to inventory the work, then locate cross-system handoffs, exceptions, approval delays, and repetitive effort. The teams that own those steps can determine where redesign is appropriate and who must be accountable for the resulting decisions.
Sarmey also points to the NIST AI Risk Management Framework and its functions—Govern, Map, Measure, and Manage—as a reference for risk work. The framework reference is part of his article’s guidance; it does not establish that a particular EWA implementation is compliant or effective.
The strategic shift for CIOs
The decision is not simply whether to fund more AI features. It is whether to redesign how the enterprise operates: which tasks should disappear, which need human judgment, what agents may do, and how the organization will know whether the result is better. EWA offers Sarmey’s sequence for asking those questions while retaining accountable people and authoritative systems in the workflow.
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.
