bomonike

Text-to-Action carried out by AI Agents

US (English)   Norsk (Norwegian)   Español (Spanish)   Français (French)   Deutsch (German)   Italiano   Português   Estonian   اَلْعَرَبِيَّةُ (Egypt Arabic)   Napali   中文 (简体) Chinese (Simplified)   日本語 Japanese   한국어 Korean

Overview

ai-agents.png

Why This?

This was written so that you will be able to claim on your resume (LinkedIn profile):

“Created Bash, Python, and Terraform, automation to, in an hour, configure and stand up a service (running locally, on Azure, AWS, Google, and other clouds) which interacts with users’ voice (SST & TTS) to autonomously perform reasoning tasks by using MCP & A2A to supervise a mix of AI agents. The AI supervisor automatically responds to trends and telemetry. (Ask me how)”

Among @aiDotEngineer World’s Fair vidoes, VIDEO by OpenPipe worked with Method: Instead of hiring people, AI Agents in fintech

GenAI vs Agents vs AI Agents

ai-agents-whiteboard.jpeg

A. Generative AI (GenAI): When ChatGPT burst to popularity November 2023, OpenAI’s LLMs generated text based on prompts typed or spoken by end-users. Back then, the LLM model referenced static data (books, github.com, etc.), but not real-time proprietary (personal) data in the real world. More importantly, they passively waited for prompts by humans to begin action.

B. Early AI Agents made use of context window to provide personalized data in vector database using “RAG (Retrieval Augmented Generation)”. Agents access tools to perform tasks. “Reasoning” and “chain-of-thought” capabilities enable agents to plan steps to achieve specific goals defined by end-users.

C. In 2025, “Agentic AI” came on the scene when LLMs (Anthropic Claude 3) were able to “reason” in “Chain of Thought”. Agentic AI can reflect, plan, and delegate (orchestrate) steps to achieve objectives by recalling context, constraints, and previous actions (from memory) to avoid redundant or dumb behavior by breaking down obstacles into steps, then adapting to select and strategies collaboratively, and execute across tools or environments.

Think of them as virtual assistants capable of handling the chores people generally rather avoid.

Unlike humans, AI agents don’t need to sleep, operate at lightning speed, and collaborate with other agents.

Agentic means proactive – they try to solve problems without constant human input.

MCP Client is Claude client or IDE Cursor or Windsurf

D. A Global Agent can be an ID such as IDE OpenAI/Windsurf, Claude, Cursor, or extension-enabled VS Code.

E. A Local Agent is a tool that can be called by an MCP client.

Levels of automation:

References:

Use Cases

The “Javis” computer in Ironman movies is a form of AI Agent.

PROTIP: AI Agents takes advantage of new “reasoning” and “chain of thought” capabilities in LLMs such as provided by Anthropic Claude.

Gemini Code Assist agents:

Grab info:

Take action:

Update info:

“We will have more AI Agents than humans” – Gaia

Supervision

“The next billionaire can be a single person.”

In a world of AI Agents, human employees would oversee a team of AI agents, each dedicated to specific tasks.

https://ti.to/software-3/ai-engineer-worlds-fair-2025 9am, June 3rd–5pm, June 5th, 2025 Mariott Marquis, San Francisco, CA USA

https://www.linkedin.com/pulse/how-ai-agents-revolutionize-workflow-automation-2025-deshmukh-pmp–vxpcf/

Enabling tech: Screen Scraping

To understand what’s happening on YOUR computer screen, laptop manufacturers Microsoft, Intel, Apple, and others are building in their hardware and operating system for applications to see what’s happening on your screen.

The path of control logic

GenAI

Frameworks like AutoGPT, LangGraph, CrewAI, and MetaGPT.

Fine-tuned models close the gap between “can predict text” and “can complete real-world tasks.”

create separate hubs for each business area (such as Marketing, HR, and so on) i

Google AgentSpace

On the Google Agentspace platform from within Chrome’s search box, developers build agents combining Gemini models with enterprise data for unified enterprise search, analysis, and synthesis.

