How Agentic AI and Enterprise Infrastructure Are Reshaping Business

The rise of agentic AI and enterprise infrastructure is changing how companies buy, sell and secure technology. In simple terms, agentic AI means software agents that act on behalf of people — searching, negotiating, buying and completing tasks. Additionally, enterprises are building new infrastructure to support those agents. Therefore, leaders must rethink agency models, payments, identity and procurement to keep pace.

## Why Omnicom's Restructure Matters for agentic AI and enterprise infrastructure

Omnicom’s decision to reorganize its agency network signals a wider shift in how marketing services will align with AI and data needs. The company is moving from many legacy creative brands into six capability-based divisions. As a result, the new structure is designed to concentrate skills and data capabilities where enterprise clients now demand them. However, this is more than a name change. It forces clients and vendors to reassess how they buy agency services and how those services integrate with AI-driven marketing systems.

Importantly, capability-based divisions can make it easier to embed data, analytics and agentic workflows into campaigns. Therefore, brands will likely expect agencies to provide end-to-end solutions that connect strategy, creative and the data plumbing needed for agents. Additionally, retiring legacy brands like MullenLowe, FCB and DDB highlights a preference for functional teams over traditional creative houses. This should speed up how quickly agencies can deliver AI-enhanced services.

For enterprise leaders, the impact is clear. Marketing relationships will center on shared data standards, faster integrations and transparent measurement for agentic interactions. Moreover, agencies that adapt their structure and tooling will be better partners for companies building agent-based commerce or customer service flows. The outlook: expect more consolidation around capabilities that directly support AI agents and data-driven marketing.

Source: Marketing Dive

Visa and AWS: Building agentic AI and enterprise infrastructure for commerce

Visa and AWS announced a partnership to make it easier for businesses to create agentic commerce systems. In practice, this means combining Visa’s payments and commerce tools with AWS’s cloud and developer services. Therefore, companies can more quickly build agents that handle searching, booking and transacting on behalf of customers. However, agentic commerce changes more than the checkout button. It touches identity, authorization, transaction orchestration and regulatory concerns.

The collaboration points to a future where payments are embedded deeply into AI agents. As a result, buyers will expect agents to compare options, select vendors and complete purchases securely. Additionally, merchants and platforms will need APIs, standardized tokens and payment flows that agents can use without human prompts. This partnership can lower the technical barriers for businesses that want to test agentic features. Moreover, it creates new revenue opportunities for platforms, merchants and technology vendors that enable smooth agent interactions.

For enterprise decision-makers, the takeaway is practical. Start mapping customer journeys where agents could add value. Then, assess payment readiness and the security controls required for autonomous transactions. Finally, plan for vendor partnerships that can provide integrated payment and cloud capabilities. Therefore, companies that move early may capture convenience-driven revenue and set customer expectations for the next wave of commerce.

Source: Digital Commerce 360

ServiceNow, Veza and the identity puzzle

ServiceNow’s acquisition of Veza highlights an urgent need: securing who and what agents can access inside an enterprise. Veza is known for an AI-native approach to access governance. Therefore, the buy shows that identity controls are central as firms adopt cloud, SaaS and autonomous agents. However, identity is still complex in many organizations. Access rules are often fragmented across systems, which makes it hard to know what an agent can or cannot do.

This deal matters because agents acting on behalf of users will require clear, machine-readable permissions. As a result, enterprises must build identity and access tools that are both precise and scalable. Additionally, vendors will need to provide visibility into access paths that agents could exploit. That visibility helps security and audit teams demonstrate control to executives and regulators. Moreover, when access controls can be expressed as data, they become easier to enforce programmatically across platforms.

For IT and security leaders, the immediate action is to inventory where agents might operate and assess current access governance. Then, prioritize consolidating identity policies and adding tooling that can handle dynamic, agent-driven workflows. Finally, consider partnerships or acquisitions that bring AI-native governance into your stack. Therefore, better identity controls are not optional; they are foundational to safe and compliant agentic adoption.

Source: CX Today

Amazon CTO on developer roles and agentic AI and enterprise infrastructure

Amazon’s CTO frames a major shift: AI will redefine ecommerce and developer roles. He describes a movement toward “AI in the human loop,” where agents assist but people remain important. Therefore, software development will change too. Developers will focus more on building safe agent experiences, integrating APIs and creating guardrails rather than only writing user-facing features.

This perspective matters for enterprise planning. For example, retailers must rethink storefronts and catalogs for agent access. Additionally, platforms must provide developer tools that support agentic interactions and safe decision making. As a result, development teams will need new skills: designing agent prompts, specifying rules for transaction flows and monitoring agent behavior for errors or bias. Moreover, companies should expect an increase in cross-disciplinary work among product, security and legal teams.

For leaders, the path forward is to invest in developer enablement and governance. Provide toolkits that simplify agent integration, and require standards for testing and observability. Finally, consider how customer experiences will change when agents can act for users. Therefore, organizations that train developers for these new roles will lead in delivering reliable agentic services and in capturing the value they create.

Source: Digital Commerce 360

Coupa brings agentic agents into procurement workflows

Coupa’s update adds agentic AI features to a spend management platform. Notably, the platform can now turn plain-language requirements into cost and scoring formulas for sourcing optimization. Additionally, it helps transform documents like statements of work into structured inputs for procurement decisions. Therefore, agentic tools are moving from consumer-facing experiments into the back office where they can save time and reduce friction.

This change matters for procurement teams. For example, converting a natural-language need into a formal evaluation metric speeds sourcing and reduces errors. Moreover, when agents can parse documents and suggest scoring rules, teams spend less time on routine tasks and more on strategy. However, this also raises governance questions. Procurement systems contain sensitive supplier and price data. As a result, identity, access and audit trails must be robust to support agentic use.

For procurement leaders, practical steps are clear. Pilot agentic features on low-risk categories and measure time saved and accuracy. Additionally, define approval gates where humans must sign off before commitments are made. Finally, align procurement tooling with enterprise identity controls so agents operate within approved boundaries. Therefore, the value is real, but it comes with responsibility to secure and audit agent behavior.

Source: Digital Commerce 360

Final Reflection: Tying Agents, Infrastructure and Governance Together

Together, these developments show a coherent picture: agentic AI is moving from concept to operational reality, and enterprises are building the infrastructure to support it. Omnicom’s restructure signals demand for capability-aligned services. Visa and AWS enable the payment and cloud plumbing required for agentic commerce. ServiceNow’s Veza buy points to identity and access as a gating factor. Amazon’s CTO frames the developer and product changes needed. Coupa brings agents into procurement where real dollars and risks live.

Therefore, the immediate work for leaders is threefold. First, organize teams and vendors around capabilities that support agents. Second, invest in the cloud, payments and API patterns that let agents transact safely. Third, tighten identity, access and audit systems so agents act within clear guardrails. Additionally, pilot projects in low-risk areas will surface practical challenges and build experience. In short, agentic AI and enterprise infrastructure offer big gains. However, they demand thoughtful governance and integration to deliver on that promise.