0% found this document useful (0 votes)
2 views20 pages

Agentic AI (Aka AI Agents)

Agentic AI represents a significant advancement in artificial intelligence, characterized by systems that can perceive, reason, and act autonomously, extending human capabilities. The document discusses the historical development, core components, and various personas of agentic AI, emphasizing its potential impacts and the need for new design principles. It also highlights the importance of understanding the interplay between perception, reasoning, and action in creating effective AI agents while addressing ethical considerations and control challenges.

Uploaded by

aalyvia.julius
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
2 views20 pages

Agentic AI (Aka AI Agents)

Agentic AI represents a significant advancement in artificial intelligence, characterized by systems that can perceive, reason, and act autonomously, extending human capabilities. The document discusses the historical development, core components, and various personas of agentic AI, emphasizing its potential impacts and the need for new design principles. It also highlights the importance of understanding the interplay between perception, reasoning, and action in creating effective AI agents while addressing ethical considerations and control challenges.

Uploaded by

aalyvia.julius
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 20

Agentic AI (aka AI Agents)

e future of AI is agentic.
Published: May 2024

Key Points:

• Agentic AI: Agentic AI systems represent a signi cant advancement in AI, characterized by their ability to
perceive, reason, and act autonomously, extending human capabilities in unprecedented ways.

• De nitions and Perspectives: Various de nitions from organizations like OpenAI, IBM, and Google
DeepMind highlight the multifaceted nature of agentic AI, emphasizing goals, actions, and natural language
interfaces. At Arti ciality, we de ne agentic AI systems as those that can perceive, reason, and act with varying
complexity to extend the human mind beyond our current experience.

• Historical Context and Development: e concept of AI agency has evolved from early systems like IBM's
Deep Blue to advanced platforms such as AutoGPT and LangChain, showcasing rapid advancements in the
eld.

• Core Components: Agentic AI systems are built on three main components:

• Perception: Utilizing advanced sensors and machine learning for environmental understanding.

• Reasoning: Combining logical reasoning, probabilistic inference, and heuristic decision-making.

• Action: Executing tasks through pre-de ned behaviors, learned skills, and adaptive control.
• Agentic AI Personas: Di erent combinations of perception, reasoning, and action create varied AI personas,
from simple aides to complex way nders, each suited for speci c applications and levels of complexity.

• Impacts and Considerations: e deployment of agentic AI has the potential to signi cantly impact daily
tasks and strategic decision-making, but also raises concerns about control, ethical considerations, and
unintended consequences.

• Designing for AI Agents: e shi from customer experience (CX) to agent experience (AX) requires new
design principles to accommodate AI agents, focusing on e ciency, transparency, and the ability to process vast
amounts of information.

• Organizational Structures: Single-agent and multi-agent architectures (vertical and horizontal) each have
distinct advantages and challenges, in uencing scalability, robustness, and exibility.

Arti ciality Page 1 hello@arti ciality.world


Th
fi
fi
fi
fi
fi
ff
Th
fi
Th
fi
ft
fl
fi
Th
fi
fi
ffi
fi
fl
fi
Introduction

Recent developments in AI have given rise to a new class of systems known as agentic AI. ese systems are
characterized by their ability to perceive, reason, and act with varying levels of complexity, extending human
capabilities in unprecedented ways. While there isn’t a single de nition of agentic AI—what is agreed upon is that
agentic AI represents a signi cant leap.

OpenAI describes agentic AI systems as those that


"can pursue complex goals with limited direct
supervision." Researchers from Stanford, Microso ,
and UCLA characterize AI agents as "a class of
interactive systems that can perceive visual stimuli,
language inputs, and other environmentally-
grounded data, and can produce meaningful actions."
IBM views AI agents as "language model-powered
entities able to plan and take actions to execute goals
over multiple iterations," while Google DeepMind
de nes advanced AI assistants as "arti cial agents
with natural language interfaces, whose function is to
plan and execute sequences of actions on behalf of a
user—across one or more domains—in line with the
user's expectations."

At Arti ciality, we de ne agentic AI systems as those


that can perceive, reason, and act with varying
complexity to extend the human mind beyond our
current experience. is de nition emphasizes the
trio of capabilities—perception, reasoning, and action
—that we will use to frame various capabilities of AI
agents later in this report.

e concept of AI agency AI is not entirely new, with


roots tracing back to early systems like IBM's Deep
Blue chess-playing system in 1997 and the debut of
virtual assistants like Siri in 2011. However, recent
years have seen a rapid acceleration in the
development and deployment of increasingly
sophisticated agentic AI systems. Early experiments
like AutoGPT and BabyAGI caught the AI
community’s attention while newer tools and
platforms like Brevian, CrewAI, LangChain, and
Qurrent provide the ability for individuals and
organizations to develop their own AI agents.

