Draft:Autonomic Business

Autonomic Business (or Autonomous Business) characterizes the "post-digital-business-period" where AI agents are increasingly deployed within enterprises, operating with partial or full autonomy to execute tasks, run processes, and deliver products and services to customers. A key pillar for realising an Autonomic Business is the development and operation of Autonomic Systems, which rely on an Autonomic Computing infrastructure. Autonomic Businesses have the ability to learn, configure, and optimize themselves and generally respond in a self-managing fashion to changing internal and external conditions. Enterprise Governance aligning strategic objectives to process and resources still remains largely human-driven thus the use of the term ‘autonomic’ (i.e. self-managing instead of self-governing). If an organisation in future is also partly or fully governed by AI agents at the enterprise level then someone may characterise that as a partly or fully ‘autonomous’ business (see also section below on Autonomic Business scale).

Origins

Autonomic Business has its origins in the mid-90s and even earlier when Multi-Agent Systems where applied in the context of business process management and task automation with standards developed by academia and industry (e.g. IEEE FIPA) on how these systems can interoperate inside and across organisations. These early approaches largely relied on rule-based systems and symbolic AI for reasoning, requiring the development of specific agent communication languages such as ACL and domain ontologies which limited their adoption and practical application. With the advent of Generative AI and Large Language Models able to process natural language, AI Agent abilities have been significantly boosted paving the way for making Autonomic Business a reality.

An Autonomic Business Scale

Similar to standards developed in the automotive industry for self-driving vehicles an autonomic business exhibit difference prevailing levels of automation from limited IT support all the way to a fully autonomous business:

• Level 0 - Limited IT Support
• Level 1 - Human Assistance
• Level 2 - Human Augmentation
• Level 3 - Conditional Autonomy
• Level 4 - High Autonomy
• Level 5 - Full Autonomy

In levels 1, 2, 3, the business is mainly human-driven with goal setting, decision making, task execution and process management partly or fully dependant on humans utilising IT tools to varied degrees to realise the organisation’s business and operating models.

In levels 3, 4, 5, the business is mainly machine-driven with goal setting, decision making, task execution and process management partly or fully dependant on AI agents interacting with humans and others AI agents to realise the organisation’s business and operating models.

Technologies

Key technologies supporting the current wave of Autonomic Business applications include

   1. Agentic AI encompassing AI Agents and Multi-agent Systems 
   2. Generative AI and LLMs
   3. Reinforcement Learning algorithms
   4. Neurosymbolic AI
   5. Digital Twins
   6. Business Process Management
   7. Robotic Process Automation

These technologies are fused and combined together when it comes to architecting AI agents overcoming some of the limitations of the rule-based systems that dominated early Autonomic Business attempts and applications.

Autonomic Ecosystems

Autonomous agents operating across businesses is another extension to the core Autonomic Business concept which to date primarily remains a research topic. EU funded R&D activities such as the AgentCities.RTD programme explored this concept. With renewed interest in AI and agents, development of "AgentCity" or "AgentVerse" type of environments in the context of SmartCities or other domains of ecosystem applications cannot be excluded in the near future.