A leadership inflection point from copilots to autonomous work
Enterprise AI has entered its “everywhere, but uneven” phase. In McKinsey & Company’s 2025 global survey, 88% of respondents report “regular AI use” in at least one business function, yet the majority are still in experimenting or piloting phases, with only about one‑third saying their organisations have begun scaling AI programmes.
At the same time, leaders are moving beyond copilots into AI agents—systems that can plan and execute multi‑step work. McKinsey defines AI agents as systems based on foundation models that can act in the real world by planning and executing steps in a workflow, and notes 23% of respondents report scaling an agentic AI system somewhere in their enterprise, with an additional 39% experimenting with AI agents.
Independent research points in the same direction. In a 2025 Boston Consulting Group brief, BCG reports that (in a study conducted with MIT Sloan Management Review) 35% of organizations say they are already using agentic AI and another 44% plan to do so soon.
This shift creates a new executive challenge: deploying agents is not merely “adding AI.” It is designing an enterprise environment where AI can reliably do work—with clear boundaries, trusted context, and observable outcomes. That discipline is Agent Experience (AX).
1. Why AI adoption is not translating into enterprise value
If this feels like a paradox—high adoption but low measurable impact—you are not imagining it. McKinsey reports that while respondents see use‑case‑level benefits and innovation signals, only 39% report enterprise adoption and say they have begun experimenting with AI agents. But use of agents is not yet widespread.
One reason is that much of what has scaled so far is horizontal AI (copilots, chat interfaces, broad assistance) that produces diffuse productivity gains that are harder to measure and harder to compound into end‑to‑end operational outcomes. McKinsey explicitly frames this as an imbalance between fast‑scaling copilots and more transformative, function‑specific “vertical” use cases that remain stuck in pilot mode, and argues that agents are one path to breaking out of that “gen AI paradox.”
OpenAI’s enterprise guidance reflects the same practical reality: real agent value comes when systems can both retrieve the necessary context and take actions through tools—while being designed with safe boundaries and escalation paths so they can operate reliably in production conditions.
So the executive takeaway is not “we need a better model.” It is: we need an environment fit for agentic work—the data, permissions, workflow pathways, and governance that allow agents to execute rather than merely advise.
2. What Agent Experience means in practice
AX is often misunderstood because it sounds like a rebrand of UX. It is not. AX starts from a different premise: an AI agent is a user with different needs and failure modes.
Salesforce’s Chief Experience Officer defines agent experience (AX) design as the development and optimization of digital environments so that AI agents can operate efficiently and effectively—and orchestrate human‑centred outcomes. Salesforce also emphasises that AX includes both designing for agents and designing of agents, because both are required to ensure agents prioritise people’s goals.
McKinsey similarly defines an AI agent as a software component that has the agency to act on behalf of a user or a system, and describes structured ways organisations may deploy agents—from copilots to workflow automation platforms to AI‑native operating models. This matters because as you climb that ladder, the agent’s “experience” becomes less about interfaces and more about operating conditions: clarity of tasks, access to trusted data, and permitted actions.
A practical way for executives to hold the distinction:
This framing aligns with Salesforce’s definition of AX and with OpenAI and McKinsey’s emphasis on tools, actions, controls, and governance for production deployments
3. The AX blueprint for B2B leaders
To make AX actionable without turning it into an engineering conversation, treat it as five leadership questions. Each one maps directly to why pilots fail and why scalable programmes succeed.
Trusted context: can agents reliably access the “truth” of the business?
Agents are only as reliable as the context they can retrieve. McKinsey’s 2025 reports repeatedly stress that scaling is hard work, and that organizations must redesign workflows rather than bolt agents onto legacy processes. That redesign starts with trustworthy, current data flows.
This is why enterprise platforms increasingly emphasise “zero‑copy” or live connectivity patterns. For example, ServiceNow positions Workflow Data Fabric and “Zero Copy Connectors” as a way for workflows and AI agents to run on real‑time contextual data rather than proliferating duplicates; ServiceNow’s materials explicitly describe accessing external data without copying it into the platform.