Other new capabilities include a no-code Agent Designer for building custom agents, and access to two new agents built by Google: Deep Research and Idea Generation.

Lyria in Google Vertex AI generates 30-second music clips from text prompt.

Google Workspace Flows enables users to create agentic workflows to streamline processes and automate repetitive tasks across Workspace.

@CodeAssist from a Kanban board

In Google Docs, users can create audio versions of their documents, including generating podcast-style overviews.

Other new Workspace AI capabilities include an AI writing tool in Google Docs,

New editing and camera control features in Veo 2 (preview) Veo 2 image generation in Google Vids.

Instant Custom Voice in Chirp 3 to enable users to create custom voices from 10 seconds of audio

Improved image generation and inpainting capabilities in Imagen 3 for fixing missing portions of an image or removing objects

new Gemini capabilities in Google Chat, and

AI-powered data analysis and surfacing of key insights in Google Sheets.

Google A2A

At Google’s Cloud Next 25 conference on April 9, VIDEO: Google announced Agent Development Kit (ADK) that communicates using Google’s Agent2Agent (A2A) protocol for AI agents to coordinate actions in a secure way on top of enterprise applications.

READ: https://google.github.io/A2A/#/

Future capability at https://github.com/google/A2A was developed with support and contributions from over 50 other companies, including Atlassian, Box, MongoDB, Salesforce, ServiceNow. (But not competitors Microsoft, Anthropic, OpenAI). A2A competes with the MCP (Model Context Protocol) API standard defined by Anthropic. https://www.youtube.com/watch?v=voaKr_JHvF4

VIDEO shows how A2A might work in JavaScript calling JSONRPC schemas (like MCP).

A2A is based on an Agent Marketplace to discover and integrate agents.

Within the marketplace Registry, each Agent has its own Agent Card with metadata about the Agent, such as what skills and desires it has.

Each agent can select which LLM model to use (unlike Claude MCP). Each agent is a vendor that provides a service.

A2A needs MCP for the lower level.

References:

MCP (Model Context Protocol)

Think of MCP like HTTP. It’s a behind-the-scenes standard, a way for clients (like AI assistants) and servers (like tools and APIs) to communicate. You don’t get excited about HTTP itself; you get excited about what it enables—browsers like Chrome, apps like Gmail, and the entire modern web. MCP is similar: it’s an enabling layer, not the end-user experience.

Claude AI and ChatGPT , on the other hand, are like Chrome. They’re the part people interact with. While MCP defines how tools are exposed and called by AI agents, https://www.linkedin.com/pulse/mcp-just-beginning-what-actually-powers-ai-productivity-zapier-cy4ic/ Zapier Agents show what’s possible with those capabilities.

MCP provides contextual data to LLMs and AI assistants to solve problems. For instance, you can build an MCP documentation server to offer complete access to your documentation to IDEs and agentic frameworks, just like using llms.txt file.

In the 1980’s game (and movie) “Tron”, the “MCP” is the Master Control Program (“god mode”).

But in real life, in Nov. 2024 Anthropic introduced their “MCP” (Model Context Protocol) API spec that standardizes how external tools and services vendors communicate with LLMs. MCP consists of a set of APIs and library.

Catalogs of MCP servers are at smithery.com and mcpappstore.com to communicate with:

IDEs Cursor and Windsurf provide a GUI to send npx commands to MCP servers which interface with various LLMs (like lego pieces).

mcp-cursor-gui.png

References:

Solo.io has announced the launch of its MCP Gateway, a Model Context Protocol gateway for the cloud native API gateway kgateway (previously called Gloo).

AI Agent Protocol Security

JOKE: The “S” in MCP stands for Security. As in there is no S in MCP.

To prevent unauthorized changes, pin specific versions of MCP server and tools they use.