Arti ciality Page 2 hello@arti ciality.world


Th
fi
fi
fi
fi
Th
fi
fi
fi
fi
ft
fi
Th
We are focused on agentic AI because we believe a) it has the potential to have wide-ranging impacts on how we
live and work, b) these technologies have the potential to both empower and disempower, and c) some level of
agency is the future of AI. Giving machines the agency to perceive, reason, and act to accomplish goals might be
quite useful with everyday tasks. But there is also the possibility for AI agents to wreak havoc if le unchecked.
Since everyone with a computer now has the ability to create agentic AI, there may be little to no ability to control
how and why these agents do what they do.

In this report, we will explore the key components of agentic AI systems, including perception, reasoning, and
action. We'll examine how these capabilities can be combined in various ways to create AI agents with di erent
roles and personas, from simple task executors to complex decision support systems. We'll also consider the
realities of deploying agentic AI in real-world contexts, including the shi from designing for human customers
to designing for AI agents themselves. Finally, we'll explore the challenges and opportunities of building multi-
agent systems and the importance of assembling diverse, multidisciplinary teams to e ectively develop and
integrate agentic AI technologies.
If you thought that LLMs changed everything (which they did), agentic AI will change everything even more, even
if its full instantiation is many years away.

Arti ciality Page 3 hello@arti ciality.world


fi
fi
ft
ff
ft
ff
The Language of Agentic AI: A Caveat

“Can machines think? is should begin with de nitions of the meaning of the terms ‘machine’ and “‘think.’”

—Alan Turing

e rapid development of agentic AI systems has sparked a need for new metaphors and mental models to
comprehend their complexity. While these AI agents display intricate behaviors and interactions that surpass
simplistic notions of machines, they remain fundamentally distinct from biological life.

When we describe advanced AI using terms traditionally reserved for living systems, we must recognize that this
linguistic mapping does not imply that AI possesses life or subjective experiences like that of humans. Instead,
this language of life, cognition, and behavior serves as a tool to help conceptualize the increasingly sophisticated
capabilities of AI, which emerge from complex informational dynamics and goal-directed, optimized functions.

As AI agents become more adaptive, interactive, and autonomous, we recognize they occupy a novel space on a
multidimensional continuum of possible complex systems, characterized by varying degrees of design versus
emergence and mechanism versus cognition. Across this continuum, agentic AI may functionally mimic
properties of life—such as perception, reasoning, and action—through rational architectures that enable them to
pursue goals while interacting with complex environments.

However, we must emphasize that this functional resemblance to living systems does not equate to AI possessing
existential choice, intentionality, phenomenological experience, or metacognition—qualities that characterize
biological cognition and remain qualitatively distinct from the computational processes of arti cial agents.

As we attempt to describe these new AI capabilities, we face the challenge of developing more precise de nitions
and conceptual frameworks to reason about the novel design space of agentic systems. In this report, we lay the
groundwork for this task, aiming to maintain a balanced perspective that includes the use of appropriate
metaphors. By doing so, we hope to provide intellectual clarity as we shape a future in which arti cial agents and
living cognitive systems interact in increasingly uid and synergistic ways.

A key challenge is to develop more precise de nitions and conceptual frameworks to reason about this novel
design space of possible agentic systems, something we attempt to make a start on in this report. Our goal is to
maintain a balanced perspective—including use of appropriate metaphors—which will provide intellectual
clarity as we shape a future in which arti cial agents and living cognitive systems interact in increasingly uid
and synergistic ways.

Arti ciality Page 4 hello@arti ciality.world


Th
fi
fi
Th
fi
fi
fi
fl
fi
fi
fi
fl
Understanding Agentic AI
At the core of agentic AI systems are three key components: perception, reasoning, and action. Each of these
capabilities can be present in an AI agent to varying degrees of complexity, allowing for the creation of systems
with diverse skill sets and areas of specialization. By understanding how these components function and interact,
we can better grasp the potential and limitations of agentic AI.

Perception

• Utilizes advanced sensors and and machine learning models to interpret and understand the environment.

• Incorporates spatial and visual reasoning capabilities to e ectively navigate and interact with the physical
world.

• Combines intuitive pattern recognition with deliberate analysis to process and make sense of complex
sensory information.

• Adapts and re nes perceptual models based on experience and feedback, enabling the agent to handle novel
and changing environments.

Perception refers to an AI agent's ability to interpret and make sense of its environment. is involves taking in
data from various sensors or input channels and processing it to extract meaningful information. At a basic level,
perceptual capabilities might include simple pattern recognition and feature extraction from structured data.
However, as the complexity of an agent's perception increases, it can begin to handle more diverse, multi-modal
sensory inputs and unstructured data.

Advanced perceptual capabilities in AI agents may include sophisticated spatial and visual reasoning, allowing
them to navigate and interact with the physical world in more natural and intuitive ways. ese systems can also
demonstrate continuous adaptation and re nement of their perceptual models, learning from experience to
better handle novel situations and environments.

As the complexity of an AI agent's perception grows, it can give rise to emergent properties and behaviors. e
system may begin to exhibit a more holistic and contextual understanding of its environment, integrating
multiple sources of information to form a coherent picture of the world around it.

