Over the past decade, digital transformation has focused on converting manual processes to digital ones, migrating infrastructure to the cloud, updating applications and creating new channels for customer and employee engagement. These efforts have resulted in tangible benefits such as accelerated cycle times, increased transparency and reduced costs. However, these initiatives have also revealed limitations: Simply digitizing a flawed process does not resolve its underlying issues; it only makes the inefficiencies operate at a faster pace.
The shift to intelligent transformation
A new shift is underway. The next phase moves beyond digital — it is intelligent. Intelligent transformation is about evolving enterprises into systems that can sense, reason, make decisions and take action with minimal friction. This evolution is driven by the emergence of AI agents: Autonomous software systems capable of achieving goals by orchestrating workflows and leveraging tools such as APIs, applications, databases and automation technologies.
From apps and dashboards to agents and actions
In the digital era, organizational value was derived from systems of record like ERP and CRM, as well as systems of insight such as business intelligence and analytics. In the intelligent era, a new dimension is added: Systems of action, defined by agentic capabilities that execute tasks across diverse functions.
AI agents don’t merely answer questions; they are capable of planning steps, invoking tools, retrieving relevant information and completing end-to-end workflows. Modern agent architectures employ tool calling or function calling, where a model determines the next action, the application executes it and the model continues based on the outcome. This approach effectively bridges language models and enterprise systems, transforming agents into practical operational levers rather than mere demonstrations.
Cross-industry value: Where agents compound outcomes
Manufacturing: From connected factories to learning factories
Industry 4.0 connected machines and visualised data through dashboards. Intelligent transformation advances this by converting signals into actions: Agents detect abnormal patterns, recommend parameter adjustments, initiate maintenance requests and coordinate the availability of parts. The cumulative benefits extend beyond uptime to faster root-cause analysis, reduced quality deviations and more predictable throughput. Even minor improvements in planning stability can lead to significant gains in service and inventory management across complex networks.
Healthcare & life sciences: From documentation burden to decision support
Clinicians and scientists often navigate context overload due to guidelines, patient histories, research literature, protocols and regulatory requirements. Agents can collate relevant context, draft structured notes, suggest trial eligibility matches or summarise evidence, all while ensuring human oversight in decision-making. Responsible implementations integrate retrieval from trusted sources, audit trails and strict permissions, as accuracy, privacy and traceability remain essential.
Retail & consumer: From personaliziation to orchestration
Personalised recommendations marked the initial stage. Now, agents can orchestrate the entire customer journey — resolving issues, amending orders, offering suitable alternatives and updating inventory systems. Internally, agents produce assortment insights, conduct promotion analyses and automate supplier communication, reducing turnaround times from weeks to hours.
The new enterprise stack: Agentic workflows built on trust
As agents advance in capability, the critical question will shift from “Can the model write?” to “Can the enterprise operate safely with agentic systems?”
A pragmatic stack is emerging:
- Experience layer: Copilots within work tools (email, CRM, service desk) and role-based agent interfaces.
- Reasoning & orchestration: Policies, routing, human-in-the-loop approvals and monitoring.
- Knowledge layer: Retrieval from curated internal content with permissions and provenance.
- Tool layer: APIs, RPA, workflow engines and systems of record enabling agent actions.
- Governance & risk: Evaluation, security, compliance and auditability by design.
Consequently, AI risk management has become a board-level concern. Frameworks such as NIST’s AI Risk Management Framework stress the importance of embedding trustworthiness across the AI lifecycle and managing risks to individuals, organizations and society.
The uncomfortable truth: Agents amplify both productivity and risk
Agents have the potential to deliver substantial productivity gains, but they also introduce new vulnerabilities: Prompt injection, data leakage, unsafe actions and excessive reliance on generated outputs. Leading platforms are continuously enhancing agent safety and implementing robust controls.
Enterprises that succeed will treat agents like other critical systems — establishing guardrails, enforcing least-privilege access, mandating approvals for sensitive actions, continually evaluating performance and maintaining traceability of data and actions.
The operating model shift: Product thinking, not project thinking
Intelligent transformation does not succeed as a one-time rollout. Agents are never static; they learn, adapt and evolve as business needs and risks change. This calls for a shift from project-based delivery to a product-oriented operating model:
- Business-led outcomes: Each agent is linked to a measurable business KPI (cycle time, NPS, yield, cash conversion).
- Fusion teams: Domain experts, engineers, data specialists and risk/compliance professionals collaborate as a unified team.
- Evaluation as a discipline: Ongoing test suites for accuracy, safety, robustness and cost.
- Change management at the edge: Successful adoption is achieved through daily workflow integration, not just on launch day.
Measuring value: 3 horizons that executives can govern
To effectively scale enterprise value, leaders should consider three horizons:
- Horizon 1: Assist. Copilots that draft, summarise, search and explain, delivering rapid adoption and immediate time savings.
- Horizon 2: Augment. Agents that complete defined workflows with approvals, offering greater ROI and governance.
- Horizon 3: Automate. New operating models that reinvent customer and enterprise value chains, presenting the largest opportunities but requiring significant change.
The common mistake is attempting to leap directly to Horizon 3. The most effective strategy is to build progressively: Use Horizon 1 to establish fluency and trust, Horizon 2 to standardise platforms and guardrails and Horizon 3 to transform the enterprise after developing confidence and capability.
Conclusion: Intelligent transformation is a leadership agenda
Digital transformation modernised technology, while intelligent transformation modernises the enterprise itself — how work is discovered, decided and delivered. AI agents serve as force multipliers, converting knowledge into action and action into learning.
Ultimately, success will not be determined by who can showcase the most impressive agent, but by who can develop the most trustworthy agentic ecosystem — one that is secure by design, outcome-oriented and embraced by employees who feel empowered rather than displaced.
In the age of AI agents, scaling enterprise value transcends traditional transformation. It requires building an organization capable of continuous self-transformation, with intelligence embedded in every decision and action.
This article is published as part of the Foundry Expert Contributor Network.
Want to join?