Autonomy boundaries: where can the agent act, and where must it ask?
OpenAI’s agent deployment guidance highlights the need to design workflows with appropriate controls, including human involvement for higher‑risk steps, and to architect tool use and orchestration so actions are reliable and reviewable.
This aligns with McKinsey’s broader argument that agents introduce new risks (from uncontrolled autonomy to lack of observability) that cannot be solved by “plugging agents into existing workflows,” but require reimagining task flows and governance with agents at the core.
Action paths: can the agent complete the job, not just recommend?
In practical terms, an agent that only drafts recommendations is still a copilot. OpenAI’s guide explicitly separates “data” capabilities (retrieving context) from “action” capabilities (interacting with systems to take steps such as updating records or sending messages), and treats both as foundational to real agent workflows.
Agent definition also centres on agency “to act on behalf” of a user or system, reinforcing that execution—not just advice—is core to agent value.
Observability and ownership: do you know what agents are doing and who is responsible?
Governance becomes more important as autonomy increases. ServiceNow’s AI Control Tower materials emphasise maintaining an AI asset inventory (connected to enterprise services and assets) to gauge risk and apply governance at scale—an example of the “control plane” idea executives need for agent sprawl.
This governance narrative is consistent with the National Institute of Standards and Technology AI Risk Management Framework (AI RMF), which is intended to help organisations manage AI risks and incorporate trustworthiness considerations across the AI lifecycle.
Workflow redesign: are you building agent‑centric processes or automating yesterday?
Both McKinsey and BCG stress that the real upside comes when organisations rethink workflows and value creation—moving beyond initial productivity gains toward differentiation. BCG explicitly notes that productivity gains are often the initial benefit, but argues the “real prize” in the agentic age is differentiation and sustaining advantage.
4. How to implement AX without betting the company
AX is best implemented as an operating discipline, not a one‑off project. The research strongly supports starting small, measuring outcomes, and scaling with governance.
McKinsey’s survey shows agent deployments are most often scaled in only one or two functions, and that scaling at the individual function level is still uncommon. That is a signal to lead with focused “lighthouse” workflows, not enterprise‑wide mandates.
A practical staged approach (consistent with OpenAI’s deployment guidance and McKinsey’s “reimagine workflows from the ground up” message) looks like this:
Select a bounded workflow with clean success criteria (cycle time, resolution time, coverage, rework rate) and a manageable risk profile.
Define autonomy boundaries explicitly: what the agent can do unaided, what requires review, and what must always escalate.
Instrument outcomes and failure modes, then expand autonomy only when reliability is proven in your environment.
This is how AX protects credibility: it replaces “AI theatre” with controlled, measurable operating gains—while preserving trust and compliance expectations.
5. Evidence from early adopters
Public examples—when used carefully—help leaders see what “AX in the real world” looks like. The key is to use verifiable, attributable claims and to label them appropriately.
ServiceNow publishes “Now on Now” internal outcomes for HR service experience, reporting 20× faster resolution to HR queries, 410,000 hours saved annually through AI‑powered search and virtual agent capabilities, $17.7M in annual cost avoidance from self‑service HR services, and 81% employee digital experience satisfaction (eSAT). These are vendor‑reported internal metrics, but they are concrete and tied to specific workflows and measurement categories executives understand.
Salesforce describes how Accenture expects agents to accelerate bids through the pipeline, aiming for 100% bid coverage, up from 25% as of December 2024. This is reported as part of Salesforce’s story about Agentforce and Accenture, so it should be read as a stated goal and programme narrative rather than an independently audited outcome—but it is still a verifiable, attributable claim from a named source.
