Are you taking advantage of the great opportunity that Artificial Intelligence (AI) offers to your IT Solution Provider or Managed Service Provider (MSP) business?
AI is being used in all areas of life now, from a personal assistant who schedules meetings for you to an AI friend that looks after your money. AI isn’t tomorrow’s technology. It’s here, today.
I recently wrote about three ways MSPs can use AI to increase their profits, lowering their cost of support by using machine learning. I received a lot of feedback on that article from MSPs asking for ideas on how to implement AI within their MSP business. So, where should you start? Here are my 7 rules for using AI effectively within your MSP business:
Create a decision framework
The first step towards implementing AI within your MSP is to create a decision framework. Put simply, don’t try to automate anything until you understand the decision process first.
For instance, your Professional Service Automation (PSA) tool should show you your top 5 ticket types. What steps do you typically take to resolve these tickets? What kinds of decisions do you typically make? By finding the answers to these questions, you can establish areas where you can easily streamline.
When choosing what to automate, ask yourself the question “How often does it happen vs how much does it cost to resolve?” Assign a cost to these decision frameworks before you begin. How long does it take an engineer to resolve that ticket right now? Create the baseline. This will show you how much time and money you’ve saved when after the tickets are automated.
Be proactive
When looking at using AI and automation within your MSP business, look beyond the basic tasks such as service restarts and disk space resets. What cumbersome tasks could be automated — perhaps using AI — to free up your time? Think sideways here.
Thinking sideways about issues means you can take a high priority, urgent ticket and turn it into a low priority ticket you can deal with when you have more time.
I saw a great example of this with an MSP I worked with recently. They found (using step 1 above) that many of their support requests were about printers not working. These support requests took time to resolve — time that both the engineer and the end-user didn’t want to waste.
When the MSP now receives a printer issue ticket, they use automation to set the default printer for the user to another printer nearby. This gets the user up and running again almost immediately. The non-functioning printer is then flagged as a lower priority ticket for the MSP to investigate at their leisure.
Look for patterns
What patterns can you spot where automation might help? If a server fails regularly, what are the underlying conditions? How can AI be used to spot those patterns and take appropriate action, when those conditions are almost met?
For instance, you spot that a Line of Business (LoB) service often saps too many resources before it crashes and takes the server down with it. Could you proactively restart that service when it sees CPU hit 90 percent? Again, this takes an urgent issue (the server crashing) and turns it into a low priority issue because you can take your time to resolve the LoB issue.
Look to automate across multiple client sites
One great way to reduce your costs of support by using AI, is to look for situations and scenarios that are true across multiple clients. You can then research an automation strategy that can be used to help streamline potential issues. For example, this could be useful when multi-tenant services such as Office 365 go offline. What steps could you automate to help the multiple clients who are affected?
Automation can save you significant time communicating with multiple clients in this scenario, reducing service desk tickets and triage resources.
Look for event triggers
Look for opportunities to automate specific triggers. For instance, one trigger might be when a new user account is created.
- Which groups can a new user be added to
- Which software needs to be installed for the new user
Likewise, when a user account is changed — which software can be uninstalled?
The Internet tool, IFTTT – If This Then That, can provide you with some great examples of automation for event triggers. Go and try it!
Continuously review and optimize
Despite the name, automation isn’t a set and forget process. As a business, you will need to review, maintain, and tweak your automation and AI processes often. To help accomplish this, put a process in place to for regularly review any automation you create. Otherwise, you may well end up with a scenario where the AI goes crazy, Terminator-style!
Photo: John Williams / Shutterstock
