Agent Sense | Agentic Workflows & Operational AI

𝐘𝐨𝐮𝐫 𝐦𝐨𝐧𝐢𝐭𝐨𝐫𝐢𝐧𝐠 𝐝𝐞𝐭𝐞𝐜𝐭𝐞𝐝 𝐭𝐡𝐞 𝐢𝐧𝐜𝐢𝐝𝐞𝐧𝐭. 𝐖𝐡𝐲 𝐝𝐢𝐝 𝐢𝐭 𝐬𝐭𝐢𝐥𝐥𝐭𝐚𝐤𝐞 the Engineers 𝐡𝐨𝐮𝐫𝐬 𝐭𝐨 𝐟𝐢𝐧𝐝 𝐭𝐡𝐞 root cause?

Engineering teams have dashboards, alerts, logs, and traces. Yet finding theroot cause of a production issue can still take hours.

Frank Chávez and I spoke with Viraat Das, Cofounder and CEO of Exla, about Libra, an ambient AI agent that continuously investigates production systems, identifies regressions, and proposes fixes.

Viraat’s recommendation for enterprise architects, and SRE teams while designing agentic workflows:

“The harder it is to reverse an action, the more a human shouldbe in the loop.”

Our main discussion points:

- Why incident detection is easier than root cause investigation- How ambient agents connect logs, traces, deployments, and code changes
-Detecting regressions that traditional monitoring may miss
- When agents should act autonomously and when human approval is necessary

Agents should be able to investigate complex problems autonomously, but execution should follow defined policies, permissions, and risk controls.

Your AI agent works, but can you prove its business value?

As enterprises invest in Agentic AI, AI automation, and autonomous AIagents, many struggle to demonstrate measurable ROI. The challenges oftenbegin before development, with unclear business requirements and poorlyunderstood workflows.

In this episode, hosts Monika Aggarwal and Frank Chavez speak withRobert Maxwell, Founder and CEO of TGT Solutions, who brings over 30years of enterprise technology experience.

Robert shares practical insights on AI agent implementation, businessprocess automation, production readiness, AI governance, and measuring AI ROI.We discuss why successful demos do not always translate into business value andwhat enterprises need to address before scaling AI agents.

The key question: How do you move from building AI agents to delivering measurablebusiness outcomes?

🎙️ Agent Sense: Practical conversations about Enterprise AI.

🌐 agentsensepodcast.com

#AgenticAI #AIAgents #EnterpriseAI #AIROI #AIAutomation

𝐀𝐠𝐞𝐧𝐭 𝐒𝐞𝐧𝐬𝐞 𝐢𝐬 𝐚 𝐬𝐡𝐨𝐫𝐭 𝐩𝐨𝐝𝐜𝐚𝐬𝐭 𝐨𝐧 𝐦𝐚𝐤𝐢𝐧𝐠 𝐀𝐈 𝐬𝐢𝐦𝐩𝐥𝐞, 𝐩𝐫𝐚𝐜𝐭𝐢𝐜𝐚𝐥, 𝐚𝐧𝐝 𝐠𝐫𝐨𝐮𝐧𝐝𝐞𝐝

An AI agent can be 95% accurate and still fail thebusiness. It can give the right answer and still take too long, use the wrong tool, cost too much, or fail to complete the task.

Frank Chávez and I talk with Jay Migliaccio, TechnicalProduct Manager for IBM watsonx Orchestrate, about how enterprises shouldmeasure and improve agent performance.

Jay Migliaccio has spent his career in product and salesroles across early-stage software companies, and is now building agentic solutions for enterprise customers. That builder's view shapes everything below.

Themes of the episode:

🔹 Why accuracy alone hides agent failures

🔹 How to measure outcomes, quality, operations,and user experience together

🔹 How observability traces a failure to the model,instructions, tools, workflow, data, or integration — IBM watsonx Orchestrateshipped a full observability dashboard this year for exactly this

🔹 Why teams need measurable goals and aperformance baseline before launch

Jay's diagnostic approach: start with the businessoutcome, trace the full execution, find the real failure point before touchingthe model or the prompt.

The Agent Performance principle:

𝑴𝒆𝒂𝒔𝒖𝒓𝒆𝒕𝒉𝒆 𝒐𝒖𝒕𝒄𝒐𝒎𝒆→ 𝑻𝒓𝒂𝒄𝒆 𝒕𝒉𝒆 𝒆𝒙𝒆𝒄𝒖𝒕𝒊𝒐𝒏→ 𝑭𝒊𝒏𝒅 𝒕𝒉𝒆 𝒇𝒂𝒊𝒍𝒖𝒓𝒆𝒑𝒐𝒊𝒏𝒕 → 𝑰𝒎𝒑𝒓𝒐𝒗𝒆𝒂𝒈𝒂𝒊𝒏𝒔𝒕 𝒂 𝒃𝒂𝒔𝒆𝒍𝒊𝒏𝒆

Agent Sense web page and all episodes: 🔗agentsensepodcast.com

#AgentSense #EnterpriseAI #AgentObservability

⚠️ Disclaimer: The views shared are based on our personalexperience and do not represent the views of IBM.

