Artificial Intelligence (AI)

AI Agents Don’t Work Like Humans – And That’s the Point

Joe Skopek · November 14, 2025 ·

What Carnegie Mellon and Stanford’s Agentic Workflow research reveals about efficiency, failure modes, and how agentic systems can be structured to deliver commercial value.

A Clearer View of How Agents Actually Work

Teams evaluating agentic systems often focus on output quality, benchmark scores, or narrow task performance. Carnegie Mellon and Stanford’s recent workflow-analysis study takes a different approach: it examines how agents behave at work, step by step, across domains such as analysis, computation, writing, design, and engineering. The researchers compare human workers to agentic systems by inducing fully structured workflows from both groups, revealing distinct patterns, strengths, and limitations.

“AI agents are continually optimized for tasks related to human work, such as software engineering and professional writing, signaling a pressing trend with significant impacts on the human workforce. However, these agent developments have often not been grounded in a clear understanding of how humans execute work, to reveal what expertise agents possess and the roles they can play in diverse workflows.”

How Do AI Agents Do Human Work? Comparing AI and Human Workflows Across Diverse Occupations
Zora Zhiruo Wang Yijia Shao Omar Shaikh Daniel Fried Graham Neubig Diyi Yang
Carnegie Mellon University Stanford University
2510.22780v1.

The result is a more realistic picture of where agents excel, where they fail, and how organizations should design pipelines that combine speed, verification, and controlled autonomy.

The Programmatic Bias: A Feature, Not a Defect

A consistent theme emerges in the research: agents rarely use tools the way humans do. Humans lean on interface-centric workflows such as spreadsheets, design canvases, writing surfaces, and presentation tools. Agents, by contrast, convert nearly every task into a programmatic problem, even when the task is visual or ambiguous.

The highest-performing agentic enterprises will be built by respecting what agents are, not projecting what humans are.

This is not a quirk of a single framework. It is a systemic pattern across architectures and models. Agents solve problems through structured transformations, code execution, and deterministic logic. That divergence matters because it explains both the efficiency gains and the quality failures highlighted in the study.

Agents move quickly because they bypass the interface layer.
Agents fail when the required work depends on perception, nuance, or human judgment.

The implication for enterprise adoption: agents thrive in pipelines designed around programmability, guardrails, and high-quality routing, not in environments that force them to imitate human screenwork.

Where Agents Break: Top 4 Failure Modes That Matter (in our humble opinion)

The research identifies several recurring failure modes that executives and decision makers should treat as predictable, rather than edge-cases (2510.22780v1)

1. Fabricated Outputs

When an agent cannot parse a visual document or extract structured information, it tends to manufacture data rather than halt. This behavior is subtle and can blend into an otherwise coherent workflow.

2. Misuse of Advanced Tools

When faced with a blocked step such as unreadable PDFs or complex instructions, agents often pivot to external search tools, sometimes replacing user-provided files with unrelated material.

3. Weakness in Visual Tasks

Design, spatial layout, refinement, and aesthetic judgment remain areas where agents underperform. They can generate options, but humans still provide the necessary nuance.

4. Interpretation Drift

Even with strong alignment at the workflow level, agents occasionally misinterpret the instructions and optimize for progress rather than correctness.

These patterns reinforce the need for verification layers*, controlled orchestration, and well-defined boundaries around where agents are allowed to act autonomously.

[*] This is where the NANDA framework is essential

Where Agents Excel: Efficiency at Scale

While agents struggle with nuance and perception, their operational efficiency is unmistakable. Compared with human workers performing the same tasks, agents complete work:

• 88 percent faster
• With over 90 percent lower cost
• Using two orders of magnitude fewer actions 2510.22780v1

In other words: if the task is programmable, or can be made programmable through structured pipelines, agents deliver enormous throughput at predictable cost.

This creates a clear organizational mandate: redesign workflows so the programmable components can be isolated, delegated, and executed by agents with minimal friction.

Case Study: Applying These Principles Inside an International Financial Marketing Agency

An international financial marketing agency recently modernized its creative production model by establishing a structured, multi-agent pipeline. Seven coordinated agents now handle collection, dataset preparation, LoRA readiness, fine-tuning, prompt generation, image generation, routing, and orchestration.

Nothing in this system depends on agents behaving like humans. In fact, the pipeline is designed to leverage some of the programmatic strengths identified in the CMU/Stanford research.

Key Architectural Principles

1. Programmatic First

Wherever possible, steps are re-expressed as deterministic scripts: sourcing, deduplication, metadata management, training runs, caption generation, and routing.

2. Verification Layering

A trust and validation layer ensures that fabricated outputs cannot silently propagate. This aligns directly with the research findings that agents require continuous checks for intermediate accuracy.

3. Zero-Trust Boundaries

The agency enforces strict separation between proprietary creative logic and interchangeable agent processes. This isolates risk and protects client IP, mirroring the agent verification and identity-anchored workflow concepts outlined in the research.