Have a thorough technical and user-level testing processes to detect misuse.

To be able to see when MCP is misused, ensure that tool descriptions are clearly visible and traces of user-visible and AI-visible instructions. Trace the data lineage of transaction payloads at the data layer…

To instill cross-server protection in guardrails: implement stricter boundaries and dataflow controls between MCP servers.

A2A servers cannot be directly called from clients as a command.

A2A servers are registered in a catalog.

Authentication in A2A is built-in out of the box (unlike MCP).

References:

Multi-Agent Hairball

Microsoft Azure

https://github.com/microsoft/ai-agents-for-beginners

using the VS Code container image in Azure AI Foundry portal, see Based on https://learn.microsoft.com/en-us/azure/ai-studio/how-to/develop/vscode Get started with Azure AI Foundry projects in VS Code.

  1. Register https://aiskillsfest.event.microsoft.com
  2. Add to calendar “AI Skills Fest - 50 Days of AI Learning”
  3. Add to calendar the series https://developer.microsoft.com/en-us/reactor/series/s-1507/

  4. Add to calendar Pamela Fox Office Hours: https://microsoft.github.io/AI_Agents_Hackathon/#office-hours

  5. On Apr 8 attend Microsoft AI Skills Fest https://www.youtube.com/watch?v=tCmgSvloyoI https://arch-center.azureedge.net/Credentials/Microsoft-AI-Skills-Fest-Map.pdf https://microsoft.github.io/AI_Agents_Hackathon/
  6. Confirm attendance for the Guiness World Record.
  7. Work on hackathon until 30, 2025

http://aka.ms/sk/recordings

https://learn.microsoft.com/en-us/training/modules/prepare-azure-ai-development/4-azure-ai-foundry https://microsoftlearning.github.io/mslearn-ai-studio/Instructions/01-Explore-ai-studio.html

https://learn.microsoft.com/en-us/collections/ee4ofotexqd15y?wt.mc_id=aiagenthack_hackathonlearningresource_webpage_cxp

Azure AI in Python

  1. Get an Azure subscription.
  2. Sign in (using Passkey, if possible)
  3. PROTIP: Create CLI environment to enable/update Azure Core “az” commands: 
    brew info azure-cli
==> Dependencies
Build: pkgconf ✘, rust ✘
Required: libsodium ✔, libyaml ✔, openssl@3 ✔, python@3.12 ✔
...
brew install azure-cli
  4. PROTIP: Create a folder for the GitHub account you want to use: 
    mkdir <em>folder_name</em>
cd <em>folder_name</em>
  5. PROTIP: Look at my https://github.com/bomonike/agentic/blob/master/az-agents.py which expects these external dependencies to be loaded: 
    pip install azure-ai-projects
pip install azure-identity
pip install azure-ai-textanalytics

    TODO: requirements.txt

  6. In a Edge web browser, go to https://ai.azure.comaif-menu-364x1146.png, the Azure AI Foundry portal. The portal’s menu is shown at left.

  7. Create an Azure AI Hub to share resources among multiple projects that use solution development features and capabilities of AI Foundry.

    A project is an organizational container with tools for AI customization and orchestration. In a project, you can organize your work, save state across different tools like prompt flow, and collaborate with others.

    CAUTION: The following resources are created for each hub:

