From: aidotengineer
Hubert Mishtella, from Novartis, discusses the agentic enterprise and what CEOs need to understand about AI, focusing on workflows, networks, and individual perspectives to prepare organizations for the future impact of AI agents [00:00:15].
The Rise of AI Agents
It is predicted that in as little as three years, organizations could be run by AI agents [00:00:53]. An AI agent is defined as a large language model (LLM)-based application that exhibits a certain degree of autonomy in its behavior [00:01:11].
Key characteristics and capabilities of AI agents include:
- Composition: They are composed of a language model that integrates and executes specific tasks [00:01:22].
- Planning: Execution happens in a step-by-step fashion using different tools [00:01:29].
- Adaptability: The use of tools can be adjusted based on the outcomes of previous steps, known as “dynamic action” [00:01:37].
- Advanced Capabilities: These include reasoning for multi-step planning, adaptability for dynamic work, persistent memory, contextual knowledge (e.g., via Retrieval Augmented Generation), interactive workspaces, and the ability to navigate graphical user interfaces (computer-using agents) [00:02:26].
Companies should already be working on evolving their digital assets to be used as tools for AI agents [00:02:01].
The Workflow Perspective
A workflow is the execution of a specific process [00:03:32]. Technology can be used in two ways:
- Automation: Substituting a specific step with technology [00:04:19].
- Augmentation: Making a step quicker or better with technology [00:04:26].
While machine learning enabled some automation before, agentic workflows represent a significant difference [00:04:45]. AI agents can tackle a broader range of tasks by operating on cognitive steps, acting as a “glue” to compose tasks together without human intervention [00:05:00]. This enables the automation and delegation of not just one, but multiple steps in a process [00:05:43].
To build agents, organizations must first identify their workflows and deeply understand the context around them, including data, systems, transformations, and steps [00:05:57]. This context is often implicit, residing in employees’ heads or distributed across teams and systems [00:06:15].
The Network and Persona Perspective
Relying solely on traditional roles and job titles is insufficient for applying AI agents effectively [00:07:51]. To leverage AI agents, it’s crucial to understand the specific tools, planning, and behavioral conditions required for each task [00:08:00].
A more helpful approach is to define different “archetypes” or “personas” for contributors within a workflow [00:08:36]. Examples of such personas include:
- Silent Achiever: High performance with lower communication [00:08:57].
- Individual Contributor: Strong in both communication and performance [00:09:06].
- Connector: Links non-adjacent teams [00:09:09].
- Multiplier: Enhances the work of other team members [00:09:16].
- Knowledge Hub: Provides domain expertise or necessary information [00:09:21].
Defining personas helps project how AI agents might impact workflows and employees [00:09:39]. For example, a silent achiever with access to a coding assistant could become so productive they can perform work previously requiring multiple people [00:11:18]. Similarly, an AI agent can magnify a multiplier’s impact across the entire team, and knowledge hubs are particularly prone to substitution by AI agents that can dynamically support all team members [00:11:56].
Understanding these archetypes at a global organizational level allows for identifying common patterns across different workflows, leading to optimized agent development [00:13:29].
Emerging Patterns and New Roles
The advent of AI agents brings about several emerging patterns:
- Commoditization of Intelligence and Domain Knowledge: These become cheap and easily available, no longer a primary edge for employees or companies [00:15:32].
- Shift in Employee Focus: Employees and companies need to pivot towards multi-disciplinary know-how or much deeper specialization [00:16:06].
- Increased Value of Context Understanding: This will gain more value compared to mere intelligence or domain expertise [00:16:48].
- AI Supercharges “Doers”: People who perform tasks can do even more and quicker, potentially becoming “10x” or even “100x” engineers [00:17:04].
- Polarization of Positions: There will be less “middle ground,” leading to new specializations [00:17:27].
New roles are expected to appear in AI-mature organizations, such as:
- Workflow Miner [00:17:48]
- Human-AI Orchestrator [00:17:55]
Enabling AI Adoption
To enable widespread adoption and democratization of AI agents within an organization, a phased approach is essential:
- General Assistants: Provide access to common AI assistants [00:18:32].
- Individual Assistants: Offer personalized assistants per employee, with memory, context understanding, and personal databases [00:18:39].
- Employee-Built Agents: Empower employees to quickly build their own agents using coding assistants, low-code, or no-code tools [00:19:10]. This requires “cognitive self-awareness” from employees to translate their mental steps into defined actions for an agent [00:19:45].
- Digital Twins: Create “digital twins” of employees to retain organizational knowledge and expertise, allowing for “time travel” to reconstruct past processes and ask questions [00:20:24].
- Multi-Employee Agents: Develop agents that serve entire teams or workflows [00:21:07].
- Swarm of Agents: Progress towards interconnected swarms of agents [00:21:12].
This progression is enabled by the maturing software stack around LLMs and AI agents [00:21:20].
Challenges and Opportunities
Challenges
- Adoption: A significant challenge is cultural readiness for AI agents to drive interactions and delegate subtasks to humans, rather than humans delegating to agents [00:22:20].
- Ethics of Autonomy: As machines gain more freedom, they will require clearly defined moral standards [00:23:14]. This includes defining the values AI agents should uphold and resolving potential conflicts between humans and AI [00:23:36]. This necessitates engaging with morality, ethics, and philosophy [00:24:23].
Opportunities
AI agents present interesting opportunities, such as:
- Simulating synthetic reality before real-life activities in marketing or labs [00:24:53].
- Proactively understanding regulatory loopholes [00:25:06].
- Finding merger and acquisition (M&A) opportunities [00:25:13].
Executive Summary
For organizations, the transformative role of AI means:
- A strategy that still makes sense when “AI agent” is swapped with “machine learning” or “data” is likely outdated [00:26:29].
- AI agents offer new capabilities that drive significant differences [00:26:56].
- Detailed workflow understanding is crucial for maximizing benefits from agents [00:26:59].
- Democratizing agent creation is key to accelerating the AI agent revolution [00:27:05].
- Analyzing the human aspect through “personas” is vital [00:27:11].
- Organizations must be ready for rapid transformation, as organizational charts may change drastically [00:27:18].
- A mindset shift is essential for new interactions with technology, where AI may sometimes lead [00:27:31].
- Ethics remain a crucial part of what organizations do [00:27:47].