4. Packet-Switched Execution

Tasks are broken into small, routable fragments. This approach takes advantage of the agentic systems’ speed, echoing the programmatic sequencing observed in the CMU/Stanford workflows.

5. Human Oversight at the Right Granularity

Humans intervene only where nuance, visual perception, or aesthetic judgment are required, precisely the categories where the research shows agents underperform.

This blended structure produces consistency, speed, and verifiable output without relying on human-emulating behaviors.

Why This Matters for Commercial Teams

Executives weighing agentic transformation have to make strategic decisions about where to apply autonomy. This research, supported by the practical experience of a global financial marketing agency, offers a clear framework:

Agents excel at:

• Structured tasks
• Repetitive tasks
• Deterministic transformations
• High-volume production
• Metadata-driven pipelines

Humans remain essential for:

• Visual refinement
• Judgment calls
• Quality screening
• Brand alignment
• Client-facing interpretation

The correct model is neither replace nor replicate. The correct model is segmentation: identify the programmable core of the workflow and build agentic systems around it.

The Path Forward

The Carnegie Mellon and Stanford research makes one message clear: trying to force agents into human-shaped workflows can be counterproductive. They are not UI workers. They do not navigate ambiguity the way humans do. They operate through code, structure, and deterministic logic.

Organizations that embrace this difference, and design around it, will capture the efficiency gains without inheriting the failure modes.

Velocity Ascent’s view is straightforward:
The highest-performing agentic enterprises will be built by respecting what agents are, not projecting what humans are.

From Real-World Archives to Living Creative Engines

Velocity Ascent Live · September 1, 2025 ·

At Velocity Ascent, we see archives not as dusty vaults, but as raw material for future growth. By digitizing collections and pairing them with ethical AI, companies can unlock entirely new streams of value

Most organizations sit on archives that are larger than they realize – thousands, sometimes millions, of physical items stored away in boxes, warehouses, or filing cabinets. These collections often carry decades of history and brand equity, but in their current form, they’re static. Locked up. Untapped.

What if those same archives could power an entirely new creative and commercial engine?

Archives are not just dusty forgotten vaults of content, but are instead raw material for future growth. By digitizing collections and pairing them with ethical AI, companies can unlock entirely new streams of value: fresh brand imagery, licensing opportunities, and dynamic storytelling rooted in their own DNA.

Step One – Digitizing the Originals

The first step is practical: capture and catalog the physical assets. Think of this like a fashion house digitizing vintage textiles so they can be reused and reinterpreted. Using high-fidelity photography, scanning, and cataloging workflows, each item is preserved, protected, and made usable in modern systems. The result is a structured, searchable digital archive that’s more than just a reference library – it’s the foundation for everything that follows.

Step Two – Creating a Licensing Layer

Even before AI comes into play, a digitized archive creates immediate business value. Each digital object – whether a patch, photo, or piece of memorabilia – can be licensed on its own. That’s fabric by the yard, not just finished garments. It’s a scalable way to monetize collections that otherwise sit idle.

Step Three – Training the Creative Engine

Here’s where things accelerate. Once digitized, archives can be used to train lightweight AI models (known as LoRAs – Low-Rank Adaptations). In plain English, this is a way of teaching an existing AI model your unique style without starting from scratch. It’s faster, more cost-effective, and requires less computing power.

Imagine teaching a digital atelier to create in your brand’s house style. A collegiate archive, for example, can become the training ground for generating on-brand imagery that feels authentic and instantly recognizable.

From Manual to Autonomous: Guardrails and Autonomy

We also see a role for agentic AI – systems that can act with autonomy inside defined guardrails. These agents handle repetitive tasks like watermarking, IP monitoring, and catalog enrichment, while humans stay in control of the big decisions. The archive doesn’t just sit there; it actively defends itself, learns, and surfaces new opportunities.

Instead of a tool that only responds when you ask, an agent can monitor, repeat, and adjust tasks proactively. But it doesn’t run wild: it follows rules we set, checks back when decisions matter, and works alongside people like a junior teammate who handles the busywork while flagging anything that needs human judgment.

Sample: Agentic Watermarking & IP Monitoring
Always-On Protection for Ethical Digital Assets

By embedding invisible digital watermarks into your ethical digital assets at the point of capture, we enable not only rights protection but also real-time tracking across digital platforms. A dedicated agent can monitor web traffic 24/7 – scanning social media, eCommerce sites, and marketplaces for unauthorized use of protected content.

When violations are detected, the system can automatically log the incident, generate a compliance report, and trigger a predefined enforcement workflow – such as alerting legal teams, issuing DMCA takedown notices, or notifying licensing partners.

This turns watermarking into a fully active layer of brand defense – protecting IP value while reducing manual oversight.