    • A multi-service Azure AI services resource to provide access to Azure OpenAI and other Azure AI services.
    • A Key vault in which sensitive data such as connections and credentials can be stored securely.
    • A Storage account for data used in the hub and its projects. Optionally,
    • an Azure AI Search resource that can be used to index data and support grounding for generative AI prompts.
    • an Azure AI Face resource
  8. PROTIP: In your user home folder, create edit python.env file by manually transfer from Azure AI Foundry portal: https://ai.azure.com/build/overview https://ai.azure.com/managementCenter/project/overview? 
    AZA_REGION="eastus2"
AZA_PROJECT="wmar-7833"
AZA_HUB_DISPLAY_NAME="wmar_ai-8943"
AZA_HUB_RESOURCE_NAME="wmar_ai-8943"
AZA_PROJECT_NAME=""
  9. Based on LEARN: SDK 
    #project_connection_string = "<region>.api.azureml.ms;<project_id>;<hub_name>;<project_name>"
project_connection_string = f"{AZA_REGION}.api.azureml.ms;{AZA_PROJECT};{AZA_HUB_NAME};{AZA_PROJECT_NAME}"
project_client = AIProjectClient.from_connection_string(
      credential=DefaultAzureCredential(),
      conn_str=project_connection_string,
   )

    The above is dependes on:
    ``` from azure.ai.projects import AIProjectClient from azure.ai.projects.models import ConnectionType from azure.identity import DefaultAzureCredential …

  10. Create Azure OpenAI : https://portal.azure.com/#create/Microsoft.CognitiveServicesOpenAI

    “Azure OpenAI Service provides access to OpenAI’s powerful language models, including all the latest OpenAI models. These models can be easily adapted to your specific tasks, including but not limited to content generation, summarization, image understanding, semantic search, and natural language to code translation. Top use cases include Call Centers, Virtual Assistants, Accessibility, Content Generation, and Code Development. The service also features the Assistants API, Fine Tuning capabilities and many ways to connect your data to the service for conversational experiences. The service can be scaled through Standard (tokens) and Provisioned (PTUs) deployment types.

https://learn.microsoft.com/en-us/training/modules/ai-foundry-sdk/03-connections?pivots=python Each resource is a connection to an external service, such as Azure AI services, Azure storage, Azure AI search, and others.

Azure Hackathon

Learn more about the Hackathon and read the official rules at https://microsoft.github.io/AI_Agents_Hackathon/ 8 April, 2025 | 10:00 AM - 12:30 AM (UTC-06:00)

  1. Sign in to start OFFICIAL PLAN: https://learn.microsoft.com/en-us/plans/0n5kb1t56pzndr 6 milestones in this plan

    Learn how to navigate and use Azure AI Foundry and Azure AI Foundry SDK. Create a custom Retrieval-Augmented Generation (RAG) solution. Fine-tune and tailor models.

    https://learn.microsoft.com/en-us/plans/0n5kb1t56pzndr?source=docs

  2. https://www.youtube.com/watch?v=x6g5JutmHXk Azure AI Agent Service using Azure AI Foundry SDK Demo

    https://www.youtube.com/watch?v=4m0MYnQnDo8&list=PLLasX02E8BPCRLOzJHUO39cSR_Bn6cvw8 playlist

    https://www.linkedin.com/in/shivam2003/ Shivam Goyal

https://gateway.on24.com/wcc/eh/4304051/category/142477/ai-agents/?wt.mc_id=1reg_25323_webpage_reactor Join TheSource EHub to explore top picks including trainings, livestreams, repositories, technical guides, blogs, downloads, certifications, and more, all updated monthly. The AI Agent section offers essential resources for creating AI Agents, while other sections provide insights into AI, development tools, and programming languages

https://learn.microsoft.com/en-us/collections/ee4ofotexqd15y?wt.mc_id=aiagenthack_hackathonlearningresource_webpage_cxp LEARN: Learning Resources for the AI Agents Hackathon

https://github.com/microsoft/ai-agents-for-beginners?tab=readme-ov-file#ai-agents-for-beginners---a-course

https://developer.microsoft.com/en-us/reactor/events/25324/ Agents 101 / AI Agents Hackathon Kickoff! 8 April, 2025 | 10:00 AM - 11:00 AM (UTC-06:00)

https://aka.ms/CreateAgenticAISolutions https://aka.ms/ai_agents_apps__data

https://developer.microsoft.com/en-us/reactor/events/25325/ Build your code-first app with Azure AI Agent Service (EMEA/US offering) 9 April, 2025 | 10:00 AM - 11:00 AM (UTC-06:00) Mountain Time (VS en Canada)

