How can MSPs leverage AI internally? Is anyone doing it now? ChannelE2E recently connected with one such MSP that leveraged the minds of data scientists to create a tool to help with its service desk. Transputec is a London-based MSP that provides support services to end customers around the world. Transputec also has an interesting connection into the world of AI startups. Founder and owner of this MSP, Sonny Sehgal, was an early investor in an AI pioneer called Vastmindz. Vastmindz is an AI unicorn, best known for its remote health and monitoring technology that has won financial backing from Microsoft and also from Apple co-founder Steve Wozniak. It’s also the company founded by Sehgal’s son, Nikhil Sehgal.It takes two to three days to train Neo on an MSP's data.Since the introduction of NeoAgent, the company has also introduced a customer-facing AI tool called Neo Pulse which provides direct assistance to end-user customers via chat on common queries.
How Can AI Help MSPs?
Sonny Sehgal asked the data scientists behind Vastmindz to create an AI tool to help with the tasks that MSPs work on everyday -- automating the resolution of help desk tickets. That’s how the new tool NeoAgent was created. Nikhil Sehgal developed Neo for MSPs, and Transputec was the first customer.The tool is delivered via a Microsoft Teams bot. Service desk workers type their queries into the bot and can get the answers to customer questions much more quickly. The bot also has the capability to normalize language used with customers, enabling technicians who have a language barrier with customers to be supplied with the most appropriate language to use in a chat conversation. Transputec is now using this tool to help its service desk team of 65 people globally resolve tickets faster, and the tool is now being offered to other MSPs as well. So far, MSPs report being able to reduce the time to resolve tickets by up to 50% and increase customer satisfaction by up to 20%.How Did Data Scientists Create Neo?
We all know that AI and machine learning needs data sets to be trained. So how did that work with this Neo tool? What data sets were used?NeoAgent is trained with the following data sets:- Internal MSP documentation such as the onboarding process for a customer, the Wifi connection process, the process for a terminated employee, etc.
- Data collected inside the MSP platform, such as ConnectWise, IT Glue, or ServiceNow. The tool uses an API to access the service desk tools and documentation tools that an MSP already uses, including past ticket data, and then uses the MSP’s internal data to train the algorithm.
- An OpenAI large language model. Sonny Sehgal noted that Transputec is a Microsoft shop, uses Azure, and now uses OpenAI.