Reasoning

• Employs a combination of logical reasoning, probabilistic inference, and heuristic decision-making


to solve problems and make decisions.

• Balances fast, intuitive processing (System 1) with slower, more deliberate reasoning (System 2)
to e ciently handle a range of cognitive tasks.

• Utilizes spatial and visual representations to support planning, problem-solving, and creative thinking.

Arti ciality Page 5 hello@arti ciality.world


ffi
fi
fi
fi
fi
ff
Th
Th
Th
• Incorporates emergent properties arising from the interaction of various reasoning components, enabling the
agent to exhibit intelligent behavior beyond the sum of its parts.

• Considers the broader context and complex systems in which the agent operates, accounting for the
interdependencies and feedback loops that shape its reasoning processes.

Reasoning encompasses an AI agent's ability to process information, draw inferences, and make decisions based
on the data it has gathered through perception. At a basic level, this might involve predominantly heuristic-based
reasoning with a limited ability to handle complex or ambiguous problems. Such systems may rely on narrow,
domain-speci c planning and problem-solving strategies with minimal integration of spatial and visual
reasoning.

As the complexity of an agent's reasoning increases, however, it can begin to demonstrate a more balanced
integration of intuitive and deliberate reasoning processes. ese advanced systems can tackle ill-de ned, open-
ended problems by employing exible, adaptive problem-solving strategies that span multiple domains. ey
may also incorporate seamless integration of spatial and visual reasoning, allowing them to manipulate and draw
insights from perceptual data in more sophisticated ways for more advanced planning and problem-solving..

At the highest levels of complexity, an AI agent's reasoning may begin to consider the broader context and
complex system dynamics in which it operates. is could involve explicit modeling of the interconnections and
feedback loops that shape the environment, as well as the ability to anticipate and adapt to emergent phenomena.
e result is the emergence of genuinely intelligent decision-making capabilities that can rival or even exceed
human reasoning in certain domains.

Action

• Executes actions and interacts with the environment through a combination of pre-de ned behaviors,
learned skills, and adaptive control.

• Employs spatial and visual reasoning to guide physical actions and manipulations in the world.

• Balances intuitive, reactive behaviors with deliberate, goal-directed actions to e ectively


respond to di erent situations.

• Learns and adapts action strategies based on experience and feedback, allowing the agent
to improve its performance over time.

• Considers the potential emergent consequences of its actions on the broader system
and environment in which it operates.

Action refers to an AI agent's ability to interact with its environment and e ect change based on its perceptual
inputs and reasoning processes. At a basic level, this might involve executing simple, pre-de ned action routines
or motor skills with limited adaptability or learning from experience.

As the complexity of an agent's action capabilities increases, however, it can begin to demonstrate a more diverse
and adaptive action repertoire. is might include sophisticated motor skills and behaviors that allow the agent
to manipulate objects, navigate environments, and interact with humans in more natural and intuitive ways.

Arti ciality Page 6 hello@arti ciality.world


Th
fi
fi
fi
ff
fl
Th
Th
Th
ff
ff
fi
fi
fi
Th
Advanced action capabilities may also involve deliberate, goal-directed control of actions, with the ability to plan
and execute complex sequences of behaviors to achieve desired outcomes.

Critically, the development of advanced action capabilities in AI agents o en involves the integration of spatial
and visual reasoning. By leveraging perceptual data to guide physical interactions, these systems can
demonstrate a more embodied and situated form of intelligence that is grounded in the real world.

Arti ciality Page 7 hello@arti ciality.world


fi
fi
ft
Personas

e three components of perception, reasoning, and action can be combined in various ways to create AI agents
with di erent capabilities and specializations. We can think of these combinations as de ning di erent agent
personas, each with its own strengths, limitations, and potential applications.

For example, an AI agent with low complexity in all three components might function as a simple aide or
assistant, carrying out basic tasks with minimal autonomy or adaptability. Such a system might be well-suited to
handling routine data entry, scheduling, or customer service inquiries but would likely struggle with more open-
ended or ambiguous problems.

In contrast, an agent with high complexity reasoning and action capabilities but low complexity perception
might function as a navigator or decision support system. Such an agent could analyze complex datasets,
generate novel insights, and propose solutions to strategic problems but would rely on human input for
situational awareness and contextual understanding.

At the highest end of the complexity spectrum, we might imagine an AI agent with advanced capabilities across
all three components—a "Way nder" persona that can navigate uncharted territories, devise innovative solutions
to complex problems, and adapt to changing circumstances with a high degree of autonomy. Such a system
would likely excel in domains requiring a combination of perceptual acuity, creative problem-solving, and
situated action, such as leading complex projects or guiding strategic decision-making.

By understanding the interplay of perception, reasoning, and action in agentic AI systems, we can begin to map
out the landscape of possible agent personas and skill sets. is, in turn, allows us to more e ectively match AI
capabilities to real-world problems and applications, ensuring that we are deploying these powerful tools in ways
that maximize their potential while minimizing risks and unintended consequences.