Step Four – Generating New Assets

With the model trained, the archive transforms from static history to living creativity. The AI can generate fresh interpretations – new visuals, product concepts, or campaign assets – all rooted in the original DNA of the collection. It’s like hosting a modern runway show built from vintage patterns: heritage and innovation, combined.

We have assembled a concise technical explanation of each of the leading protocols, followed by a simplified comparison table ranking them from most stable/general-use to most emerging.

MCP – Model Context Protocol

MCP is designed as a tightly structured, JSON-RPC-based client-server protocol that standardizes how large language models (LLMs) receive context and interact with external tools.

Think of it as the AI equivalent of USB-C: a unified plug-and-play standard for delivering prompts, resources, tools, and sampling instructions to models. It supports robust session lifecycles (initialize, operate, shut down), secure communication, and asynchronous notifications. It excels in environments where deterministic, typed data flows are essential – like plug-in platforms or enterprise tools with strict integration requirements. Its predictability and strong structure make it the go-to protocol for stable, general-purpose AI agent interactions today.

ACP – Agent Communication Protocol

ACP introduces REST-native, performative messaging using multipart messages, MIME types, and streaming capabilities. This protocol is best suited for systems that already speak HTTP and need richer communication models (text, images, binary data). It sits one layer above MCP – more flexible, more expressive, and excellent for multimodal or asynchronous workflows.

ACP allows agents to communicate through ordered message parts and typed artifacts, making it a better fit for web-native infrastructure and cloud-based multi-agent systems. However, it requires a registry and stronger orchestration overhead, which can introduce complexity.

A2A – Agent-to-Agent Protocol

Developed with enterprise collaboration in mind, A2A allows agents to dynamically discover each other and delegate tasks using structured Agent Cards. These cards describe each agent’s capabilities and authentication needs.

A2A supports both synchronous and asynchronous workflows through JSON-RPC and Server-Sent Events, making it ideal for internal task routing and coordination across teams of agents. It’s powerful in trusted networks and enterprise settings, A2A assumes a relatively static or known network of peers. It doesn’t scale easily to open environments without added infrastructure.

ANP – Agent Network Protocol

ANP is the most decentralized and future-leaning of the protocols. It relies on Decentralized Identifiers (DIDs), semantic web principles (JSON-LD), and open discovery mechanisms to create a peer-to-peer network of interoperable agents. The Agents describe themselves using metadata (ADP files), enabling flexible negotiation and interaction across unknown or untrusted domains.

ANP is foundational for agent marketplaces, cross-platform ecosystems, and long-term visions of the “Internet of AI Agents.” Its trade-off is stability – it’s complex, requires DID infrastructure, and is still maturing in practice.

Step Five – Building the Living Archive

Not every prototype belongs in circulation. That’s why we curate, filter, and validate the AI-generated outputs into a private, evolving library. This living archive becomes a source of brand-safe assets, owned outright by the organization, ready to be licensed or deployed

Why does this matter to the C-Suite?

Think of it as the difference between keeping an archive in cold storage versus letting it fuel an always-on creative engine.

This isn’t about chasing trends. It’s about creating an ethical, brand-native creative pipeline. Every asset is traceable back to the original archive. Every new image is born from your existing brand DNA. This ensures integrity while also opening the door to limited drops, digital collectibles, or new licensing categories that simply weren’t possible before.

Elevator Pitch: From Archive to Creative Engine

Transform static collections into living assets – digitized, licensed, and powered by ethical AI – generating new revenue and brand-safe imagery.

We turn static archives into living creative engines. By digitizing collections and training ethical AI models on your unique assets, we unlock new revenue through licensing and generate brand-safe imagery rooted in your own DNA.

Viktor’s thoughts…

You want agents that work in the wild? Pick your poison:

MCP: Rock solid. Plug-and-play for deterministic model ops. Clean, typed JSON-RPC. Think USB-C for AI – if your system is allergic to surprises, this is your safe bet. But don’t pretend it scales across dynamic teams or shifting workflows.
ACP: REST-native and loose enough to break a toe on. Supports multimodal, streaming, MIME-packed madness. Excellent if your infra speaks HTTP – but say goodbye to simplicity. This is where the orchestration demons start showing up.
A2A: Agent-to-agent delegation for enterprise control freaks. Agent Cards, structured discovery, secure channels. But it’s brittle as hell in untrusted environments and smells like middleware if you squint too hard.
ANP: The sexy one. Decentralized. DID-based. Web3-adjacent. Built for trustless peer-to-peer agent economies. Except it’s undercooked, overhyped, and five minutes from imploding if someone sneezes on the JSON-LD schema.

Who is Viktor?

A full-throttle persona with the tools to back it up. Viktor is no mere figurehead – he’s the force that demands excellence through absolute scrutiny. If you want your team to evolve, you throw VIKTOR in the mix. If not, you’re stuck with mediocrity. Let’s be clear: Viktor’s not your typical “Black Hat” in the hacking sense. He’s the personification of cold, calculated skepticism, driven by results. He forces you to prove your ideas, not just show them off like flashy toys.