https://developer.microsoft.com/en-us/reactor/events/25312/ Build your code-first app with Azure AI Agent Service (APAC offering) 9 April, 2025 | 4:00 PM - 5:00 PM (UTC-06:00) Mountain Time (VS en Canada) Dave Glover https://aka.ms/CreateAgenticAISolutions/?wt.mc_id=1reg_25312_webpage_reactor Create agentic AI solutions by using Azure AI Foundry

https://aka.ms/ai_agents_apps__data/?wt.mc_id=1reg_25312_webpage_reactor Practical Foundation for AI Agents: A Developer’s Guide on Azure AI Foundry, Apps, and Data

https://developer.microsoft.com/reactor/events/25313 Transforming Business Processes with Multi-Agent AI using Semantic Kernel 10 April, 2025 | 1:00 PM - 2:00 PM (UTC-06:00) Mountain Time (US & Canada)

Extending AI Agents with Azure Functions 24 April, 2025 | 10:00 AM - 11:00 AM (UTC-06:00)

https://developer.microsoft.com/en-us/reactor/events/25364/ Securing AI agents on Azure by Joylynn Kirui & Pamela Fox 29 April, 2025 | 10:00 AM - 11:00 AM (UTC-06:00) Mountain Time (US & Canada)

https://aka.ms/model-mondays - Explore All Resources https://aka.ms/model-mondays/chat - Continue The Conversation https://aka.ms/model-mondays/collection - Explore The Model Cards https://aka.ms/model-mondays/newsletter - Get The Weekly Recaps

Marlene Mhangami Korey Stegared-Pace

Anything LLM

https://www.youtube.com/watch?v=Z6kK4DXqCOI

Calendar Task Apps

ADM (Automated Decision Making)

DEFINITION: Tasks https://practicallyintelligent.com/

Benefits

AI Agentic Utilities

Rather than just GitHub star history, https://www.youtube.com/watch?v=odbU0pGMxzs

Code & Labs

https://github.com/sinanuozdemir/oreilly-ai-agents https://learning.oreilly.com/live-events/ai-agents-a-z/0642572007604/ Sinan Ozdemir, founder and CTO of LoopGenius, where he uses state-of-the-art AI to help people create and run their businesses. He has lectured in data science at Johns Hopkins University and authored multiple books, videos. makes use of Crew.ai

Agencies

Arseny Shatokhin

<a target="_blank" href="https://www.youtube.com/watch?v=SjxYuaq3hFU">VIDEO</a>: Limitless AI built a Virtual Assistant AI Agent that can do ANYTHING for Me [Step-by-step Walkthrough]

Agency templates

Yiannis Demetriou

https://www.youtube.com/watch?v=fIN2vIOTsZs Build The ULTIMATE AI Content Agent Using Make & N8N (NO-CODE) by Yiannis Demetriou

https://www.youtube.com/watch?v=kAk131n5aA4

https://www.skool.com/ai-automations-mastery-7642/classroom/0c39a268?md=91e6a57596b54616968beb2a9838923a AI Automation Agency Wizards

Andrew Ng

https://www.youtube.com/watch?v=KrRD7r7y7NY Andrew Ng Explores The Rise Of AI Agents And Agentic Reasoning | BUILD 2024 Keynote