Walmart provides a strong pattern: the company describes consolidating multiple agents into four “super agents”—distinct entry points for customers, associates, partners/suppliers, and developers—because many separate agents can become “overwhelming and confusing.” That is a pure AX insight: experience design becomes the discipline of reducing cognitive and operational load as agents proliferate.
On the productivity side, CIO Dive reports that Walmart said in earnings call that AI coding assistance and completion tools saved developers about 4 million hours in a year, supporting the idea that well‑embedded AI (with clear workflow integration) can generate measurable operational gains.
Across these examples, the pattern is consistent with McKinsey and BCG’s research: early wins come when organisations pick specific workflows, redesign the environment, define boundaries, and govern scale—rather than treating agents as a feature to “deploy.”
Conclusion: AX is the operating system for the agentic enterprise
Leaders should treat AI agents as a new kind of workforce capability: powerful, fast, and increasingly accessible—but only as reliable as the environment you design around them. McKinsey’s 2025 data shows adoption is high, scaling is hard, and agentic deployment remains early across most functions.
AX is the discipline that closes that gap. It operationalises what the best research now repeats: agents unlock value when organizations redesign workflows (not bolt on tools), provide trusted context, define explicit autonomy boundaries, and implement governance that makes speed sustainable.
For entrepreneurs and executives, the message is straightforward: competitive advantage will not go to the company with the flashiest demo. It will go to the company that builds agent‑ready products and operations—so agents can reliably execute outcomes with traceable, governable behaviour.
Data sources
All links are already embedded in the article.
- McKinsey 2025 “State of AI” survey findings, including adoption, scaling, and agentic AI statistics, plus McKinsey’s definition of AI agents and deployment patterns.
- BCG perspectives on agentic AI adoption and the shift from productivity gains to differentiation.
- eGlobalis Agentic AI and Customer Innovation: Why Governance Is Now the Key Differentiator
- eGlobalis: Designing CX for Non‑Human Customers: AI Agents, APIs, and Machines as Users
- eGlobalis: AI in CX Is Not the Problem — Escalation Failures Are the Real Trust Gap
- Salesforce definition of Agent Experience (AX) and a published Agentforce/Accenture example with bid‑coverage goals.
- OpenAI deployment guidance on agent building blocks (data, action, orchestration) and safe rollout considerations.
- ServiceNow materials on Workflow Data Fabric / Zero Copy Connectors, AI Control Tower governance framing, and “Now on Now” HR outcomes.
- NIST AI Risk Management Framework (AI RMF) to anchor governance language in an independent standard.
- Walmart’s first‑party description of “super agents” as unified entry points, and CIO Dive reporting of the “4 million developer hours” claim from Walmart earnings commentary.
AI Assistance Disclosure
AI tools were used solely for language refinement, grammar, and structural clarity. All ideas, analysis, and conclusions are the author’s own.
A version of this post was originally published on eGlobalis.
What an eye-opening article on Agent Experience (AX)! It’s made me reflect deeply on how we’re redesigning not just CX, but the experience of these “new autonomous employees” in B2B.
As an architect of connected experiences, I see AX as the natural evolution toward true omnichannel: where AI agents access trusted contexts and execute within clear boundaries, freeing humans for strategic and empathetic work. This shift from copilots to autonomous agents echoes my recent contributions to the Contact Center Observatory, where AI humanizes processes without losing that personal touch.
Thank you, Ricardo Saltz Gulko, for illuminating this leadership inflection point
Brilliant take on AX as the OS for agentic enterprises! Love how it bridges McKinsey data to practical blueprints like trusted context and autonomy boundaries.
Ricardo, this is a very interesting perspective.
Many discussions about AI agents focus on capabilities or architecture, but your point about Agent Experience highlights something important: AI agents also need an environment where they can actually operate.
From a contact center perspective, this idea feels quite familiar.
Human agents have always depended on the surrounding structure — clear workflows, reliable information, and defined boundaries for decisions.
When those elements are missing, the burden often shifts to the frontline.