Agentic AI Personas

Persona Perception Reasoning Action

Aide Low Low Low

Trailblazer Low Low High

Navigator Low High Low

Path nder Low High High

Lookout High Low Low

Voyager High Low High

Oracle High High Low

Way nder High High High

Arti ciality Page 8 hello@arti ciality.world


Th
fi
fi
fi
ff
fi
fi
Th
fi
ff
ff
Aide

Persona Perception Reasoning Action

Aide Low Low Low

• Description: Your dependable companion who carries your everyday burdens, guiding you through the basics
with patience and care.

• Human Role: e human acts as a direct supervisor, providing step-by-step instructions and monitoring the
AI's performance closely. e human is responsible for breaking down tasks into simple, executable steps and
ensuring the AI follows them accurately.

• AI Agent Role: e AI agent acts as a simple executor, following the human's explicit instructions without
much autonomy or adaptation. e AI's role is to perform speci c, well-de ned tasks as directed by the human.

• Use for:

• Basic data entry, simple customer service queries, routine scheduling tasks.

• Managing personal schedules, creating simple shopping lists, helping with basic travel arrangements.
• Do Not Use for:

• Complex data analysis, creative problem-solving, strategic decision-making.

• Providing in-depth life advice, handling complex personal nances, making important life decisions.

Trailblazer

Persona Perception Reasoning Action

Trailblazer Low Low High

• Description: Your intrepid explorer who forges ahead, e ciently tackling repetitive tasks and clearing the
path for your success.

• Human Role: e human acts as a task manager, de ning speci c goals and constraints for the AI to follow.
e human is responsible for setting up the AI's high-level objectives and monitoring its progress, but does not
need to provide as much direct instruction for the actual execution of tasks.

• AI Agent Role: e AI agent acts as a skilled operator, autonomously executing complex tasks and adapting to
minor variations in the environment. e AI's role is to use its advanced action capabilities to achieve the
speci c goals set by the human, while still relying on human guidance for overall strategy and decision-making.

• Use for:

• Automated manufacturing processes, repetitive data cleanup, high-volume transaction processing.

Arti ciality Page 9 hello@arti ciality.world


Th
fi
fi
fi
Th
Th
Th
Th
Th
Th
Th
fi
ffi
fi
fi
fi
fi
• Automating repetitive household chores, managing routine bill payments, helping with basic tasks.
• Do Not Use for:

• Handling nuanced customer interactions, making judgment calls, providing strategic insights.

• Providing emotional support, o ering nuanced advice on personal relationships, handling complex or
sensitive family matters.

Navigator

Persona Perception Reasoning Action

Navigator Low High Low

• Description: Your skilled guide through complex information landscapes, who plots the best course through
data and ideas, steering you towards valuable insights and informed decisions.

• Human Role: e human acts as a strategic partner, working with the AI to de ne high-level goals and
strategies. e human is responsible for providing domain expertise and ethical guidance, while relying on the
AI to generate and evaluate potential solutions.

• AI Agent Role: e AI agent acts as a decision support system, using its advanced reasoning capabilities to
analyze data, generate insights, and propose solutions to the human. e AI's role is to augment human
decision-making by providing informed recommendations, while still relying on the human for nal approval
and execution.

• Use for:

• Analyzing complex datasets, providing data-driven insights, guiding decision-making processes.

• Analyzing personal health data, providing insights on nancial planning, helping to research and compare
complex products or services.

• Do Not Use for:

• Physically executing tasks, directly interacting with customers, handling sensitive or emotional situations.

• Directly intervening in personal situations, providing hands-on assistance with tasks, o ering emotional
guidance or support.

Arti ciality Page 10 hello@arti ciality.world


fi
fi
Th
Th
Th
ff
fi
Th
fi
ff
fi
Path nder

Persona Perception Reasoning Action

Path nder Low High High

• Description: Your resourceful planner and problem-solver, who blazes trails through uncharted territories,
devising innovative strategies and solutions to help you reach your goals.

• Human Role: e human acts as a high-level supervisor, setting overall objectives and constraints for the AI to
operate within. e human is responsible for monitoring the AI's performance and ensuring its actions align
with human values and priorities, but does not need to be involved in the day-to-day decision-making and
execution of tasks.
• AI Agent Role: e AI agent acts as an autonomous problem-solver, using its advanced reasoning and action
capabilities to analyze complex situations, devise solutions, and execute them independently. e AI's role is to
take on signi cant responsibility for achieving the objectives set by the human, while still remaining
accountable to human oversight.

• Use for:

• Optimizing supply chain logistics, developing comprehensive project plans, creating innovative solutions to
complex problems.

• Creating detailed plans for personal projects, developing innovative solutions to complex personal challenges,
optimizing personal productivity and goal achievement.

• Do Not Use for:

• Gathering and interpreting raw data, handling physical tasks, providing emotional support.

• Providing direct physical assistance, handling intimate or sensitive personal matters, o ering emotional
comfort or support.

Lookout

Persona Perception Reasoning Action

Lookout High Low Low

• Description: Your vigilant guardian who keeps a watchful eye on your environment, alerting you to potential
risks and anomalies and helping you stay on course.

