One of the most valuable observations from Yann Bordelanne's deep dive into IBM Bob is that the conversation around AI is evolving.
The real question is no longer:
Can AI generate code?
The more meaningful question is:
How can AI help technical teams solve real-world problems more effectively?
To answer that, Yann evaluated IBM Bob against practical IBM Maximo scenarios that developers, administrators, and consultants encounter daily. Rather than focusing on simple code generation, he explored how Bob could support troubleshooting, automation, documentation, configuration management, integration design, and eventually direct interaction with Maximo itself.
One lesson appeared repeatedly throughout the article:
The quality of the output depends heavily on the quality of the context provided.
When Bob was asked to generate a simple SQL script, it produced a technically reasonable response. However, when supplied with additional information such as work order details, business rules, naming conventions, and existing internal scripts, the quality and relevance of the response increased significantly.
The same pattern appeared across multiple areas:
Good context transformed generic answers into practical deliverables.
Yann tested IBM Bob against several common Maximo tasks.
Maintaining BIRT reports often requires navigating unfamiliar datasets and report structures.
Bob helped identify components, understand report logic, and suggest modifications such as:
While human review remains essential, AI can reduce the time spent exploring report structures.
Application XML files can be complex and difficult to navigate.
Bob was used to locate specific controls, understand surrounding structures, and assist with modifications such as adding fields or changing layouts.
Even small XML updates become easier when the assistant understands the broader structure of the configuration.
One particularly interesting scenario involved investigating a Java library conflict.
Instead of manually reviewing extensive error logs, Bob helped identify:
This doesn't replace experienced engineers, but it can accelerate the early stages of troubleshooting.
Maximo teams frequently need to translate existing business logic into automation scripts.
By providing Java implementations as context, Bob was able to assist with the creation of Python and Jython automation scripts that align with Maximo conventions.
Perhaps the most exciting part of the article explores the future potential of IBM Bob through the Model Context Protocol (MCP).
Instead of simply generating Maximo-related code, Bob could potentially interact directly with authorized Maximo capabilities.
Using Maximo Application Suite 9.2 MCP support, future workflows could include requests such as:
The workflow becomes:
User Request
↓
IBM Bob
↓
Maximo MCP Server
↓
Authorized Maximo Tool
↓
Validation & Security Checks
↓
Result or Approved Action
The important detail is that actions remain governed by Maximo's existing business logic, security controls, permissions, and validation mechanisms.
A major theme throughout Yann's analysis is responsible AI adoption.
Direct database access is rarely the right approach.
Instead, Bob should interact through:
Additional safeguards include:
This ensures AI becomes a controlled participant in enterprise workflows rather than an unrestricted actor.
What I found most interesting about Yann's article is that it frames IBM Bob as an augmentation tool rather than a replacement for developers, consultants, administrators, or architects.
Today, Bob can assist with:
Tomorrow, through MCP and controlled enterprise integrations, it may also become a conversational gateway to business systems themselves.
The future of AI in enterprise software isn't just about generating better code.
It's about combining:
Context + Tools + Governance + Human Validation
to create meaningful business value.
For Maximo environments, the combination of IBM Bob, MCP, and secure enterprise controls offers a compelling glimpse into what agentic enterprise management could look like.
Credit: Full credit to Yann Bordelanne for the original article, hands-on experimentation, and detailed exploration of IBM Bob across real IBM Maximo scenarios. This post is a summarized and paraphrased interpretation of his findings and insights.