Season 2, Episode 1 - Agent Memory Is More Than RAG

AI agents can access huge amounts of enterprise data, yet they often forget the context that matters.

In this episode of Agent Sense, Monika Aggarwal and Frank Chavez talk with Ran Aroussi, founder and lead architect of MUXI, about what it takes to build useful agent memory.

We discuss why storing data and using RAG does not create memory, what agents should remember, how memory can be distilled and updated over time, and why enterprise memory needs context, relationships, time, and source traceability.

Ran also shares his perspective on building memory as an organizational capability, with access controls and governance built into the architecture.

A practical conversation for architects, engineers, and AI leaders building agents for production.

Guest: Ran AroussiFounder and Lead Architect, MUXICreator of yfinance

Agent Sense is hosted by Monika Aggarwal and Frank Chavez.

#AgentSense #AIAgents #AgentMemory #RAG #EnterpriseAI #MUXI

Most AI agents wait for a person to open a chat and give them a prompt. Ambient agents work differently. They listen for events across enterprise systems and start working when something changes.

In this episode of Agent Sense, Monika Aggarwal and Frank Chavez discuss what makes an agent ambient and how the architecture changes when AI works continuously in the background.

Using an SRE Ambient Agent as a mature example, they discuss how an agent can monitor operational signals, investigate incidents, invoke other agents and tools, involve a human when needed, and verify the outcome.

The Ambient Agent design principle:

Start with an event → Bound the action → Earn autonomy.

AI can summarize a meeting. Enterprise work starts after the summary.

In this episode, Monika Aggarwal speaks with Artem Koren, co-founder and Chief Product Officer of Sembly AI, about how meeting context can move beyond notes and trigger real work.

They discuss how AI can support follow-ups, workflow updates, CRM actions, and executive reporting across teams, countries, languages, and meeting platforms.

Artem also explains why humans must steer agent execution and why enterprises should start with a controlled process, clear success measures, and a specific business outcome.

Agent Sense keeps enterprise AI simple, practical, and grounded.

Agent Sense reflects the personal perspectives and experiences of its hosts and guests. It is not an official IBM podcast.

Agent Sense — Episode 8: Governed Catalog of AI Assets

Enterprise AI agents does not fail from lack of ideas. They fail from agent sprawl: duplicate builds, unclear ownership, no way to know what is safe to reuse. Jyotsna Narayanan, Principal Product Manager for watsonx Orchestrate at IBM, joins host Monika Aggarwal to break down what a governed catalog of AI agents, tools, and MCP servers needs to include, how to evaluate reuse readiness using references like OWASP Top 10 and CoSAI, and where to start.

Topics: AI agents, agentic AI, enterprise AI governance, AgentOps, MCP, Model Context Protocol, watsonx Orchestrate, agent catalog, AI reuse, enterprise automation, production AI.

Vendor-agnostic, not monetized, based on field experience. Views are personal and do not represent IBM.

The episode is about hashtag#9minutes long. In this episode, my co host Frank Chávez and I are joined by Vitalii Duk from Dynamiq to discuss how agents are moving from answering questions to doing work across tools, systems, and runtime environments.

“𝑪𝒉𝒂𝒕 𝒎𝒆𝒂𝒏𝒔 𝒂𝒏 𝒂𝒈𝒆𝒏𝒕 𝒄𝒂𝒏 𝒂𝒏𝒔𝒘𝒆𝒓 𝒚𝒐𝒖𝒓 𝒒𝒖𝒆𝒔𝒕𝒊𝒐𝒏. 𝑪𝒐𝒎𝒑𝒖𝒕𝒆𝒓 𝒖𝒔𝒆 𝒎𝒆𝒂𝒏𝒔 𝒕𝒉𝒆 𝒂𝒈𝒆𝒏𝒕 𝒄𝒂𝒏 𝒕𝒂𝒌𝒆 𝒂𝒄𝒕𝒊𝒐𝒏 𝒂𝒄𝒓𝒐𝒔𝒔 𝒂𝒑𝒑𝒍𝒊𝒄𝒂𝒕𝒊𝒐𝒏𝒔, 𝒕𝒐𝒐𝒍𝒔, 𝒂𝒏𝒅 𝒘𝒐𝒓𝒌𝒇𝒍𝒐𝒘𝒔. 𝑻𝒉𝒂𝒕 𝒊𝒔 𝒘𝒉𝒆𝒓𝒆 𝒆𝒏𝒕𝒆𝒓𝒑𝒓𝒊𝒔𝒆 𝑨𝑰 𝒔𝒕𝒂𝒓𝒕𝒔 𝒕𝒐 𝒃𝒆𝒄𝒐𝒎𝒆 𝒐𝒑𝒆𝒓𝒂𝒕𝒊𝒐𝒏𝒂𝒍.”

