Great idea – if you’re leaving Xitter, start a blog! You don’t get the immediate feedback, but it’s *your* platform and you can control what gets said and done there.

***

I have a friend who *really* wants to jump on the LLM / generative AI bandwagon with his tech services business. Trouble is he hasn’t put in the work to understand what these tools are really capable of (and more importantly, what they’re incapable of). His ideas are all over the place. On one hand I want to help him, but on the other…it could turn into a big project. I’m thinking about it. At this stage of life, everything is a tradeoff with the limited time one has.

Thinking out loud, some of the things that current-gen AI models can do for a tech consulting firm are:

• Writing proposals. Proposal writing is tough and is an unbillable expense. Most people aren’t good at it. With someone in charge who knows the right queries to use, proposals could be done in days instead of weeks, using 1-2 people instead of 5-6. And every successful proposal is righteous input for the LLM (large language model), improving it fast.
• Writing documentation. Technical manuals, reports, how-to essays…so much writing. This could be a mix of billable and unbillable work.
• Writing code, like database queries, Excel and PowerBI formulae, HTML code, etc. A single analyst can do the work of 3-4 using generative AI as a force multiplier.
• Automation of routine data entry. Not sure if this is an RPA (robotics process automation) or LLM type problem, but one way or the other it’s not hard to automate.
• Natural language exploration of large datasets. Previously you had to combine a business expert with a data analyst to explore what’s really going on in a large dataset produced by a business process. Any good LLM can be that data analyst and answer the business expert’s questions quickly and authoritatively.
• Rapid onboarding. Tech firms have to invest time in training their new hires, and an LLM can be a useful Q&A resource for newbies. How do we do this? Where can I find that? Etc.
• Contract review. Contracts can be large, complicated and hard to understand. An LLM could be a good first reviewer (before paying a lawyer) for the business layman trying to understand all the terms and conditions.
• Customer service or employee service chatbot creation.
• Project management support. Generating a detailed project plan from just a few inputs. Analyzing a project plan and identifying the missing details. Generating PM reports.
• Teach clients how to do all the above – an applied AI practice, if you will.

That’s not a complete list, but not bad for just thinking out loud.