Follow ZDNET: Add us as a preferred source on Google.

ZDNET’s key takeaways

• Companies must demonstrate sustained early wins from AI investments to build momentum. 
• Companies must invest in quality, governed data, and shared workflows. 
• The key to successful agentic transformation is shifting from siloed AI to systemic AI.

Scaling agentic AI in business requires a strong data foundation. Companies need trusted quality data as the backbone of agentic AI deployments. Business leaders must identify high-impact workflows to assign to AI agents as a key capability to scaling adoption. And scaling agentic AI starts with rethinking how work gets done. 

A strong data foundation and governance are key, but how can companies mature from pockets of AI agent innovation and pilots to realizing business-wide value from AI? 

According to the Accenture research, companies need to create the intelligent superhighway — governed data, explicit decision logic and codified workflows, cloud‑native, modular architectures, and a future-ready workforce.

Five ways AI can create business-wide value

Accenture found that nearly 9 in 10 (86%) organizations plan to increase AI investments in 2026 based on their belief that AI will help increase revenues. That said, only 21% of companies are redesigning end-to-end processes with AI at the core. Accenture research based on more than 6,000 AI engagements identified five ways AI can create business-wide value.

1. Define AI’s timeline for business impact 

Treat AI as a multi-year enterprise build, not a quarter-to-quarter experiment; this requires long-term planning and doing. This also means sustained investments and the ability to identify and communicate short-term wins. Business leaders must define doable value targets to build organizational momentum. Accenture found that meaningful value from AI investments on the income statement takes 12 months or more.

2. Development of operational readiness 

According to Accenture, 70% of technology budgets still support legacy systems that slow the flow of information. To achieve operational readiness, companies must codify end-to-end processes so AI can operate quickly and at scale. The right form of AI must also be applied to how work is done. Not all work requires AI agents. The best use of AI agents is when the workflow requires reasoning; otherwise, traditional automation can do the job. Accenture noted that many firms over-apply agentic AI and leaders must avoid this trap. 

3. Strong data foundations for AI

Accenture found that when data provides consistent context, it drives better decisions. Invest in governance and semantically consistent data, which requires a modern AI-enhanced cloud stack, AI guardrails, and redesigned workflows. AI-ready cloud environments are modular in design and support machine learning, generative, and agentic AI orchestration. A strong data foundation uses clean data to deliver the right context — a shift from probabilistic to a more deterministic set of outcomes. 

Companies need a coherent data strategy and access to high-quality proprietary datasets. It is the data and the metadata (data about the data) that deliver the contextual intelligence for AI agents to execute tasks in a trustworthy manner. Accenture identified two working patterns: rebuild entire processes in which agents orchestrate workflows across systems, or invoke agents only when AI boosts performance. 

4. Talent matters – it’s about people and technology

Only one in three executives believes their talent strategy is fully integrated with their AI strategy. We must reinvent talent at work. It’s not technology that disrupts, it’s people. Accenture found that while more than 40% of organizations are upskilling their people, fewer than 10% are redesigning roles. Companies must invest in training and reskilling. Companies must also keep humans in the lead. 

At Salesforce, we found that becoming an agentic enterprise is less about a technology transformation, and more about a relational transformation. Relational transformations consist of the six ‘Rs’:

1. Redesigning process with humans and AIs.
2. Reskilling our people.
3. Redeploying people to new high-impact roles.
4. Restructuring our teams and organizations (financial implications).
5. Recalibrating new performance metrics.
6. Reclaiming latent value (the stuff we ignored in the past that can create value for our stakeholders). 

Business value reclamation is born as your company becomes increasingly autonomous through digital labor. 

5. New AI operating models are the only path to scale value

AI cannot scale inside a pre-AI operating model.  A future-ready AI operating model is more about shared capabilities and not siloed departments. This means companies must invest by buying, promoting, or building ecosystem partners. The future-proof AI ecosystem will give your company access to talent, better tools and stronger opportunities to co-innovate. 

Obstacles to business-wide scale of AI  

According to Accenture, transitioning from experiments to enterprise-wide value is a journey across three dimensions: Siloed AI to prove and diagnose, Structural AI to build the system for scale, and Systemic AI to embed intelligence in the core. Accenture defines each dimension:

1. Siloed AI: Productivity gains appear in pockets (often in enabling functions), but progress is constrained by fragmented data, ad hoc governance, and weak end-to-end links. Win quick credibility and diagnose the blockers by modernizing priority data domains, standing up joint business-tech governance, and beginning talent reinvention.
2. Structural AI: Momentum shifts from experiments to institutional capability as companies build the enterprise architecture and operating model for scale. Organizations that act across the critical enablers — value leadership, talent, digital core, responsible AI and continuous improvement — are far more likely to scale high-value use cases.
3. Systemic AI: Companies in this phase pair technological sophistication with deep shifts in talent strategy, role design and leadership behavior. Intelligence is embedded in the enterprise core. They treat reinvention as a continuous capability rather than a one-time transformation. Only a smaller set of organizations advance to systemic AI, where intelligence becomes embedded in the enterprise core, according to Accenture.

Accenture found that fewer than one in five organizations have modernized their data, platforms, governance and talent systems enough to support broad AI deployments. Accenture research reveals that obstacles to business-wide scale of AI lie in outdated operating models. A key finding from Accenture was that organizations that unlock AI’s full potential treat adoption as a strategic requirement — cloud readiness increasingly separates AI transformational leaders from laggards.

Security is also a top priority. Building resilient AI systems requires security to be embedded by design. The Accenture research shows that while early wins with AI agents are needed to build organizational confidence, it is systemic AI that will determine long-term success and overall business value. 

I love this quote from the Accenture report: “AI rewards commitment, not impatience. Nobody wants a racecar in a traffic jam.” To learn more about the Accenture research, you can visit here.