We share a practical frame:
🔹 Start with bounded, observable, and measurable agent use cases
🔹 Define the agent’s authority boundary before giving it access to systems
🔹 Use orchestration to manage work across tools, agents, approvals, and exceptions
🔹 Add observability, audit, and human review before scaling agent action
🔹 Treat Day 2 and Day 3 readiness as part of the design, not an afterthought

𝙏𝙝𝙞𝙨 𝙞𝙨 𝙩𝙝𝙚 𝙚𝙣𝙩𝙚𝙧𝙥𝙧𝙞𝙨𝙚 𝙖𝙧𝙘𝙝𝙞𝙩𝙚𝙘𝙩𝙪𝙧𝙚 𝙨𝙝𝙞𝙛𝙩: 𝙖𝙜𝙚𝙣𝙩𝙨 𝙖𝙧𝙚 𝙢𝙤𝙫𝙞𝙣𝙜 𝙛𝙧𝙤𝙢 𝙧𝙚𝙨𝙥𝙤𝙣𝙨𝙚 𝙩𝙤 𝙚𝙭𝙚𝙘𝙪𝙩𝙞𝙤𝙣. 𝙏𝙝𝙚 𝙫𝙖𝙡𝙪𝙚 𝙞𝙨 𝙣𝙤𝙩 𝙤𝙣𝙡𝙮 𝙞𝙣 𝙬𝙝𝙖𝙩 𝙩𝙝𝙚 𝙖𝙜𝙚𝙣𝙩 𝙘𝙖𝙣 𝙙𝙤. 𝙄𝙩 𝙞𝙨 𝙞𝙣 𝙝𝙤𝙬 𝙨𝙖𝙛𝙚𝙡𝙮, 𝙤𝙗𝙨𝙚𝙧𝙫𝙖𝙗𝙡𝙮, 𝙖𝙣𝙙 𝙧𝙚𝙥𝙚𝙖𝙩𝙖𝙗𝙡𝙮 𝙞𝙩 𝙘𝙖𝙣 𝙤𝙥𝙚𝙧𝙖𝙩𝙚 𝙞𝙣 𝙧𝙚𝙖𝙡 𝙗𝙪𝙨𝙞𝙣𝙚𝙨𝙨 𝙬𝙤𝙧𝙠.

Episode 6 of Agent Sense continues from Episode 5, where we talked about MCP, A2A, and enterprise integration.

In this episode, Monika Aggarwal and Frank Chávez discuss why enterprise agents need governed access to core systems before they can scale in production.

MCP helps agents connect to tools and systems. But connection alone is not enough. As agents start working across ServiceNow, Workday, SAP, HR, IT, finance, and customer operations, enterprises need a control layer.

That is where the MCP Gateway comes in.

We discuss how an MCP Gateway helps manage identity, policy, approvals, audit, and traceability. It gives agents access to approved tools without opening direct, unmanaged paths into core enterprise systems.

In about 4 minutes, we cover:

🔹 Why direct agent access to core systems creates risk
🔹 How MCP Gateway supports controlled enterprise access
🔹 Why public MCP servers are useful for testing, but not enough for production
🔹 How approved tools help agents scale across business workflows
🔹 Why traceability matters when agents take action

Episode 5 was about connection.
Episode 6 is about controlled access.

Disclaimer: The views shared are based on our personal experience and do not represent the views of IBM.

Tags for Spotify search:

Agentic AI, Enterprise AI, MCP, MCP Gateway, AI agents, ServiceNow, Workday, SAP, AI governance, agent governance, operational AI, enterprise architecture, AI integration, agentic workflows.

Theme: As AI systems evolve from single models into networks of autonomous agents, integration is the primary bottleneck.

Core Integration Challenges:

1. Fragmented Tool Access

2. Context Loss Across Agents

3. Tight Coupling & Low Reusability

4. Lack of Standardized Communication
   

   Integration needs standards.
   MCP standardizes how agents connect to systems.
   Agent to agent communication standardizes how they pass work.

   
   

   🔷 MCP connects agents to enterprise systems.
   🔷 A2A connects agents to each other so work can move across the enterprise.