Employees of an Enterprise typically consume Enterprise-wide capabilities that Fulfillers and AI Agents enable. The FullFillers and AI-Agents are platform-agnostic and consume data from Systems of Record (SoR), and interact with Systems of Engagement (SoE).


Agentic AI itself does not employ a hard-coded process flow; instead, it reasons and determines its next steps/actions based on real-time triggers, responses, data, and events. This makes Agentic AI for the Enterprise less predictable but more flexible and capable than its predecessors (machine learning and hard-coded workflows). That said, AI Agents will also utilize traditional technical capabilities, such as APIs and microservices, to accomplish the work.

Employees can ideally utilize a single point of contact, a 24/7 multi-modal, multi-lingual chatbot with a human backup..The ChatBot (or its human backup) engages in a conversation with the Employee and, during or after that dialogue, can interact with multiple AI agents, each with its specific purpose and skills.
The AI agents and Fulfillers collaborate with other AI agents and enable orchestrated Enterprise Service Management (ESM) Practices such as:

Foundation, inventory, and Identity & Access:
  • Request & Update Enterprise Master Data (locations, companies, organisations, costcenters, etc)
  • Request & Manage Registration, lifecycle, capabilities, and ownership of a Digital Product (Digital Portfolio Mngt)
  • Request & Maintain Catalog of Orderable/Consumable products and services (Catalog Mngt)
  • Request & Deliver an Enhancement to an existing Digital Product (ideation & Planning)
  • Request & Provide Logical Access to a Digital Product (Identity and Access Mngt)
  • Review & Maintain Information/instructions (Knowledge Mngt)
  • Review & Maintain Legislation, Policies, Controls, and Risks
  • Discover & Monitor Inventory of Digital Products (CMDB, IT/OT Operations Mngt)
  • Determine & Manage License/Asset Utilization (IT/OT Asset Mngt)
Technical/IT Support:
  • Ask & Respond to a Generic question/request (Universal Request Mngt)
  • Request & Orchestrate a Password or User lock-out reset
  • Register & Communicate Unavailability (Planned Maintenance & Major incidents)
  • Raise & Resolve a Service interruption/degradation (Incident Mngt.)
  • Raise & Process a Legal Privacy Breach (GDPR)
  • Raise & Process an IT Security issue (IT Security Mngt)
  • Request, Process a Change to a Digital Product (Change Mngt)
  • Request & Deliver a Digital Product (Product or Service Request Mngt)
  • Observe & Manage an Event (outages, transactions, performance, logs)
Enterprise Support:
  • Request & Manage an On-site visit by a support technician (Field Service Mngt)
  • Book & Reserve a Facility (room/vehicle/asset)
  • Request & Authorize Site Access for Site Visitors
  • Raise & Process an HR Case/Request (HR Operations)
  • Raise & Process a Facility Case/Request (Facility Operations)
  • Raise & Process a Legal Case/Request (Legal Operations)
  • Raise & Process a Finance Case/Request (Financial Operations)
  • Raise & Process a Customer Case/Request (Customer Service Mngt)
  • Request & Manage Joiners, Movers, and Leavers (JML)
  • Review & Maintain Business Continuity Plans & DR (Intergrated Risk Mngt)
  • Detect and Resolve Policy and Compliance Issues (Audit & follow-up)
The above includes registration of entitlements, commitments, subscriptions, transactions, usage, traffic, contracts, consumption, purchase orders, deliveries, invoices, etc..

WHY IS THIS IMPORTANT?

Each of the bullets above represents a separate best practice in enterprise management.

Not all of the practices above will be seen as an enterprise-wide practice by all enterprises. Some enterprises will exclude all practices under Enterprise Support and reduce the enterprise-wide scope to IT only. Another enterprise will enforce enterprise-wide practices, but will allow organisations within the Enterprise to choose their platforms for each individual capability. In the three decades that I have worked in ICT, I have not seen two enterprises that made the same choices. That said, If one can soon purchase market-standard off-the-shelf AI agents that can collaborate with other market-standard off-the-shelf AI Agents, standardization of commodity work across the enterprise will likely follow..

In a selected ESM scenario, it is recommended to create practice-specific agents and assign them to a single accountable person, typically one of the ESM practice leads. As you may have noticed, the fine-grained ESM practices outlined herein are structured in a way that makes them applicable to any large enterprise. Each Enterprise capability may be assigned to a different accountable owner/lead who may be based in one of the Enterprise Domains. The creation and maintenance of AI Agents shall be performed by the responsible DevSecOps team of the platform team that hosts the AI Agents. Since each platform may be used for different practices (and vice versa), DevOps teams may need to liaise with multiple practice leads (also known as chapter leads), and practice leads may need to liaise with various DevSecOps teams.

The recommended model assumes that all parties within the enterprise adhere to the enterprise-wide practices that apply to all employees. I.e., someone in location/affiliate/opco x adheres to the same enterprise practices and uses the same enterprise tools as someone in another country/affiliate/opco. The practice lead has the mandate to enforce this across the enterprise (non-negotiable). That said, only the above mentioned Commodity ESM practices are enterprise-wide. Business-specific ERP, CRM, M&S, and Finance practices may be implemented differently across various companies and locations.

If you do not believe in matrix governance, or if you want each of your DevSecOps teams to liaise with only one Product Owner, or if you believe the platform owner should be accountable for all the things enabled by their platform, then you should choose a platform/domain-specific governance. In such a scenario, it is recommended to utilize domain-specific platforms and make the Tribe lead for the domain accountable for all matters within their domain or tribe, including compliance with applicable legal and corporate guidelines.

Companies that lack an appetite for enterprise-wide standard practices and consistency should not invest in commodity platforms or practices for employee support; instead, they should reserve that money to address inefficiencies and the complexities that come with local autonomous organizations that each work in a different way with different tools.
.