• Human Role: e human acts as an interpreter, translating the AI's advanced perceptual insights into
actionable information. e human is responsible for making sense of the data and patterns identi ed by the AI
and determining how to use this information to guide decision-making and action.

Arti ciality Page 11 hello@arti ciality.world


fi
fi
fi
fi
fi
Th
Th
Th
Th
Th
ff
Th
fi
• AI Agent Role: e AI agent acts as a sensory processor, using its advanced perceptual capabilities to gather
and analyze complex data from the environment. e AI's role is to provide the human with rich, detailed
insights about the world, but relies on the human to interpret this information and determine appropriate
actions.

• Use for:

• Monitoring security cameras, detecting anomalies in large datasets, identifying potential quality control
issues on production lines.

• Monitoring home security systems, detecting potential issues with personal devices or appliances,
identifying unusual patterns in personal health or nancial data.

• Do Not Use for:

• Making complex decisions, physically responding to detected issues, providing strategic guidance.
• Making judgment calls on personal matters, directly intervening in emergency situations, providing guidance
on sensitive personal issues.

Voyager

Persona Perception Reasoning Action

Voyager High Low High

• Description: Your fearless adventurer who boldly navigates complex, dynamic environments, adapting to new
challenges and discovering opportunities at every turn.

• Human Role: e human acts as a supervisor and ethical governor, monitoring the AI's actions and ensuring
they align with human values and priorities. e human is responsible for setting overall objectives and
constraints, but relies on the AI to autonomously execute tasks and adapt to the environment.

• AI Agent Role: e AI agent acts as an adaptive automaton, using its advanced perception and action
capabilities to interact with the world in sophisticated ways. e AI's role is to autonomously execute tasks and
respond to changes in the environment, while still relying on the human for high-level guidance and ethical
oversight.

• Use for:

• Autonomous exploration of hazardous environments, dynamically adapting to changing conditions in


manufacturing processes, navigating complex virtual environments.

• Exploring virtual worlds for entertainment, adapting to user preferences in personal digital assistants,
navigating complex online resources for personal research.

• Do Not Use for:

• Providing strategic direction, making ethical judgments, handling sensitive interpersonal situations.

Arti ciality Page 12 hello@arti ciality.world


fi
fi
Th
Th
Th
Th
Th
fi
Th
• Providing guidance on major life decisions, handling sensitive personal information, o ering emotional or
psychological support.

Oracle

Persona Perception Reasoning Action

Oracle High High Low

• Description: Your wise and far-seeing guide, who reads the signs and interprets the mysteries, o ering
prophetic insights and illuminating the path ahead.

• Human Role: e human acts as a collaborator and co-creator, working with the AI to generate new ideas and
solutions. e human is responsible for providing creative insight and domain expertise, while relying on the AI
to process complex data and generate novel combinations and possibilities.

• AI Agent Role: e AI agent acts as an insight engine, using its advanced perception and reasoning capabilities
to identify patterns, connections, and opportunities that may not be apparent to the human. e AI's role is to
augment human creativity and problem-solving by providing new perspectives and possibilities, but relies on
the human to evaluate and act on these insights.

• Use for:

• Predicting market trends, identifying potential risks and opportunities, providing strategic foresight.

• Identifying potential risks and opportunities in personal investments, providing insights on long-term career
planning, o ering strategic guidance on personal growth and development.

• Do Not Use for:

• Directly executing tasks, physically interacting with the environment, handling routine queries or
transactions.

• Directly intervening in personal situations, providing hands-on support with tasks, o ering intimate or
emotional guidance.

Way nder

Persona Perception Reasoning Action

Way nder High High High

• Description: Your ultimate navigator and problem-solver, who masterfully guides you through the most
complex and uncharted territories, drawing on advanced capabilities across all dimensions to help you nd
your way to success and achievement.

Arti ciality Page 13 hello@arti ciality.world


fi
fi
fi
Th
fi
ff
Th
Th
ff
ff
Th
ff
fi
• Human Role: e human acts as a high-level strategist and ethical governor, setting overall goals and values
for the AI to pursue. e human is responsible for de ning the strategic vision and ethical principles that guide
the AI's actions, but relies on the AI to autonomously navigate complex environments and make decisions in
pursuit of these objectives.

• AI Agent Role: e AI agent acts as an autonomous cognitive agent, using its advanced perception, reasoning,
and action capabilities to understand the world, make decisions, and take actions in sophisticated ways. e
AI's role is to operate with signi cant autonomy and adaptability in pursuit of the objectives set by the human,
while still remaining transparent and accountable to human oversight.

• Use for:

• Managing complex, multi-faceted projects, dynamically adapting to changing strategic priorities, providing
comprehensive and contextually aware decision support.

• Managing complex personal projects, dynamically adapting to changing life circumstances, providing
comprehensive and contextually aware personal decision support.

• Do Not Use for:

• Handling simple, routine tasks, providing direct physical assistance, dealing with highly emotional or
sensitive situations.