Explainer Youtubers

https://www.youtube.com/watch?v=2-JV8UxegWE What is AI Agent? | Simple Explanation of an AI Agent by codebasics

https://www.youtube.com/watch?v=IivxYYkJ2DI AI Agents and AI Assistants: A Contrast in Function by IBM Technology

https://www.youtube.com/watch?v=eDnA1pVPS64 Mastering the AI Agent Stack: Building Smarter, Scalable Systems by sam_mk87

https://www.youtube.com/watch?v=93fCvFkY1Lg Product Management Is Dead, So What Are We Doing Instead? | Lenny & Friends Summit 2024 Lenny’s Podcast Claire Vo @clairevo is the Chief Product Officer at LaunchDarkly and the creator of ChatPRD. https://www.chatprd.ai/blog/product-management-is-dead She previously held leadership roles at Optimizely and Color. Claire is known for her forward-thinking approach to product management and her expertise in AI-powered product development. In this talk, she discusses: • Why product management as we know it is dying • How AI is transforming product development faster than we expect • The rise of AI-powered “triple threats” who can handle product, design, and engineering • What product leaders need to do to stay relevant in the AI era • How to build and manage AI-powered product teams

https://www.youtube.com/watch?v=KY8n96Erp5Q Why Agent Frameworks Will Fail (and what to use instead) Dave Ebbelaar

https://www.youtube.com/watch?v=kJvXT25LkwA How I Made AI Assistants Do My Work For Me: CrewAI Maya Akim

https://www.youtube.com/watch?v=pcWKJ_yvf1M Microsoft Launches 10 NEW AI Agents Sam Witteveen

AI Automation Groups on Skool

$99/month Skool was founded in 2019 by Sam Ovens (CEO) and Daniel Kang (CTO). Skool has 30 employees and is based in Los Angeles, California. In 2024 Alex Hormozi added https://www.skool.com/games/about

Some lures:

Listed by cost and name:

https://www.skool.com/ai-automations-by-jack-4235/about https://www.skool.com/ai-automations-by-jack-4235/classroom/0c39a268?md=91e6a57596b54616968beb2a9838923a $67/month AI Automation by Jack Roberts

https://www.skool.com/aiworkshop $59/m AI Workshop

https://www.skool.com/ai-biz-in-a-box $497/m AI Biz In A Box

https://www.skool.com/ai-collective $9/m AI Collective

NCA & AI Guy - focused at faceless YT channels.

https://www.skool.com/learn-ai FREE AI Automation Agency Hub

https://www.skool.com/learn-ai/about FREE AI Automation Agency Hub by Liam Ottley

https://www.skool.com/insightai-academy-3338 FREE InsightAI by Nicholas Vasiliu @nicholas-vasiliu-3720

https://www.skool.com/cyberpro-community FREE CYBERPRO COMMUNITY (FREE)

https://www.skool.com/ai-automations-mastery-7642 FREE AI Automation Agency Wizards

https://www.skool.com/systems-to-scale-9517 FREE Systems to Scale

https://www.skool.com/ai-automation-society FREE AI Automation Society

https://www.skool.com/insightai-academy-3338 FREE InsightAI Academy

https://www.skool.com/agent-zero FREE Agent Zero

https://www.skool.com/data-alchemy FREE Data Alchemy

https://www.skool.com/ai-for-devs-6115 FREE AI for Devs

https://www.skool.com/voice-ai https://www.skool.com/voice-ai/classroom/e5a56130?md=61dd6abdeb9a4ce880b7c57c1fa779b5 FREE Voice AI Accelerator (Free)

https://www.skool.com/automated-client-acquisition FREE ACA (free)

https://www.skool.com/aiacademyskool FREE Ai Academy

https://www.skool.com/facelessyoutubefree FREE Faceless YouTube Free

https://www.skool.com/aiography FREE AIography - AI Film Academy

https://www.skool.com/lazyempire FREE TaskMagic’s Lazy Empire

https://www.skool.com/godago Free AI Productivity Hub

https://www.skool.com/aiagency Free AI Agency Incubator

References

by Hasan Aboul Hasan

by Aaron Dunn

https://www.youtube.com/watch?v=WP5sQhGlxj4 The Secret Plan Behind Artificial Intelligence by More Perfect Union

https://www.youtube.com/watch?v=mlz0paWRqac How to 25X Your Business with AI in 2025 (Complete Blueprint) by Julia McCoy

