In my last few posts, I’ve covered how managed service providers (MSPs) can use artificial intelligence (AI) to cut costs, increase margins, and provide new services to customers. In each one, I have touched on how AI is not the be-all, end-all to any problems, however.
Most people are aware of the person who used a generative AI (GenAI) engine to ask how to create a pizza where the cheese didn’t slide off. They were provided with the glorious response of adding glue to the mix (please don’t try this yourself).
Others will have seen images created by AI in which people have six (or more) digits on their hands or limbs at all the wrong angles. GenAI engines are also particularly bad at dealing with the creation of text in images.
There is also a growing trend where some news sites will use GenAI to create articles in order to minimize the number of journalists needed, thereby keeping costs down. Many such articles show up without having been read by an editor, with parts repeated (often several times) and non-sequiturs abounding throughout.
All of these point to one thing: AI cannot be used as a standalone tool within the MSP’s environment. Checks must be put in place. Sure, all these are symptoms of an immature technology. Things will only improve over time, rapidly in the case of AI. However, the last thing MSPs want to see is inadequate AI pushing customers to other providers who have maintained a higher cost but more effective human-based set of services.
Practical uses for AI
Let’s consider the various places where AI can be used and assess how mature such usage is now.
- Data analysis and event management. This is more machine learning than full AI. It represents an ideal starting point for MSPs on their AI journey. It allows for real-time data analysis and the initiation of actions to address any identified issues. This elevates automation capabilities to the next level. This area is pretty mature in the market. Many vendors offer orchestration and platform management tools that can be pretty much left to their own. It is still good practice, however, to ensure that regular reports are provided to humans. This is so that they can keep an eye on trends. They can also ensure that high-impact events are flagged in real time, allowing individuals with better insights into the bigger picture to take action.
- Prospect/customer advisory services. Here, AI can use natural language processing to understand what a prospect or customer is looking for when they visit an MSP’s site or contact the first-level helpdesk. Advanced rules-based engines (not really AI, but most vendors selling such capabilities have moved to call them AI) can then direct them through to best-effort responses or to a human if the engine decides that it would be a better route to go. Here, ensuring that the engine does not go down rabbit holes, with the prospect/customer giving up the will to live and giving up on their session, is key. We need to place intelligent rules into the system, such as timeouts. If a prospect/customer has provided a stated number of responses, the system should automatically route them to a human to maintain the contact.
- Service maintenance. Using AI to analyze a customer’s usage of one or more services is highly beneficial. This analysis enables the provision of tailored feedback to customers. It can suggest potential cost savings and recommend additional services for a more comprehensive solution. However, customers do not want confusing advice. Account managers must verify any AI-generated content before sending it to ensure clarity and relevance.
- Detailed reporting. AI can generate more complex reports that can target the recipient more effectively. GenAI can create technical reports for system admins within customer environments. It can also provide insights into how the MSP’s services are delivering value at a business level to the line of business managers and C-level staff. It’s important to keep in mind that this can be a high-stakes game. Account managers must review all output before sending it to the customers.
AI’s staying power
Overall, it is apparent that AI in its many forms is here to stay and that MSPs must get to grips with it. This includes what AI is suitable for now and what it is likely to be suitable for in the future. At this stage, it makes the most sense to err on the side of caution. Use account managers to thoroughly check and amend AI output before it goes beyond the company walls.
Disclosure (or non-disclosure, really): None of the above has used AI in any shape or form. Currently, the author still prefers to use his own experience and capabilities to craft articles. Therefore, readers can only blame the author himself for any issues they identify.
Photo: Rawpixel.com / Shutterstock