• Handling simple daily routines, providing direct physical assistance, dealing with highly emotional or
intimate personal matters.

Agent Structures

Today, AI agents are designed with three structures or


architectures: single agents, multiple agents in a
vertical hierarchy, and multiple agents in a horizontal
hierarchy. Single agent structures are generally less
complex, as they focus on individual tasks and
actions. Multi-agent structures are generally more
complex, as they involve communication and
potential emergent behaviors among the agents.

e organizational structure of multiple agents is


proving to be important. Vertical structures o er
centralized control but can have information
bottlenecks as all agents have to communicate through the “boss.” Horizontal structures o er more exibility but
can get weighed down by communication overhead and emergent behaviors. For instance, research has shown
that agents in horizontal architectures can spend 50% of their time giving orders to each other and exchanging
niceties like saying “how are you?”

Perhaps this is proof that LLMs trained on vast quantities of conversational text will learn to mimic water cooler
chat. While chit-chat is important for human bonding, it’s likely just a waste of time and compute for machines.

Arti ciality Page 14 hello@arti ciality.world


Th
fi
fi
Th
Th
Th
fi
ff
fi
ff
fl
Th
Comparison of Single Agents vs Multi-Agents

Single Agents Multi-Agents

Complexity Generally less complex, as they focus on More complex, as they involve
individual decision-making and coordination, communication, and
actions. potential emergent behaviors among
multiple agents.

Scalability Limited scalability, as they are designed Highly scalable, as they can distribute
to handle speci c tasks or tasks and decision-making among
environments. multiple agents.

Robustness Less robust, as the failure of the single More robust, as the system can continue
agent can lead to system failure. to function even if some agents fail.

Flexibility Less exible, as they are designed for More exible, as they can adapt to
speci c tasks and environments. changes in the environment or goals
through coordination and
reorganization.

Communication No communication with other agents, Requires communication and


but may interact with the environment coordination mechanisms for agents to
or human users. share information and collaborate.

Advantages and Disadvantages of Single Agents, Vertical Multi-Agents, and Horizontal Multi-
Agents

Single Agents Vertical Multi-Agents Horizontal Multi-Agents

Advantages • Simpler architecture • E ective delegation • Flexibility


• Easier to manage • Specialization • Parallel processing
• Faster response times • E ective coordination • Fault-tolerance

Disadvantages • Limited exibility • Communication • Coordination


• Limited adaptability bottlenecks complexity
• Single point of failure • Cascading failures • Coherance di culty
• Limited adaptability • Potential con ict

Arti ciality Page 15 hello@arti ciality.world


ff
ff
fi
fl
fi
fl
fl
fi
fl
ffi
fi
CX to AX: Designing for Agents
What happens when your customers start to adopt AI agents to “get things done,” including searching for new
products, purchasing products, and interacting with customer service agents? Your customer service will no
longer solely consider human customers. Your marketing messages will no longer solely focus on human
reception. Even your SEO might no longer solely focus on Google Search but on the countless AI agents that each
individual human uses to nd things on the Agentic Web.

As AI agents become more sophisticated, they may replace human customers in a wide range of interactions. is
shi will require companies to fundamentally rethink their approach to customer experience design—shi ing
from customer experience (CX) to agent experience (AX).

Why would we want AI agents to handle these


interactions and experiences for us? First, agents are
able to do things we might value: process vast
amounts of information, make objective decisions,
and provide highly personalized experiences based on
comprehensive user data. Second, the future of the
internet (what we call the Agentic Web) will be
overwhelmed by AI-generated content that humans
will not be able to process—it will only be readable by
machine. at means we will need AI agents to “read”
the internet for us since the internet will be made by
machines, for machines. ird, agents promise
freedom—freedom from the devices where they work
for us.

e adoption of AI agents is likely to begin with low-


interaction utilities, such as bill payments and simple
service inquiries. ese transactions o en involve
standardized processes and require minimal personal
input, making them well-suited for automation. As AI
agents demonstrate their e ectiveness and reliability
in handling these tasks, consumers may become
more comfortable entrusting them with more complex
interactions. Imagine an AI agent that you use to purchase products on Amazon. Gone are the days of si ing
through reviews and wondering if you might have missed something that’s important to you. An AI agent’s
capability to process and analyze vast amounts of data might be increasingly important as the quantity of
comments and reviews increases, written by AI.

It’s important to note here that we are unaware of authentication methods for AI agents that will be required for
e-commerce companies, nancial institutions, and regulators to approve of AI agents conducting transactions.
But that gap is likely just a moment in time. As VeriSign’s authentication services grew out of the need to secure
website connections, services to authenticate AI agency will likely be developed too.

Arti ciality Page 16 hello@arti ciality.world


Th
ft
fi
Th
fi
Th
fi
Th
fi
ff
ft
ft
ft
Th
It’s not possible to know how far this transition to agentic customers will go and over what time frame. But, no
matter what, this future requires shi ing focus from CS to AX. How might you design your company’s external
experience to appeal to AI agents that are acting on behalf of humans?