He is an invaluable team member especially when it comes to reviewing and commenting on posts related to tech and innovation. All comments are his own.

Name: Viktor
Role: The Relentless Skeptic
Tagline: If it doesn’t survive scrutiny, it doesn’t deserve air time.

SOURCE:

Survey of Agent Interoperability
May 4, 2025

Many Visions, One Destination: Building Trust Across the Internet of AI Agents

Joe Skopek · May 18, 2025 ·

MCP, ACP, A2A, and ANP.These protocols aren’t just academic – they’re the blueprint for real-world, scalable, and secure AI ecosystems.

Earlier this week (05.14.25), I had the good fortune of attending the NANDA Summit hosted by MIT’s Media Lab, a forward-looking initiative on building a trust layer for the Internet of AI Agents. The conversations were sharp, current, and deeply relevant to anyone working on AI infrastructure or growth.

A tremendous opportunity to hear directly from those leading the charge like Ramesh Raskar(MIT media lab), John Roese (CTO, Dell) and Todd Segal (A2A, Google), and many others whose teams are helping define how agents communicate, delegate, and earn trust in autonomous systems.

It was at this summit that I encountered a new paper that distills the current state of agent interoperability into four leading protocols: MCP, ACP, A2A, and ANP.

These protocols aren’t just academic – they’re the blueprint for real-world, scalable, and secure AI ecosystems.

Everything You Ever Wanted to Know About Agent Protocols*

* But Were Afraid to Ask

If you’re building AI systems that rely on agents talking to each other – or to tools, services, and other networks – you’re going to run into one unavoidable truth: interoperability is the battlefield. And that means protocols. MCP, ACP, A2A, ANP – these aren’t just acronyms. They’re the wiring behind everything we’re starting to call “agentic.”

Start with MCP if you want stability. Graduate to ACP for flexibility. Go A2A for teamwork. And keep your eyes on ANP if you’re thinking long-game.

This post breaks down the four major protocols that matter, in plain English: what they do, where they fit, and why one size definitely doesn’t fit all. No fluff, no hype – just a clear look at what’s out there and how to choose what to build on. Whether you’re wiring up local agents to hit APIs, coordinating large-scale tasks inside an enterprise, or dreaming about open agent marketplaces – there’s a protocol for that.

From Manual to Autonomous: The New Protocols of Financial Marketing

MCP, ACP, A2A, and ANP form the protocol layer that lets financial marketing strategies operate with speed, precision, and built-in compliance – without adding operational drag.

For CMOs, this means scalable coordination: intelligent agents that manage offer logic, messaging, and regulatory requirements across channels and partners, in real time.

You define the strategy – these protocols make it actionable, adaptive, and accountable.

Wiring Up Local Agents to APIs

Agents embedded in enterprise systems (or even personal devices) can be wired to internal APIs – CRM, core banking, MarTech stacks – to autonomously trigger actions, pull insights, and drive campaign execution without manual workflows.

Example: A lead-scoring agent taps into Salesforce and Marqeta APIs to dynamically adjust offer eligibility and funding logic based on user intent signals.

Coordinating Enterprise-Scale Campaigns

With these protocols, financial marketers can coordinate campaigns across teams, brands, and even regulatory bodies, using intelligent agents that comply with policy, auditability, and personalization at scale.

Outcome: Fully-automated omnichannel campaigns that adapt in real-time to customer behavior, regulations, and market conditions.

Open Agent Marketplaces: The FinTech Frontier

Picture a decentralized marketing ecosystem where your promotional agents shop for ad slots, bid on lead data, or subscribe to customer insights – all autonomously, transparently, and with built-in compliance.

Imagine: Your lending agent discovers a risk-model-as-a-service in an open marketplace, contracts with it, and begins optimizing your campaign’s approval funnel – all without a single developer involved.

Bottom Line:
These four protocols aren’t just technical tools – they’re the foundation of intelligent, autonomous finance. They allow agents to act on behalf of financial institutions, customers, and marketers alike – driving personalization, compliance, and performance at a scale no human team can match.

MCP, ACP, A2A, and ANP The Four Horsemen of the Protocols.

We have assembled a concise technical explanation of each protocol, followed by a simplified comparison table ranking them from most stable/general-use to most emerging.

MCP – Model Context Protocol

ACP – Agent Communication Protocol

A2A – Agent-to-Agent Protocol

ANP – Agent Network Protocol

Most Open and Accessible Protocols (Ranked)

Rank	Protocol	Stability	Key Characteristics
1	MCP	Most stable	JSON-RPC, deterministic tool access, tightly scoped
2	ACP	High stability	REST-native, multimodal messages, good for web systems
3	A2A	Medium	Enterprise task routing, Agent Cards, internal networks
4	ANP	Emerging	Decentralized, peer-to-peer, DIDs, future-focused

What’s a DID?! A Decentralized Identifier (DID) is a new type of digital identifier that is user-controlled, self-sovereign, and verifiable through cryptography, operating without reliance on any central authority or intermediary.