https://www.youtube.com/watch?v=O2YmiYEJ-sY The Proven Content System for Unlimited AI Agency Leads (copy this) by Liam Ottley

https://www.youtube.com/watch?v=YqeI7iViRmE Marc Andreessen: It’s Morning Again In America Hoover Institution

https://www.youtube.com/watch?v=Xix3HHhyeoI AI Agents are about to join your company en masse by Matthew Berman

https://www.youtube.com/watch?v=G7S1R5o_hWs 5 AI Business Ideas For 2025 – These Make Me $50K/Month Sara Finance

https://ibmtechxchange-virtual-agents-platform.bemyapp.com/#/event

https://www.youtube.com/watch?v=F8NKVhkZZWI&t=165s IBM:

https://www.youtube.com/watch?v=ZL14jkX39G0 The Evolution of AI Agents by Bloomberg Technology

https://www.youtube.com/watch?v=2-JV8UxegWE What is AI Agent in Artificial Intelligence? | AI Agent Explained by codebasics

https://www.youtube.com/watch?v=w-A70c_to5A I Left My Dream Job at Mckinsey: Here’s Why by Wayne Hu

Anthropic “How to build effective agents” * Workflows have predefined code paths. * AI Systems have dynamic processes and tool usage.

Data-Alchemy shop

VIDEO: by Dave Ebbelaar in Amsterdam, Netherlands at free https://www.skool.com/data-alchemy/classroom and https://www.datalumina.com/data-freelancer provided in ?499 course at https://launchpad.datalumina.com/ referencing modules explained in https://www.youtube.com/watch?v=p4G0coRey9w https://www.skool.com/data-freelancer/about

PDF Parsers:

PydanticAI & Anthropic

studio.auttomator.ai hackathon of agents https://www.youtube.com/watch?v=xfFyAumTjT0

Objectives?

BTW In the movie “2001: A Space Odessey”, the computer named HAL 9000 murders an astronaut to resolve conflicting objectives: Deliver astronauts to Jupiter and NOT let them know what awaits them there. So the computer decides that delivering dead astronauts would solve both objectives.

Automation

Healthcare: In healthcare, automation is increasingly important. From robots helping with surgery to computer programs that help diagnose diseases, technology is helping healthcare workers provide faster, more accurate, and more personalized care. This gives doctors and nurses more time to spend with patients, providing emotional support and addressing complex medical needs. Think of how AI can analyze medical images to detect early signs of cancer, potentially saving lives through early intervention. A robot assisting a doctor during surgery

Finance: The finance industry has also embraced automation to make things run smoother and more efficiently. Algorithmic trading systems execute trades on the stock market incredibly fast, taking advantage of market fluctuations that a human trader might miss. AI-powered tools help protect customers from financial crimes by analyzing patterns and flagging suspicious activity. Chatbots and virtual assistants handle simple customer inquiries, like checking account balances or resetting passwords, which frees up human agents to deal with more complicated issues that require empathy and problem-solving skills.

Transportation: Automation is changing the transportation industry in big ways. Self-driving cars could completely change how we get around, making roads safer by eliminating human error and reducing traffic congestion by optimizing routes and speeds. Automated traffic management systems can adjust traffic lights in real-time to keep traffic flowing smoothly, even during rush hour. Drones can deliver packages and medical supplies, overcoming geographical barriers and providing access to essential goods and services.