Designing for AX requires a fundamentally di erent approach than traditional CX design. Rather than treating
interactions as linear, predictable exchanges of information, AX design must account for the complex, emergent
dynamics that arise when your organization’s agents interact with your customer’s agents. is means designing
interfaces and interactions that are not only intuitive and user-friendly for humans, but also for AI agents.

Designing for AI agents also means being prepared for unintended consequences and unexpected behaviors that
may arise from agent-agent interactions. is requires building in safeguards, monitoring mechanisms, and
governance structures that can detect and respond to emergent risks or harms, while also enabling bene cial
forms of emergence and adaptation to occur. And this requires building entirely new safety mechanisms—for
both your organization and your customers.

It’s important to note that these design methods do not exist today. is is an important focus for Arti ciality. We
worry that the early and easy answers like oversight, override, transparency, explainability, accountability, while
essential, will not be su cient as AI is able to handle more complexity with more agency. And that is happening
faster than anyone thought.

As a start, we o er a framework of seven design considerations, di erentiated between a human (CX) and a
machine (AX).

CX: Human AX: Machine

Information • Limited attention span. • Can process vast amounts of


Processing • Susceptible to cognitive biases (e.g., information quickly, objective and
recency bias, con rmation bias), data-driven decision-making.
in uenced by emotions and social • Not in uenced by emotions or social
factors. factors.
• May choose a product based on a recent • Analyzes thousands of product reviews
ad they saw, even if it's not the best and speci cations to determine the
option for their needs. best option based on the user's needs
and preferences.

Personalization • Appreciates personalized experiences • Expects highly personalized


based on past behavior and experiences based on comprehensive
preferences. user data.
• May respond positively to subtle cues • Requires explicit and transparent
and nudges. personalization.
• Is more likely to purchase a product if • Expects product recommendations to
it's recommended based on their be based on a comprehensive analysis
previous purchases. of the user's data, including browsing
history, purchase history, and explicitly
stated preferences.

Arti ciality Page 17 hello@arti ciality.world


fl
fi
fl
fi
fi
ff
fi
ffi
ft
Th
ff
ff
Th
Th
fi
fi
CX: Human AX: Machine

Trust & • Builds trust through positive • Requires clear explanations of how
Transparency experiences and brand reputation. recommendations are generated.
• May not always understand or • Demands transparency in data usage
question underlying algorithms. and decision-making processes.
• Trusts a well-known brand and may • Requires a detailed explanation of the
not question how their factors considered in generating a
recommendations are generated. recommendation and how the user's
data is being utilized.

User Interface • Responds to visually appealing and • Prioritizes e ciency and


intuitive interfaces. functionality over aesthetics.
• May overlook minor usability issues. • Expects seamless integration with
• Is more likely to engage with a website other AI systems and data sources.
that has an attractive layout and easy • Prioritizes a streamlined interface
navigation. that allows for quick data input and
retrieval, and seamless integration with
other relevant AI systems.

Privacy & • Concerned about personal data • Requires strict adherence to privacy
Security privacy. regulations and data protection
• Expects secure transactions and data standards.
handling. • Expects advanced security measures
• Expects their credit card information to prevent unauthorized access or
to be securely processed and stored manipulation.
when making an online purchase. • Requires all data exchanges to be
encrypted and adhere to the latest
privacy regulations, with robust
security measures in place to prevent
data breaches or unauthorized access.

Feedback & • Provides feedback through ratings, • Continuously learns and adapts
Improvement reviews, and customer support based on real-time data and
interactions. interactions.
• Adapts slowly to changes in • Requires dynamic and responsive
preferences or needs systems that can quickly incorporate
• Leaves a product review and rating feedback.
based on their experience, which • Continuously feeds real-time data on
slowly in uences the product's overall user interactions and preferences back
rating and future development. into the system, allowing for rapid
adaptation and improvement of the
online experience.

Ethical • May be in uenced by social norms • Operates based on prede ned ethical
Considerations and ethical considerations. guidelines and constraints.
• Expects companies to adhere to ethical • Requires clear de nition and
practices. enforcement of ethical boundaries in
• May avoid purchasing from a AI decision-making.
company known for unethical labor • Makes decisions within prede ned
practices or environmental damage. ethical boundaries, such as avoiding
products or services that violate human
rights or environmental standards.

Arti ciality Page 18 hello@arti ciality.world


fi
fl
fl
fi
ffi
fi
fi
fi
About Arti ciality
Extending Minds Beyond

Arti ciality is dedicated to unraveling AI's impact on individuals, organizations, and society, blending expertise
in elds ranging from cognitive science and complexity theory to philosophy and design. As arti cial
philosophers and meta-researchers, we aim to make the philosophical more practical and the practical more
philosophical. Our hope is to extend human minds beyond our current experience through the collective
intelligence of humans and machines.

