Tangible results with AI are modest to date – at least if the objective is to realize business transformation on the strength of this groundbreaking technology.
That’s a key assertion in a set of 2026 AI forecasts and recommendations from the consulting firm PwC. Some of those recommendations – including a firm statement that AI initiatives must be driven from the top-down to realize tangible business outcomes – may be a bit surprising, but the firm makes a strong case for the actions it calls for.
PwC’s recommendations – I’ll highlight three below but recommend reviewing the complete analysis — are timely and worth detailed discussion as we embark on 2026 and companies aim to build on the AI work they’ve done to date to transform their business and realize the most impactful outcomes. Examples of those types of results include “surging top-line growth and significant valuation premiums,” according to PwC.
Take It From The Top
In PwC’s view, companies have erred by taking — or accepting — a ground-up approach to AI (think shadow AI or employees picking and using their own tools) “that they then try to shape into something like a strategy.” PwC goes on to say this results in projects mismatched with business priorities and not leading to business transformation. “Crowdsourcing AI efforts can create impressive adoption numbers, but it seldom produces meaningful business outcomes.”
PwC calls for enterprise-wide, top-down AI strategy driven by senior leaders determining where and how a company invests in AI. The go-forward plan should be executive, employee, and strategic alignment around shared goals. It says the melding of tech, deployment protocols, and star human performers will yield high-ROI opportunities enhanced or driven by AI. It cites demand-sensing and forecasting, product design, and core business functions including finance, IT, and HR as examples of these opportunities.
My take: employees may have gone off on their own because their employers wouldn’t or couldn’t move quickly enough, but they’ve undoubtedly realized efficiency gains and quite possibly more that shouldn’t be downplayed. But executive leadership on AI should indeed drive even higher impact as the technology is infused into new or existing processes that are core to business success.
Workforce Evolution and Impact
Closely related to the organizational considerations detailed above will be the emergence of what PwC predicts will be new “AI generalists” – that is, individuals who understand a wide range of tasks in enough depth to oversee agents and align the work of those agents with business goals. This is a clear shift away from the long-term specialization of work across industries and career paths.
Given that agents are increasingly being created to perform specialized tasks such as writing or adapting software code for specific languages, a company may opt for engineers who understand technology architecture and are able to oversee agents that know the various languages. In finance, to use another example, employees would focus on developing dynamic pricing models or optimizing vendor payment terms, while agents take over common functions such as purchase order matching or anomaly detection.
“As agents—available for anyone to buy or rent—take on more ‘midlevel’ work, differentiation comes from senior professionals who excel at strategy and innovation,” PwC says, while agents can replace entry-level workers for task-based work and mid-level employees evolve to orchestrate and manage agents.
PwC delivers two recommendations specific to evolving people and the workforce: look for candidates who are open-minded enough to be generalists and perform agent orchestration. And, companies must recognize the need for new skills (agent orchestration), new incentives (aligned to business outcomes) and new roles (oversight and strategy) as their workforces evolve.
While much is made of the threat to employees and jobs from AI, this specific set of ideas provides a roadmap that retains – but significantly repositions – people amid the advent of agents; those people must evolve from specialists to generalists with a strong AI comfort level.
The (Technical) Need for Orchestration
In 2025 the concept of “vibe” coding by non-software developers emerged and gained significant momentum. But when it comes to vibe coding, PwC cites the need for “industrializing” – putting into production with enterprise-level monitoring — the innovations that this practice yields.
A key part of industrializing AI innovations is a technology referred to at various times as orchestration or governance of agent estates.
PwC identifies key traits of an orchestration layer:
- Provides dashboards and commands that let users drag-and-drop agents into new workflows
- Facilitates use of AI tools from different vendors into unified processes
- Supports integrated code reviews and tools including encrypted credential vaults and sandboxes for prototyping
That’s a robust set of features that will undoubtedly make agent estates enterprise-ready – and give senior executives confidence in the data governance that’s been established.
Citing one practical example, PwC notes that orchestration can help IT create new capacity for strategic work by automating or assisting with many common, routine tasks.
Concluding Thoughts
I’m more bullish than PwC on the results delivered by AI agents to date, but the firm’s approach to cutting through the AI hype to focus on what’s needed going forward is refreshing nonetheless. Its organizational recommendations come at an optimal time in the AI deployment curve, especially as companies look to infuse AI as broadly as possible, align it with core business objectives, and manage the technology’s workforce implications.
The report’s focus on orchestration technology highlights the importance of this layer in getting the most out of the technology while exerting the enterprise-grade controls that protect the business, its employees, and its customers.
Related Insights:
- Microsoft’s Agent and Copilot Breakthroughs of 2025
- How Copilot at Scale Is Redefining the AI Operating Model
- Cloud Wars Live: AI Rewards Process Discipline, Not Hype
- PwC and Oracle Team Up To Turn Cloud Savings Into AI Budget
- PwC Leader Explains How AI Impact People, Finance, and ERP Migrations
AI Agent & Copilot Summit is an AI-first event to define opportunities, impact, and outcomes with Microsoft Copilot and agents. Building on its 2025 success, the 2026 event takes place March 17-19 in San Diego. Get more details.
Tom Smith
Analyst
Cloud Wars, Agent and Copilot
Related Posts
