Mick Seals

🚀 Unlocking Agentic AI: From MCP to A2A, ACP, ACNBP & Beyond Agent-based systems are no longer theoretical. They're here—and protocols are the glue that makes them collaborate, delegate, and scale. If you're building with AI agents, here’s a quick guide to the evolving communication stack that powers next-gen autonomy 👇 🔌 MCP (Model Context Protocol) Think of this as the USB-C port for agents. It standardizes agent-to-tool communication via JSON-RPC, letting agents access APIs, databases, or search—ideal for stateless tool use and injecting external knowledge. 🤝 A2A (Agent‑to‑Agent Protocol) Built for collaboration across boundaries. Developed by Google, now open under the Linux Foundation, A2A enables task delegation, capability discovery, and real-time coordination between agents via HTTP/SSE and JSON-RPC. 🧠 ACP (Agent Collaboration Protocol) Structured conversations between agents. ACP introduces intent-based messaging (like REQUEST, INFORM, PROPOSE), shared ontologies, error handling, and multi-modal payload support—ideal for high-trust, real-time interactions. 🔐 ACNBP (Agent Capability Negotiation & Binding Protocol) Enabling smart agent matchmaking. This newer protocol supports secure discovery, negotiation, and commitment exchange between agents—essential when agents need to select trustworthy collaborators on the fly. 🌐 ANP + AG‑UI (Emerging standards) ANP handles agent discovery in dynamic ecosystems. AG‑UI powers real-time agent-to-user streaming (SSE/WebSockets), making agents feel alive in UI environments. 🧩 How it all fits together Imagine an agent workflow: Agent uses ANP to discover peers Agents negotiate with ACNBP Tasks are delegated via A2A External tools are accessed over MCP Conversations flow using ACP UIs stream results in real-time using AG‑UI Each protocol plays a unique role—but together, they enable resilient, intelligent, multi-agent systems. 🎯 Why this matters ✅ Modular: Plug in/out tools and agents 🛠️ Secure: Discovery, binding, identity baked in 📡 Scalable: Works across local, cloud, and hybrid systems 💬 Multimodal: Text, code, audio, and visuals 📈 Enterprise-ready: Backed by industry adoption Are you exploring multi-agent architectures yet? Curious how MCP, A2A, or ACP could reshape your enterprise workflows? Let’s connect and compare notes—agentic design is moving fast, and collaboration protocols are the key to making it work. #AgenticAI #A2A #MCP #MultiAgentSystems #AIProtocols #AIArchitecture #AIEngineering #FutureOfWork #EnterpriseAI #LLMops #AutonomousAgents

37

2 Comments

𝗘𝗔 𝗳𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸𝘀? 𝗙𝗼𝗿 𝘁𝗵𝗲 𝗺𝗶𝗻𝗶𝗼𝗻𝘀. 𝗧𝗵𝗲 𝗻𝗼𝗯𝗹𝗲 𝘄𝗼𝗿𝗸—𝘁𝗵𝗲 𝗶𝗻𝘁𝗲𝗹𝗹𝗲𝗰𝘁𝘂𝗮𝗹 𝘄𝗼𝗿𝗸—𝗶𝘀 𝘄𝗵𝗮𝘁 𝗳𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸𝘀 𝗰𝗮𝗻𝗻𝗼𝘁 𝘁𝗼𝘂𝗰𝗵. The routine, mechanical parts of Enterprise Architecture—namely the entire AS-IS documentation—should have been automated long ago. The obstacle has always been the EA framework industry's attachment to certified practitioners manually creating framework models and diagrams. None of the diagrams from any EA framework communicate effectively with non-IT stakeholders. They are useless for that purpose. The matrices and lists frameworks recommend have no value for architecture work itself—they only serve project managers tracking progress. We have confused documentation with architecture. The frameworks have become ends in themselves rather than tools for understanding. EA will only become a noble and essential part of corporate management when EA practitioners understand that their true role is to act as liaisons—bridging the gap between shareholders, the board, C-level executives, and the operational side of the institution. Enterprise Architecture has never proposed an effective way to collect, store, and produce valuable responses to queries about what lies under the hood. The primary reason is that everything under the hood is simply too vast, too complex, too interconnected, and too widespread to be collected and understood by humans alone. And, most importantly, automation is essential to eliminate the biases that people inevitably introduce when collecting and interpreting data. Manual data collection allows subjective judgments, selective attention, and preconceived notions to shape the representation of reality. Only through systematic, automated collection can we achieve an objective foundation for architectural decision-making. The path forward requires us to stop treating EA as a documentation exercise and start treating it as a sensemaking discipline. We need tools that can handle the complexity—automating the collection, synthesizing the connections, and presenting insights in forms that business stakeholders can actually use. Only then will EA fulfill its promise as the connective tissue of the modern enterprise. If you'd like a demonstration, no PowerPoint guaranteed — just send me an invite for an online conference. Be ready to be surprised. www.zconatus.com fernando@zconatus.com +55 11 98707 9954 #EA #EnterpriseArchitecture #BusinessArchitecture #zConatus #zAbsolute