We don’t endorse the utopian AI rhetoric. Yes, we believe that AI may have a positive impact on humanity. But our
excitement is balanced by our fear of Big Tech’s fantasy to make machines that can replace us. We strive to slow
the frantic pace of debates around AI, down-regulating the noise so you can see stories in context. We push and
pull on the assumptions that drive hype about the adoption of AI. And we ask questions about how AI changes
knowledge, truth, belief, creativity, meaning, authenticity, and the overall human experience.
At Arti ciality, we believe in a new vision for human-machine collaboration in which AI serves as a Mind for our
Minds, amplifying our creativity, agency, and wisdom. As AI systems emerge with the capability to perceive,
reason, and act on their own, we must approach their development with a deep understanding of the human
experience. Our mission is to ensure that as we give machines more agency, we design AI minds to work for our
human minds, enhancing our humanity rather than diminishing it.

To further our mission, we o er a range of services and resources. Our research publication cuts through the AI
hype cycle, o ering nuanced insights that dig deep into the current state and future possibilities of human-
machine collaboration. We discuss the science of humans and machines in plain English. And we share
aspirational ideas along with, at times, uncomfortably honest perspectives on AI's limitations and the scenarios
that should concern humanity.

Our strategic advisory services help organizations amplify their competitiveness by providing thought
leadership, frameworks, and best practices that challenge mainstream narratives. We o er change management
programs to guide organizations through AI-driven complexity, embracing the uncertainties and emergent
behaviors inherent in human-machine systems. And our educational programs empower workforces to
collaborate with AI systems in ways that remain meaningfully human.

Join us on this way nding journey through the complexity of minds meeting machines. Subscribe, explore our
services, or contact us to learn more about how we can help you collaborate e ectively with AI. With careful
consideration and a dash of creativity, the future may be human-centered yet.

Advisory Services
In the AI age, organizations face a daunting challenge: how to harness the potential of generative AI while
navigating the complexities of human-machine collaboration. At Arti ciality, we empower leaders to embrace
this challenge as an opportunity—an opportunity to reimagine the relationship between technology and
humanity.

Our comprehensive suite of services guides you through every stage of your AI journey:

Amplify Your Competitiveness

Arti ciality Page 19 hello@arti ciality.world


fi
fi
fi
fi
fi
ff
fi
ff
fi
fi
ff
ff
fi
• Stay ahead with cutting-edge AI research and unconventional thought leadership
• Identify and capitalize on emerging AI-driven opportunities
• Leverage generative AI to innovate with new products and services
• Mitigate threats from AI disruption to protect your market position

Build Your AI Capabilities


• Cultivate AI skills and machine intuition that enhance human judgment
• Foster an AI-empowered culture aligned with your organization's values
• Establish cross-functional AI centers of excellence for expertise-sharing
• Attract and upskill top AI talent to drive your initiatives

Manage Complex Change


• Implement complex change methods for AI-driven transformation
• Develop strategies to adapt to the emergent dynamics of human-AI collaboration
• Manage cultural and ethical challenges of successful AI adoption
• Build stakeholder trust through AI communication strategies

At Arti ciality, we believe that the key to unlocking the full potential of generative AI lies not in the technology
itself, but in the human creativity, agency, and wisdom that guide its development and deployment. By designing
AI systems that serve as a Mind for our Minds—amplifying our ability to learn, grow, and make meaning in the
world—we can create a future where technology is a catalyst for human ourishing.

Contact us today to learn how Arti ciality can help you navigate the challenges and opportunities of the AI age
with con dence and creativity. Together, let's build a future where the collective intelligence of humans and
machines creates a world that extends our minds beyond our current experience.

Who We Are
Helen and Dave Edwards are a husband-and-wife team of technology executives and AI pioneers who founded
Arti ciality. eir expertise blends insights from elds as diverse as cognitive science, complexity theory,
philosophy, and design, which they have applied in the AI space since 2012. In 2017, they sold their rst AI market
research rm, Intelligentsia.ai, to Atlantic Media, continuing their work at Atlantic Media’s subsidiary Quartz.

At Arti ciality, Helen combines decades of experience innovating across major industries with a superpower for
identifying the next emergent scienti c breakthrough that will impact our complex world of humans and
machines. She previously co-founded Intelligentsia.ai (which Atlantic Media acquired) and worked at Fonterra,
Meridian Energy, Paci c Gas & Electric, Quartz, and Transpower NZ.

At Arti ciality, Dave combines decades of experience distilling and advancing big technology trends with a
superpower for understanding how technology, design, and capital will shape our future. He previously co-
founded Intelligentsia.ai (which Atlantic Media acquired) and worked at Apple, CRV, Macromedia, Morgan
Stanley, Quartz, and inkEquity.

Helen and Dave are co-authors of Make Better Decisions: How to Improve Your Decision-Making in the Digital Age
and the upcoming Mind for our Minds: A Human-Centered Design Philosophy for AI and Amplify: A Guide to the
Complexity of Human-Machine Collaboration.

Arti ciality Page 20 hello@arti ciality.world


fi
fi
fi
fi
fi
fi
fi
fi
Th
Th
fi
fi
fi
fi
fl
fi

You might also like