top of page
Search

Voice of AI: Agents, take over! From data to decisions, Next Best Action!

  • Writer: Ralph Schwehr
    Ralph Schwehr
  • Dec 1, 2025
  • 5 min read

Something has shifted. Again. While many are still talking about "AI in customer dialogue," a new layer is moving into focus in the enterprise stack: agents that not only respond but also act, based on a data foundation that delivers "Next Best Actions" in real time. What was recently a vision has become tangible in the last two weeks: from Salesforce and Microsoft to Freshworks, Pega, Genesys, and Iterable.


As outlined in the last issue of Voice of AI, the value creation level of AI is shifting downwards, towards infrastructure, data, and governance. This week, we see the reverse movement upwards: This very foundation is now being used to operationalize agents in everyday CRM.



Data is the agent's fuel: Salesforce + Informatica

With the completion of the Informatica acquisition, Salesforce is sending a clear message: Anyone who wants to deploy agents on a large scale in their CRM must first solve the data problem. Informatica brings precisely that to the stack, from robust data access and quality assurance to MDM and governance.

For " Next Best Action, " this means that instead of fragmented customer profiles, consolidated profiles are created that can be enriched and evaluated in real time. Agents who, for example, deliver a retention offer, issue a service credit, or escalate a risk case, access the same reliable data foundation, including clean policies and traceability.

In short: Salesforce isn't simply buying "more data," but rather an enterprise-ready foundation for agent-based decision-making.


Agents are becoming part of everyday work: Microsoft, Zendesk & Freshworks

In parallel, Microsoft is refining its agent capabilities at Ignite 2025. Agent 365 forms a security and management layer for enterprise-wide agents, Work IQ becomes the intelligence layer, and new agents in Word, Excel, and PowerPoint make it clear: the leap is from "conversation" to "coordination of work."

Exciting for CRM and service teams: Agents become the executing unit for " Next Best Action ," across system boundaries. A decision from the CRM can be directly translated into documents, tickets, or workflows, while observability, policies, and role models ensure that nothing escalates unnoticed.

Zendesk picks up exactly where it left off: With the new integration into the Microsoft ecosystem (Agent 365 & M365 Copilot), service agents land where employees already work, in Outlook, Teams, Word & Co. Ticket creation, status updates, escalations: all controllable from the familiar interface, including audit trails and rights management.

Freshworks is taking the next step towards becoming a "digital employee." With its vertical "Freddy AI" agents for e-commerce, travel, logistics, and fintech, and over 50 predefined workflows, end-to-end cases are handled: refunds, order updates, CRM field changes—automated yet embedded in clear processes. Agents no longer just explain, they get things done.


The brain behind the hands: Pega, Genesys & Iterable

Agents without intelligent decision-making logic are just fast-clicking robots. This is precisely where Pega, Genesys, and Iterable come into play.

Pega has been recognized as a Leader in Real-Time Interaction Management (RTIM) in the latest Forrester Wave, a significant boost for those who understand "Next Best Action" not as a PowerPoint slide, but as a dynamic, ongoing engine. Pega's RTIM approach makes it clear: the real magic lies in reassessing, every second, which action delivers the highest value for which customer, across marketing, service, and risk.

Genesys, in its weekly Cloud CX release notes, demonstrates how AI capabilities in the contact center are continuously maturing. Micro-improvements in routing, prediction, knowledge integration, and bot handover combine to form a real-time nervous system that provides agents, both human and machine-based, with the right decisions.

In its latest blog post, Iterable describes the shift from static campaigns to "living journeys": customer journeys that respond to signals instead of rigidly following schedules. This is precisely where agents come into their own: they are the operational layer that puts NBA decisions into practice, from personalized offers to preventative measures before churn becomes visible.


From Experiment to Blueprint: How Agents & NBA (Next Best Action) 2026 Interact

The common theme of the week is clear:

  1. Strengthen the data foundation (Salesforce + Informatica)

  2. Operationalizing agents (Microsoft, Zendesk, Freshworks)

  3. Decision logic & journeys in real time (Pega, Genesys, Iterable)

This shifts the discussion surrounding AI in CRM: away from "What can the model do?" and towards "How do we orchestrate data, decisions, and agents along the entire journey?" . 2026 will not be the year of the next chatbot, but the year of robust agent governance: roles, observability, KPIs, risk frameworks, and a clear picture of which three to five agents in your organization truly create value.


💡 Key takeaways in brief

  • No clean data, no NBA (Next Best Action): Salesforce secures the data engine that scalable agents in CRM need with Informatica.

  • Agents are moving to where the work happens: Microsoft, Zendesk and Freshworks are bringing AI agents to Office, tickets and vertical workflows.

  • NBA engines remain the strategic lever: Pega, Genesys and Iterable demonstrate how real-time decisions control agents, not the other way around.

  • Contact Center + Journeys = Real-time nervous system: Service and marketing platforms merge to create the stage for "Next Best Action".

  • The decision for 2026 will be made in the blueprint: Companies that align their processes, data and governance for agents now will secure their advantage.



🔍 Source overview

🎯 Conclusion & Call to Action

The agent wave is here, but it will only become a competitive advantage where data, decisions, and governance work together seamlessly. Those who roll out agents today without NBA logic, a data foundation, and clear guidelines risk shadow processes and a loss of trust. Defining the blueprint now will lay the foundation for scalable automation over the coming years.


👉 Let's discuss your agent & NBA blueprint for 2026 In a joint roadmap workshop we will clarify:


  • which three to five agents in your CRM & Contact Center have the greatest leverage,

  • how data, processes and governance are prepared for this,

  • and how you can productively pilot your first agent in 30 days .


A reply to this email is sufficient.


Note: Feel free to share this blog, we'd appreciate it!

 
 
 

Comments

OAKAI®

E-Mail: info@oakai.de

© 2024 OAKAI®

Imprint

Data protection

  • LinkedIn

Thanks for subscribing!

NEWSLETTER

Sign up for the OAKAI newsletter.

bottom of page