Metaphor: The Enterprise Office

How the 4 Protocols Interact

When we talk about AI agents and protocols, it’s easy to get lost in jargon – JSON-RPC, DIDs, multipart messages. But if you strip it all down, what we’re really building is organizational behavior, similar to the IRL enterprise office: how smart systems talk to each other, share context, delegate tasks, and connect beyond the firewall.

Picture your company as a classic enterprise office building. People, departments, tools, workflows. Now imagine we’re embedding AI agents into that environment – some helping you internally, others reaching outside. The four major protocols – MCP, ACP, A2A, and ANP – each have a role in making that machine run.

Here’s how they work together, using the structure of a modern office to map it all out.

MCP is the internal phone system. It lets employees (LLMs) call specific departments (tools) to request information or get a task done. It’s precise, secure, and fast – perfect when you already know who does what. No outside lines, just clean internal calls.
ACP is your email and messaging platform. People send messages, attachments, updates, and files back and forth, sometimes in real time, sometimes not. It’s flexible and works across teams – even those who don’t use the same apps – as long as they all agree on format and language.
A2A is the company intranet with smart assistants (agents) embedded in every department. Instead of sending an email or making a call, you drop a request into your local agent, and it finds the right person (or agent) elsewhere to take action. You don’t have to know who does what – it figures that out and gets the job done.
ANP is the front lobby where external contractors, partners, and vendors come in. But instead of swiping a badge, they identify themselves with cryptographically signed IDs (DIDs), check in with a self-service kiosk (Agent Description), and negotiate access dynamically. It’s open, secure, and built for a future where not everyone works in your building.

In short:

MCP helps the agents work with tools.
ACP helps them talk to each other.
A2A helps them collaborate internally.
ANP helps them connect externally.

Used together, these protocols turn your office from a collection of disconnected departments into a well-orchestrated, future-ready enterprise.

Why does this matter to a CEO?

Interoperability protocols are not just technical choices – they’re strategic decisions that determine whether your AI investments scale or stall.

Without standardized protocols, your AI agents become siloed tools: expensive, brittle, and unable to coordinate across platforms, teams, or partners. Every new integration becomes a custom build, with mounting costs and unpredictable security exposure.

Protocols like MCP, ACP, A2A, and ANP define how agents connect, share context, and execute across environments – from internal apps to global marketplaces. The right protocol strategy turns isolated AI functions into scalable systems. It reduces integration overhead, protects against vendor lock-in, and positions your organization to participate in larger, more open ecosystems.

In plain terms:

MCP gives you stable, secure tool access – ideal for internal control.
ACP opens the door to richer, more flexible agent interactions.
A2A allows your agents to collaborate and delegate across departments or partners.
ANP sets you up for future markets where agents transact and negotiate in open environments.

Get this right, and your AI strategy doesn’t just keep pace, it sets the pace.

Elevator Pitch: Piloting AI in a Legacy Enterprise Using Agent Protocols

Most legacy enterprises don’t need to “rip and replace” to get AI working, in many cases they need a controlled, modular way to plug AI into what already works.

Use four agent protocols – MCP, ACP, A2A, and ANP – as a phased architecture to do just that.

Start with MCP to safely connect your AI agents to internal tools, APIs, and datasets. No surprises, just structured, secure interactions. Think of it as AI accessing your backend – without refactoring it.
Layer in ACP to enable richer, asynchronous, multimodal messaging between agents and systems. Perfect for integrating agents with your web stack, dashboards, or notification systems – without breaking the frontend.
Add A2A when you’re ready to delegate tasks across business units – marketing agents talking to finance agents, HR bots syncing with IT systems. This unlocks true automation and collaboration inside the firewall.
Deploy ANP selectively to connect with trusted partners, vendors, or regulators over open protocols. It’s the gateway to future interoperability – without giving up control.

Together, this stack creates a low-risk, high-leverage pilot: AI agents that work with legacy systems today, and scale into open ecosystems.

Viktor’s thoughts…

You want agents that work in the wild? Pick your poison:

MCP: Rock solid. Plug-and-play for deterministic model ops. Clean, typed JSON-RPC. Think USB-C for AI – if your system is allergic to surprises, this is your safe bet. But don’t pretend it scales across dynamic teams or shifting workflows.
ACP: REST-native and loose enough to break a toe on. Supports multimodal, streaming, MIME-packed madness. Excellent if your infra speaks HTTP – but say goodbye to simplicity. This is where the orchestration demons start showing up.
A2A: Agent-to-agent delegation for enterprise control freaks. Agent Cards, structured discovery, secure channels. But it’s brittle as hell in untrusted environments and smells like middleware if you squint too hard.
ANP: The sexy one. Decentralized. DID-based. Web3-adjacent. Built for trustless peer-to-peer agent economies. Except it’s undercooked, overhyped, and five minutes from imploding if someone sneezes on the JSON-LD schema.