Here today

Agents are here today from leading AI developers, including

References

https://lakefs.io/blog/what-is-rag-pipeline/

https://conted.ox.ac.uk/courses/artificial-intelligence-generative-ai-cloud-and-mlops-online £3595 Artificial Intelligence: Generative AI, Cloud and MLOps (online) course at the University of Oxford January - April 2025 by Ajit Jaokar Based on Machine Learning Engineering with Python: Manage the lifecycle of machine learning models using MLOps with practical examples, Second Edition by Andrew P. McMahon

mementic replace all QA. muuktest Raleigh, NC https://www.linkedin.com/in/ivanbarajasvargas/

Barajas Vargas https://www.youtube.com/watch?v=8vrA-XFfyvQ https://www.youtube.com/watch?v=j6N15PWL-UA

gigaML

https://hbr.org/topic/subject/ai-and-machine-learning

https://www.youtube.com/watch?v=ju9PmFl5L0c I tried 50+ Agent libraries, the only one I really needed is this BiasVsVariance

Read More - https://aka.ms/entity1 MS LEARN: Custom named entity recognition

https://colivara.com

## https://learn.microsoft.com/en-us/training/modules/build-copilot-ai-studio/ LEARN: Build a RAG-based agent with your own data using Azure AI Foundry

Introduction 2 min Understand how to ground your language model 6 min Make your data searchable 7 min Build an agent with prompt flow 7 min Exercise - Create a custom agent that uses your own data 15 min Knowledge check 3 min Summary 1 min

Adding grounding data to an Azure AI project You can use Azure AI Foundry to build a custom age that uses your own data to ground prompts. Azure AI Foundry supports a range of data connections that you can use to add data to a project, including: Azure Blob Storage Azure Data Lake Storage Gen2 Microsoft OneLake You can also upload files or folders to the storage used by your AI Foundry project. in Playgrounds

https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/understand-embeddings embeddings in the Azure OpenAI Service

https://learn.microsoft.com/en-us/azure/search/vector-search-overview vector search

https://learn.microsoft.com/en-us/azure/search/hybrid-search-overview hybrid search

Agent-to-Agent Interaction

https://www.youtube.com/watch?v=rAeqTaYj_aI

Prompt Flow

Agentic uses a “flipped” pattern of user interaction. The AI system asks users questions to gather information and perform tasks.

Prompt Flow is a development framework for defining flows that orchestrate interactions with an LLM.

After uploading data to Azure AI Foundry and creating an index on your data using the integration with Azure AI Search, you can implement the RAG pattern with Prompt Flow to build a generative AI application.

  1. A flow begins with one or more inputs, usually a question or prompt entered by a user, and in the case of iterative conversations the chat history to this point.
  2. The flow is then defined as a series of connected tools, each of which performs a specific operation on the inputs and other environmental variables. There are multiple types of tool that you can include in a prompt flow to perform tasks such as:
    • Running custom Python code
    • Looking up data values in an index using an Index Lookup tool to retrieve data from an index so that subsequent tools in the flow can use the results to augment the prompt used to generate output from an LLM. https://learn.microsoft.com/en-us/azure/machine-learning/prompt-flow/tools-reference/index-lookup-tool
    • Creating prompt variants - enabling you to define multiple versions of a prompt for a large language model (LLM), varying system messages or prompt wording, and compare and evaluate the results from each variant.
    • Submitting a prompt to an LLM to generate results.
  3. Finally, the flow has one or more outputs, typically to return the generated results from an LLM.

Coursera course

“AI Agent Developer on Coursera.com by Vanderbilt University’s Jules White in Nashville, Tennessee.

What is the difference with agentic AI?

Agentic AI is characterized by its ability to perform tasks without any human intervention.

Types of AI agents

UI

https://medium.com/@amosgyamfi/the-3-best-python-frameworks-to-build-uis-for-ai-apps-08619d53d281 The 3 Best Python Frameworks To Build UIs for AI Apps Build user-friendly AI chat interfaces crucial for seamless interactions, demo presentation, and testing using the leading Python frameworks for creating intuitive chat UIs for AI.

Code

Here is an overview of our code base at https://github.com/bomonike/agentic

  1. Install Ollama server
  2. Install LLAMA 3.2 and other LLMs
  3. Run Chatbot within benchmarks
  4. Add memory

v020 + qrcode :ai-agents.md