My mental excercise starts today by reading a blog post whose title I captured some week ago.
"Services: the new software"
The author Julien Bek provided us with some critical insight and analysis upon the ongoing changes on the Agentic AI market. First I will quote some of his most insightful and concise summaries from his blog. Later we will look at how this is impacting and changing the very specific IT Service Management and Observability space. What are the latest movement there related to Juliens observations.
1. "Writing code is mostly intelligence. Knowing what to build next is judgement".
As another blog author points out, traditional relationship between human-being and software defines software and application as the tool where we build in intelligence, while we leave the decision and judgement to people.
What we are seeing right now is AI is taking over the intelligence work rapidly in multiple domains, with software engineering at most (over 49% of tool calls are made by agent). Other domains will follow, like legal, finance, accounting and customer service.
2. "A copilot sells the tool. An autopilot sells the work"
As a support the AI tools were put into the hands of IT professional to increase productivity and efficiency, but with AI agents the AI tools are no longer a copilot but an autopliot to accomplish the complete work for customer. It delivers outcome, and this is what a service is!
Many SaaS tools live on a copilot model to help professionals conduct data administration more efficiency, with build-in standard and automation. While with Agentic AI tools custom build or AI-agent native vendor, it delivers the outcome to customer instead of just support. Like accounting software has been a help to do the book keeping, while Agentic AI will deliver the book closing as well.
3. "The higher the intelligence ratio in any field, the sooner autopilots will win."
For different expert domains, the more specialised it is, the most complicated topic is, the greater extend customer will appreciate an autopilot. This is valid for traditional highly specialist areas like Legal, Tax advisory, Insurance brokerage as well as IT managed services. In plain English, the less customer understands the problem, the more likely customer will jump on a AI agent journey as long as the outcome is equivalent.
4. "Today’s judgement will become tomorrow’s intelligence. "
As AI agent improves every day in understanding the context and correcting the historical mistakes, more and more of the judgement part can as well convert into rules and conditions for AI to automate. This is the same journey when computers start to play chess.
For the future convergence between the judgement and intelligence, what areas will be easy to do first? The author Julien Bek gives his forecast.
If a task is outsourced today, it is likely to be on top of list for AI autopilot. Why?
The reason is simple:
1. Customer has accepted the work to be done externally
2. There is clear scope and budget for that
3. The buyer is already purchasing an outcome
"Replacing an outsourcing contract with an AI native service provider is a vendor swap. Replacing headcount is a reorg".
Julien has even provided an opportunity map for the business domains where AI automation has the greatest potentials and quickest gains with autopilot.
Now coming to our domain area of IT Service Management and Observability, what are the current developments for the vendors to take on this co-pilot to autopilot transition?
I asked again Claude to provide me with a summary of the major functionalities as well as availabilities.
My takeaway after summarizing all this info:
• If you want to build a native AI agent driven product that delivers outcome instead of selling as a tool, Julien has pointed out the opportunity
• The convergence of copilot to autopilot means that observability tooling is more and more focused on delivering the most important service outcome - stability and uptime with the help of AI agents and less manual intervention.
• This is what people has been buying IT Service Management tooling and process to achieve (though manually), which means the more monitoring becomes automated and AI driven, the less dependent IT organization will have on the ITSM tool functionality. The close cooperation between ServiceNow and Dynatrace shows ServiceNow is feeling the heat.