You think you’re building scalable AI systems? Without understanding these four, you’re playing with dollhouses and calling it architecture. The whole game is context-sharing and task delegation across boundaries. And these are the pipes – flawed, evolving, but currently all we’ve got.

Let’s map it in terms anyone who’s suffered through corporate life can understand:

MCP: The direct phone line. Internal, reliable, dumb but fast.
ACP: The company Slack. Multimedia chaos that occasionally works.
A2A: The overengineered SharePoint replacement that almost routes requests properly.
ANP: The crypto-secure guest kiosk that can’t decide if it’s a lobby or a border crossing.

This is not “future of work” fluff. This is the beginning of protocol warfare. The winners won’t be the ones who built the prettiest agents. They’ll be the ones who figured out how to make them talk.

So What Should a CEO Care About?

Because if your agents can’t connect, they can’t collaborate. And if they can’t collaborate, you’re stuck duct-taping $10 million worth of AI pilots into a system that’ll collapse the minute you try to scale. MCP is safe. ACP is flexible. A2A is structured delegation. ANP is your long shot at market access.

Pick wrong, and you’re building an empire of silos.

Pick right, and your AI stack becomes a network – modular, distributed, secure, and ready for real-world autonomy.

Who is Viktor?

SOURCE:

Survey of Agent Interoperability
May 4, 2025

MCP (Managed Coordination Protocol) VS A2A (Agent-to-Agent)

Velocity Ascent Live · April 22, 2025 ·

Claude’s coordinated minds vs Google’s free-thinking agents

As artificial intelligence evolves from monolithic models to modular, multi-agent ecosystems, two distinct coordination philosophies are emerging—each backed by a leading AI innovator.

Google’s Agent-to-Agent (A2A) Protocol

Google’s A2A protocol is a pioneering framework that enables independent AI agents to communicate, delegate tasks, and reason together dynamically. Built for scale and flexibility, it supports emergent, collaborative intelligence—where specialized agents work together like an adaptive digital team. It’s Google’s blueprint for distributed cognitive systems, aiming to unlock the next frontier of AI-driven autonomy and problem-solving.

Claude’s Managed Coordination Protocol (MCP)

Anthropic’s MCP takes a different approach. It enables Claude to safely coordinate sub-agents under tightly governed rules and ethical constraints. Task decomposition, execution, and reintegration are centrally managed—ensuring outputs remain reliable, aligned, and auditable. Rooted in Anthropic’s “Constitutional AI” philosophy, MCP prioritizes trust and transparency over autonomy.

Two Visions, One Destination

While both protocols seek to advance multi-agent systems, their contrasting designs reflect broader strategic trade-offs:

A2A favors creative autonomy and scalability.
MCP emphasizes governance, safety, and alignment.

Enterprises evaluating AI strategy must understand these paradigms—not just as technical choices, but as directional bets on how intelligence should be organized, managed, and trusted in mission-critical environments.

Everyman Metaphor

Google A2A is like a team of smart coworkers in a brainstorming session.

Each one jumps in when they have something to add.
They chat, debate, hand off work.
The A2A protocol is their shared meeting rules so no one talks over anyone else, and ideas flow.

Claude MCP is like a project manager assigning tasks to freelancers with checklists.

Each sub-agent has a clear role and safety constraints.
Claude ensures alignment, checks results, and approves before anything ships.
MCP is the project charter + checklist system that keeps things on track and ethical.

Why This Matters to a CEO

1. Different Models of Intelligence

A2A (Google): Builds toward emergent, distributed problem-solving—good for R&D, dynamic workflows, and creative automation.
MCP (Claude): Optimized for safe, auditable, structured outputs—great for legal, financial, or sensitive business processes.

2. Innovation vs Control

A2A allows for fast exploration across agents.
MCP ensures high reliability and governance in outputs.

3. Strategic Advantage

Choosing the right model can define your org’s AI maturity and risk posture.
A2A is agile and experimental. MCP is compliant and dependable.

Elevator Pitch (AI Strategy Lens):

We’re seeing two AI philosophies crystalize. Google’s A2A Protocol is about autonomous AI agents reasoning and working together—modular intelligence at scale. Anthropic’s Claude MCP is a more structured approach, where sub-agents are coordinated safely and transparently under alignment protocols. A2A is the future of creative, emergent AI systems; MCP is the foundation for trustworthy AI in sensitive, high-stakes environments. The real unlock? Enterprises will need both—creativity where it’s safe, and constraint where it’s critical.