In that sense, AX is not only about designing systems for AI agents.
It also raises a broader question about how organizations design the operational environment for both human and non-human agents working together.
Thank you for sharing this perspective.
Agentic AI is pushing CX into a new era where the experience isn’t just delivered by humans — it’s shaped by autonomous systems acting on a customer’s behalf. What stands out to me is that the real differentiator isn’t the tech anymore. It’s whether organizations build the governance to make those autonomous decisions trustworthy.
From a CX lens, governance isn’t bureaucracy. It’s what ensures customers get fast, consistent, explainable outcomes instead of surprises. When autonomy is paired with clear guardrails, you get the kind of reliability that actually strengthens trust and frees teams to focus on higher‑value work.
We’re no longer just designing journeys; we’re designing the conditions under which AI can safely participate in them. The companies that treat governance as part of the customer promise — not an afterthought — will be the ones who scale innovation without sacrificing confidence
This is an excellent review of the evolving AX world. I do believe it will evolve further as we experiment more and implement more. We need to think of the agents as human agents from a training and feedback standpoint. We need to manage them and not assume that a one time training is sufficient and the rest they can train themselves. the evaluation of outcomes on a regular basis will be crucial to ensure they represent the true voice of the company.
Thank you! Takei-san, you’re pointing to something that becomes very visible in real implementations. In several B2B AI deployments I’ve been involved in, the main issue was not the model or the agent logic — it was the absence of clear operating conditions. When that happens, execution breaks down quickly: decisions stall, handoffs become inconsistent, and the system starts relying on human intervention far more than intended. This is where Agent Experience becomes a structural discipline.
It forces organizations to define how decisions are made, when escalation is triggered, and how accountability flows across systems. Without that, AI introduces variability instead of control. The real shift, in my view, is moving from designing “solutions” to designing operating models that can support both AI and humans consistently.
That’s where scalability and trust actually come from.
Thank you for pushing the conversation in that direction. -R
Lior, thank you — really appreciate your perspective. I agree, and I’d add that the real shift is not just training agents, but governing them. As organizations scale agentic models, the challenge becomes defining clear ownership, decision boundaries, and accountability for outcomes. Without that layer, even well-trained agents can create inconsistency. The companies getting this right are the ones treating AX as an operational discipline, not just a technical capability.
Great point — this is exactly where the next level of maturity will be defined. —R
Debbie, thank you — really thoughtful points. What I’m seeing is that once autonomy comes into play, good design alone isn’t enough. The system has to stay consistent and make the right calls even without human oversight. That’s where governance becomes practical — not as structure for its own sake, but as what keeps decisions aligned at scale.
Your point about it being part of the customer promise is spot on. That’s exactly where this is heading. — R
Takei-san, ありがとうございます — I really value this super practical perspective, especially from the contact center lens. What you describe is exactly where many organizations are still underestimating the challenge. It’s not only about enabling agents, but about how the surrounding model absorbs complexity. When that structure is weak, the pressure inevitably shifts somewhere else — often to the frontline.
I also really like your point on the shared environment between human and non-human agents. That intersection is where a lot of friction — or real progress — will happen. Thank you! –R
Thank you for bringing this angle into the discussion. –R
Mª Angeles Dominguez Santalla, thank you very much — I truly appreciate you taking the time to share such a thoughtful reflection.
What stood out to me in your perspective is the connection between autonomy and context. In many implementations, the challenge is not the intelligence of the agent, but the quality and reliability of the environment it operates in. That’s where the real differentiation starts to emerge. I also like how you frame this as a leadership moment. The shift is no longer about adding AI into existing structures, but about rethinking how work is designed across human and autonomous collaboration.
Really valuable contribution — thank you for adding this angle to the discussion. –R
Great article on AX. This part “AX is best implemented as an operating discipline, not a one‑off project. The research strongly supports starting small, measuring outcomes, and scaling with governance.”, really makes sense in my head as it should be seen more like a continuous deployment process.