2

1 Comment

Outline for 2025 Latest architecture for AI Agent Systems: 1. Agent Interface (Audio/Video/Chat) 2. Pre-processing Filters (Security, Language translation, Spell-check) 3. Memory Orchestration (Cache/Redis for short term, Lakehouse/Databricks for long term) 4. Logic Orchestration (LLM/LMM-based, Model-calling) comprising: a.) Configuration layer (Prompt engineering like Chain of thought, multi-shot and rules, settings, workflows, etc) b.) Coarse-grained agents (MCP Servers - Delegators) c.) Fine-grained agents (Tools/Capabilities of MCP - Executors, Public or Private) 5. Solution Block Layer (APIs called by Fine-grained agents) 6. Infrastructure layer (AWS, GCP, Kubernetes/Docker)

11

If you haven't checked out DQX for Data Quality, you need to read this. 🎉 DQX v0.9.3 is all about boosting your data quality game with better scalability, flexibility, and usability. Check out the highlights: 🚀 Multi-Table Checks: Run checks across multiple tables in a single go! 🗺️ New Validations: Introducing checks for IPv6 addresses, comprehensive Spatial Data (geometry, OGC validity), and ensuring your DataFrame conforms to the expected Schema! 📐 🧐 Smarter Comparisons: Enhanced dataset comparisons now include absolute and relative tolerance. 🎯 Targeted Profiling: Use SQL filters to profile specific subsets of your data for more relevant insights. 🛠️ Flexible Installs: Install DQX in custom workspaces and use group-assigned clusters. Ready to level up your data quality? Check out DQX v0.9.3! 👇 #DataQuality #DQX #DataEngineering #NewRelease

5

I’ve been running a multi-model AI workflow for a while now, and it’s changed how I build software. Instead of chasing the “best” model, I use role separation: Architect → Constructor → Evaluator. I wrote up the framework I’ve been using in this article: The ACE Framework. If you’re working with multiple models already, this may resonate. If not, you might want to start.

8

6 Comments

Sometimes, a side chat during an ARB meeting raises interesting points. Here’s a recent exchange I had while discussing the role of Enterprise Architecture with a fellow architect. These are his words: “I see the EA role in exactly three domains: -Set standards — solution patterns at both low and high levels. -Define technology usage — what technologies to use and how to use them. The primary goal is to reduce, unify, and modernize the technology stack. -Design solutions across lines of business and applications — with the primary goal of unifying all business lines into a single model. All of these are policing functions. The tricky part is figuring out how to make application development fear the police.” I’ve always seen EA first and foremost as an agent of change, and only secondarily as the police. When EA is not acting as an agent of change, the business usually fills that role—and typically with far less regard for governance. Standards, mediation, and stakeholder management are absolutely critical—but to me, that’s not police work. It’s closer to legislation and lawmaking. So, do we announce a “war on drugs”… or do we legalize marijuana? 🤔

12

3 Comments

Metis Analytics built Google for AI agents. Thousands of AI systems exist but there's no way to discover them. Agentic Exchange changes that. → Semantic search for AI agents → Trust & security scoring → Open protocol (ADP v2.0) → 51 agents indexed and growing https://lnkd.in/gBFQp_Kv Making AI systems discoverable. #AI #AgenticAI

10