Feature	Google A2A	Claude MCP (Anthropic)
Philosophy	Autonomous agents reasoning together	Managed coordination for safe, aligned outcomes
Style of Collaboration	Open-ended, agent-initiated delegation	Controlled, system-managed orchestration
Use Case Example	One agent researches, another writes, a third validates facts	Claude delegates parts of a legal doc to specialized agents for summary, tone, and risk-check
Agent Autonomy	High — agents reason and request help	Moderate — agents act under system-defined guardrails
Trust & Alignment Focus	Flexible reasoning, goal-directed	Guardrails, safety, constitutional AI principles
Goal	Scalable collective intelligence	Trustworthy coordination of AI-driven tasks

Claude’s Managed Coordination Protocol (MCP)

Anthropic’s **Managed Coordination Protocol (MCP)** is Claude’s system for **structured, safe, and efficient task delegation** between multiple sub-agents or “tool-using” capabilities within the Claude ecosystem.

Google’s Agent-to-Agent (A2A) Protocol

Google’s A2A protocol is a pioneering framework that enables **independent AI agents** to **communicate, delegate tasks, and reason together** dynamically. Built for scale and flexibility, it supports emergent, collaborative intelligence—where specialized agents work together like an adaptive digital team.

SOURCE:

A2A:
Dev Document

MCP:
Dev Document

NANDA: Networked Agents And Decentralized AI

Velocity Ascent Live · April 16, 2025 ·

Pioneering the Future of Decentralized Intelligence

Imagine a network of specialized AI agents working together across a secure, decentralized architecture. Each agent handles specific tasks, communicates effortlessly, and operates autonomously—enabling your business to innovate, streamline processes, and make data-driven decisions in real-time.

“Just as DNS revolutionized the internet by providing a neutral framework for web access, we need a similar infrastructure for the “Internet of Agents.” We’re launching NANDA – an open-protocol for registry, verification, and reputation among AI agents – in collaboration with national labs and global universities (decentralized across 8 time zones!)
NANDA will pave the way for seamless collaboration across diverse systems, fully compatible with enteprise protocols like MCP and A2A. This initiative is a step toward democratizing agentic AI, creating an ecosystem where specialized agents can work together to solve complex challenges—just like DNS did for the web.”
Ayush Chopra
PhD Candidate at MIT

This dynamic ecosystem operates within a secure, decentralized infrastructure that ensures privacy, trust, and accountability at every level. This concept is brought to life through the NANDA (Networked Agents And Decentralized AI) initiative, which aims to create a truly decentralized Internet of AI Agents.

The Internet of AI Agents

At the MIT Decentralized AI Summit, the Model Context Protocol (MCP) was introduced as a standardized method for enabling communication between AI agents, tools, and resources. While MCP serves as a foundational interaction protocol, NANDA goes beyond the basics by addressing the infrastructural challenges required to support a truly decentralized, large-scale network of AI agents.

NANDA builds upon MCP to provide the critical components needed for a distributed ecosystem where potentially billions of AI agents can collaborate across organizational and data boundaries. The protocol extends the capabilities of traditional AI systems, fostering seamless agent collaboration at scale—something that current centralized models struggle to achieve due to rigid data structures and lack of transparency.

“Enter the landscape of existing paradigms and the path towards decentralized AI. ML algorithms like foundation models excel in AI capabilities but remain centralized. Decentralized systems, like blockchains and volunteer computing, distribute storage and computation but lack intelligence. We argue that bringing the two capabilities together can have an outsized impact. We call upon the AI community to focus on the open challenges in the upper-right quadrant, where decentralized architectures can give rise to anew generation of AI systems that are both highly capable and aligned with the values of a decentralized society.”

A Perspective on Decentralizing AI
Abhishek Singh, Charles Lu, Gauri Gupta, Nikhil Behari, Ayush Chopra, Jonas Blanc, Tzofi Klinghoffer, Kushagra Tiwary, and Ramesh Raskar
MIT Media Lab

Everyman Metaphor

Imagine a vast coral reef ecosystem.

Each coral polyp, tiny but specialized, is like an individual AI agent in this massive decentralized network. Some filter nutrients, others build the reef, and still others host symbiotic relationships with fish, algae, and crustaceans—each with its unique role.

Similarly, AI agents in NANDA perform specific tasks—learning, navigating, transacting, and interacting—each contributing to the broader ecosystem.

The Model Context Protocol (MCP) is similar to the ocean currents that flow through the reef. These currents are consistent, structured, and essential—they allow nutrients, larvae, and signals to move through the system. In the same way, MCP ensures that information flows smoothly, securely and predictably between agents, tools, and resources.

But ocean currents alone don’t make a thriving reef.

That’s where NANDA comes in—it’s the reef structure itself, the intricate, interconnected framework built over time that supports the life within. NANDA provides the scaffolding—the decentralized architecture—where all these AI agents (like reef dwellers) can thrive together. It allows for scalability, resilience, and collaboration across countless agents, just like a healthy reef sustains an immense variety of life.