What resonates most here is treating AI agents as real employees. The AX blueprint you outline moves the conversation from ‘which model’ to ‘which operating conditions,’ which is the shift most B2B leaders are missing. It’s a very practical way to translate AI hype into value creation.
I appreciate how you separate UX from AX. Defining agents as users with their own constraints, risks, and success criteria is a good reframing for enterprise design teams.
Ricardo, this really resonated – especially the point that most companies don’t actually have an AI problem, they have a workflow and ownership problem that AI is now exposing.
What I found valuable here is the focus on boundaries and observability. It sounds simple, but in reality that’s where things tend to fall apart. Without clarity on what an agent is allowed to do – and who is accountable when it goes wrong – “autonomy” quickly turns into confusion rather than progress.
Also appreciated the way you framed this as an environment issue rather than a model issue. It makes it much more tangible. It’s not about chasing better AI, it’s about setting things up so the AI can actually do something useful and reliable.
Feels like a very practical way to cut through a lot of the noise around AI right now – Thanks!
João, I like how you connected that to continuous deployment — that’s exactly the right lens.
AX isn’t something you “roll out,” it’s something you keep refining — adjusting behaviour, boundaries, and outcomes as you go. That’s where governance becomes critical, not as control, but as a way to scale safely. Thank you so much –R
Tzachi thank you, really appreciate that perspective. That shift from “which model” to “which operating conditions” is exactly where real value starts to show up — otherwise it stays in experimentation mode.
And fully agree on separating UX from AX. Once agents are treated as users with constraints and accountability, the whole design approach becomes much more grounded and executable.
Great reflection. thnks so much again –R
Raphael, thanks a lot and I really appreciate you calling that out. You’re right — boundaries and observability sound simple, but that’s exactly where most implementations break. Without clarity on responsibility, autonomy quickly creates noise instead of progress. That’s why AX has to be designed around accountability from day one, not added later. Great reflection. Thank you –R
João, I like how you connected that to continuous deployment — that’s exactly the right way to think about it. AX needs that ongoing cycle of iteration, measurement, and adjustment. Treating it as a one-time rollout is where it usually breaks. Thanks a lot –R
Tzachi, really appreciate that. The shift to operating conditions is exactly where things become real. And treating agents as “employeees” — with constraints and accountability — is what makes AX scalable. thank you so much –R
Raphael, really liked this. You captured it well — boundaries and observability sound simple, but that’s exactly where things break. Without clear ownership, autonomy turns into confusion fast. Framing it as an environment issue is key. That’s what makes AI actually reliable and useful in practice. Great line of thoughts. Thank you very much again–R
This is a great read. Thinking about my own reaction to a comment you made about a McKinsey report from 2025 close to the start of what you wrote… what can I say?… My reaction was “2025! That’s ancient history.” While AI has been advancing at amazing speed, Agentic AI has been moving even faster. Personally, I think the key is to concentrate on the areas where Agentic AI complements what humans are doing, rather than replacing it. Since this is about B2B, I think most companies only assign humans to the largest accounts. Agentic AI complements that for the rest. Yes, people will say that what the agents are doing is only 70 to 80% accurate, but that’s so much better than doing nothing.
Hi Maurice, thank you for posting your thoughts, your perspective really lands — the pace is so fast that even very recent references already feel outdated, especially with how quickly agentic AI is evolving. In Samsung there is a group formed around 20 months ago, to keep catching up with such fast pace progress of AI.
The real shift isn’t replacement, it’s how agentic AI and humans operate together.. yet. In B2B, the strongest model is still hybrid: AI driving speed, orchestration, and scale, while humans focus on judgment, accountability, and relationships. To prevent human frustation.
That’s where AX becomes critical. If we don’t design and extablish governance clear roles, boundaries, and escalation between both, we don’t gain efficiency — we create friction. Thank you so much –R