Account Managers (AMs) at Managed Service Providers (MSPs) need to stay informed about everything happening with their clients—who the key contacts are, common issues, and how the client feels about the service they receive. Gathering and maintaining this information is time-consuming and often requires dedicated effort before customer meetings to compile insights, review common tickets, and gather other relevant data.
While this information is stored in a ticketing system (PSA), scheduled reports typically only cover ticket statistics. Common issues and customer sentiment must be gathered manually. This adds further time pressure, especially if the AM doesn’t fully understand the technical issues being logged or if trends aren’t immediately obvious. They may need to consult with technicians or professional services to gain a clearer understanding of recurring problems.
What does it do?
Using our automation platform, PSA system, and Azure OpenAI, I’ve created a workflow that enables account managers or senior management to select a time frame and a client. Minutes later, they receive an emailed report that provides an overview of the client’s issues, highlights key concerns, and includes AI-generated recommendations based on those issues.
The automation will then review the titles of every ticket for that client, in that date range and generate the following information:
• Key client concerns
• The sentiment of the client
• Common issues reported
• Recommendations for the customer
How does it work?
Approved users of this automation are provided with a simple web form. They only need to select the company and the dates of interest, then click submit. That is all that is required from the end user.
The automation queries the PSA system for tickets corresponding to each of the selected months, storing the ticket subjects in monthly variables such as jan_tickets. The process continues through each selected month until all relevant variables are populated.
These twelve variables are then submitted to a contained instance of OpenAI, hosted securely within our tenancy. This ensures that all conversations with the AI remain internal and protected.
OpenAI has proven to be a great addition to the professional environment and has already been integrated to such an extent that some teams rely on it heavily. The ticket data is passed to the AI along with the instruction:
We’ve found that if you tell the AI the exact hypothetical role it’s playing, the output is significantly better than when you simply give it instructions.
The AI then returns this output to the workflow process in JSON format, which can be manipulated by the workflow. Using this data, an email is constructed and sent to the person who submitted the form.
Conclusion
This workflow is a tool designed for anyone needing a high-level overview of a client’s current landscape. It can be run for previous years to observe changes over time and assess whether the recommendations made in 2024 were considered or remain relevant.
This workflow serves as a guide rather than a formal report to be shared with clients. It provides the account manager with a course of action and recommendations to consider. All information is valuable, as long as it is accurate.