Want to Profit from the AI Boom? Prompt Engineering is the Place to Start

Can AI consulting help generate MRR and boost your bottom line? It can if you study AI, apply it in-house, then craft your sales pitch in the right way. The perfect place to start is honing your approach to communicating with LLMs (large language models) like ChatGPT.

“Prompt engineering” is the process of providing thoughtfully constructed input text that produces the most useful output possible. Think of it like asking a non-technical customer a question about their computer problem. If your question is vague, their answer will be, too. When a prompt engineer refines their inputs to match the “learning style” of the LLM, they get faster, more accurate information.

But is there a market for MSPs to consult in this space? Experts say yes, it’s one of many AI services that MSPs are successfully bundling and selling. Research from PwC predicts that AI will contribute $15.7 trillion to the global economy by 2030, with businesses adopting AI seeing productivity gains of up to 40%. Key to that being able to communicate with and problem solve collaboratively with AI tools.

“Prompt engineering might be the fastest growing profession we’ve seen,” noted Lars Nyman,the CMO of of CUDO Compute, a London-based company that provides infrastructure for AI projects. 

Practical Advice for AI Consulting

“Every role in an MSP can speed their job by asking ChatGPT to help,” said Ronnie Parisella, project manager and consultant at DataBit. Once you’ve improved your own processes, he adds, you’ll know how to use AI to help your customers. You have to dive in and use the technology daily, Parisella said. Right now, “people use about 1% of an LLM’s capabilities.” 

Go for the low-hanging fruit, advised Parisella, starting with the text content you produce on a daily basis. “Everything text based can be improved. Look at marketing plans, blog posts, websites, templates for client proposals, everything,” he added. Even accounting for proofreading and editing time, an MSP can save 90% of the time it takes to draft written copy according to Parisella. “Then, take what you’ve learned to your customers.” 

Corey Kirkendoll, president and CEO of 5K Technical Services,  advised MSPs start using prompt engineering for specific use cases. “Give AI a single directive – here’s what I need – to get the output as close as possible to what you’re trying to solve,” he said. One key to that is adopting and assigning personas. A good prompt might be, “You’re a social media expert, and I’m looking to do X.” 

Imagine your generative AI model is a new employee. “Give them step-by-step instructions. Describe the workflow you want and break it down into smaller parts. It may take multiple prompts to make it clear what you want,” Kirkendoll said.

When people see tangible benefits to using an LLM, the door to AI consulting swings wide open. Kirkendoll walks his customers through practical examples of how he’s leveraging digital labor. That leads to a brainstorming session where the customer identifies problems that AI might be able to solve. 

Kirkendoll also plans to offer customers with AI applications that address common use cases. “Salespeople hate to update their CRM,” he added. “Instead, AI can be used to summarize every phone call and insert those summaries into that customer’s CRM records automatically. It gives the rep back at least half that time.”