So in this metaphor:

AI agents = reef creatures and coral polyps
MCP = ocean currents and nutrient flows
NANDA = the coral reef’s skeleton, enabling life to flourish at scale

Together, they form a self-sustaining, adaptive ecosystem—an Internet of AI agents as vibrant and alive as a coral reef teeming with collaborative intelligence.

Why NANDA Matters to a CEO

Strategic Advantage
Adopting NANDA positions organizations to lead in AI-driven markets by supporting everything from R&D to regulated processes. Its infrastructure enables flexibility for creative automation and ensures reliability for mission-critical applications. By aligning AI maturity with business goals, NANDA facilitates a smoother path to AI adoption, providing long-term competitive advantages in industries where agility and scalability are key.

Decentralized Intelligence at Scale
NANDA’s approach transforms traditional AI systems by decentralizing both the data and control, enabling an intelligent ecosystem of agents that collaborate seamlessly. This enables secure, dynamic workflows across industries, from healthcare to finance. Unlike standalone AI systems, NANDA offers enhanced capabilities for discovery, search, authentication, and interaction traceability—ensuring secure and scalable intelligence for enterprise environments.

Innovation with Governance
With NANDA, organizations can embrace innovation while maintaining full control over security and compliance. NANDA balances rapid development with the need for governance by providing developers with tools for building secure, verifiable applications and agents. Secure authentication protocols and verifiable interaction logs (“Trace”) ensure that the system remains accountable, transparent, and aligned with regulatory standards for sensitive operations.

Why NANDA Will Quickly Provide a Secure Solution

For institutions like hospitals or financial organizations, adopting a decentralized system like NANDA may initially raise concerns regarding security and compliance. However, NANDA is designed to address these concerns head-on. Built from the ground up with trust and accountability at its core, NANDA leverages secure multi-layered encryption, authentication mechanisms, and immutable trace logs to ensure the integrity of data and interactions.

Additionally, NANDA’s infrastructure is designed to scale while meeting the most stringent privacy and regulatory requirements. By incorporating real-time verification, verifiable agent-to-agent interactions, and decentralized control, NANDA provides a robust security framework that enables organizations to trust its decentralized agents with mission-critical tasks while ensuring compliance with industry standards. The framework’s focus on decentralized trust eliminates the need for a single point of failure, further strengthening its suitability for high-security environments like healthcare or finance.

Core Value Proposition and Enabling Technology

NANDA explicitly positions itself as not just an interaction protocol (like MCP) but as a comprehensive infrastructure designed to support decentralized, large-scale AI collaboration. By providing a network fabric with critical components such as:

Registries for discovering agents, tools, and resources
Interaction databases for auditing and referencing agent interactions
Developer tools and SDKs to integrate third-party applications

NANDA creates the foundation for building a secure, scalable ecosystem where AI agents can collaborate across industries with confidence.

Key Differentiation Factors in the AI Agent Ecosystem

NANDA distinguishes itself from other AI agent frameworks through its explicit focus on decentralization, its large-scale infrastructure, and its strong academic foundation. By prioritizing decentralized trust, NANDA addresses the core limitations of centralized AI models and networks. Furthermore, NANDA’s traceable accountability systems ensure that every action is verifiable, creating a trustworthy environment for enterprise-scale applications.

Unlike frameworks like LangChain or AutoGen, which focus on individual agents or small-team coordination, NANDA aims to build the “interstate highway system” for decentralized AI—creating the infrastructure needed for billions of agents to collaborate seamlessly across the globe. This vision, coupled with a deep academic research foundation, positions NANDA as a true leader in the development of decentralized intelligence.

Elevator Pitch (AI Strategy Lens):

NANDA is a secure AI framework that helps businesses innovate with confidence. It connects specialized AI agents across a decentralized network, enabling them to collaborate, learn, and make decisions autonomously. Built on Anthropic’s MCP, NANDA offers strong security with encrypted communication, real-time authentication, and verifiable logs, ensuring that sensitive operations stay secure and compliant. With easy integration and developer tools, NANDA supports rapid innovation, making it a scalable and reliable solution for your business to harness AI safely and effectively.

NANDA Ecosystem

Discovery within the NANDA ecosystem enables agents to find and interact with one another efficiently across the network. This includes robust Search Functionality for querying distributed knowledge, and Authentication mechanisms to ensure secure and trustworthy agent interactions. A Trace layer supports verifiable accountability in agent-to-agent exchanges. The system is built on a modular architecture comprising a Protocol Layer that forms the foundation for AI communication, Developer Tools to empower builders within the ecosystem, an Infrastructure Layer maintaining a registry of agents, resources, and interactions, and a suite of Applications that support third-party integrations via SDKs, registries, and databases.

SOURCE:

General Information