How to Blend AI with Operations
This past Friday, I spent two hours training a team member on the process flow for one of our managed services offerings—a product line I created six years ago. Over time, the processes involved had evolved and been refined, but due to some serious lack of oversight, client satisfaction had dropped significantly. In fact, the client described the service as a waste of money—and after looking into it, I couldn’t disagree.
The solution was clear: we had to get back to basics. When you find yourself in a slump—whether in poker when you’re on tilt or simply feeling uncertain—the best move is to return to your playbook. You rely on proven strategies, re-focus, and regain momentum.
A Blend of People and Technology
As I explained to my team member, our managed services are as much about managing people as they are about managing technology. Yes, we support complex systems like the SailPoint identity management platform, but our real value lies in building confidence through clear, human-to-human communication and well-defined processes.
At this point in time, there isn’t a single AI agent that can plug in, solve all issues, and perform root cause analysis entirely on its own—especially for a platform as nuanced as SailPoint. However, AI can function as a powerful assistant right at our desk, ready to help push investigations along.
A Real-World Example
During our training session, we tackled a recurring issue from our weekly managed services report: termination processes were failing. In a healthcare and compliance context, that’s a big deal. We’d been alerting the customer about these failures but hadn’t offered a clear plan to resolve them or shown that we were actively investigating.
Step 1: Identify the Pattern
We reviewed the five most recent terminations and saw they each failed intermittently, all with a 403 HTTP error message.
Step 2: Refine Through Logs
With that pattern in mind, we dug into the logs. We isolated the source of the 403 error to the connection between SailPoint and Azure.
Step 3: Bring in AI
At this juncture, we asked AI for the most likely causes of a 403 error message in this specific context. AI gave us five potential reasons.
Step 4: Overlay Human Expertise
We applied our knowledge of how SailPoint and Azure interact, and we narrowed it down to a likely timing issue.
Step 5: Communicate with the Client
Next came the human-to-human side. We wanted to reassure the client we were on top of the problem. My team member drafted an email, but I suggested we leverage OpenAI to ensure clarity and completeness. We asked AI to craft a concise summary of our investigation and a polite request for the client to check on their end. Then we copied the right stakeholders and sent the email.
Where AI Fits—and Where It Doesn’t
In this entire 10-step process, AI helped with two important tasks:
Identifying possible causes of the error
Drafting the client email
Will AI be able to substitute for more steps in the future? Quite possibly. But as both an AI expert and a SailPoint product guru, I can confidently say it will require focused AI architecture to replicate the nuanced human expertise, decision-making, and communication that we’re providing. That means there’s still job security—for now!
Key Takeaways
Return to Basics: When performance slips, go back to the fundamental processes and strategies that worked before.
Use AI Strategically: AI can accelerate investigations and improve communication, but it doesn’t replace the need for human insight—especially in complex systems.
Maintain Human Touch: Managed services rely on building trust with clients. Clear, empathetic communication is vital.
Iterate and Refine: Even a well-established service offering can drift off course. Constant oversight and refinement help keep clients satisfied.
By combining human expertise with AI’s capabilities in a structured way, we can continuously improve our managed services—and ensure our clients see real value, not wasted money.