Maximizing Efficiency in VC: Streamlining Due Diligence with ChatGPT

Grunt Work is Not a Good Use of Your Time

I’ve written about how chatGPT can be used for better decision making. But it is also great for the tedious, time consuming parts of VC too. I’d argue that by spending less time on grunt work it leaves time for better decisions too.

Hiring a Diligence Consultant in 30 Minutes

For my hypothetical investment into Mistral, I’ll need to hire a consultant to complete the technical and market diligence. This means I need a job specification and need to find the appropriate consultant. You’ve probably heard of using chatGPT to write job descriptions before but I want to show you how you can use the diligence scope to make this even faster.

Crafting a Job Specification Using the Diligence Scope

Returning to the diligence scope, I tell chatGPT that I need to hire someone to perform the technical and market diligence, outlined in the scope. I simply ask chatGPT to write a description of the diligence and the skills needed to complete it. The results are ok, would definitely save me quite a bit of time, but are generic. My diligence scope is less specific than I would normally write, as I’m just using Mistral as an example. This made the job specification too generic. This shows the importance of tweaking and refining your chatGPT results as you go, so that all subsequent steps are more personalised and useful.

chatGPT's first attempt at a job description, held back by an insufficiently detailed diligence scope

Prime Your AI Assistant

To fix the generic job description, I asked chatGPT what it knows about benchmarks and competitors in the area, to “prime” it. Then asked for a more detailed specification covering the types of benchmarks and types of analysis. I have not provided more detail here but it always improves it if you can provide specific examples. I think the rate of improvement of Mistral’s models and their competitors is likely important so it should include that too. I note the market section is too generic and it needs to be more AI focused. This is much better but needs some final tweaking with a more detailed description of the expertise I’m looking for. I can iterate back and forth on this description to get it just right, add in other items I think of myself. Then, a real time saver, I ask chatGPT to make a list of Linkedin Sales Navigator search terms to find the right candidate.

A more detailed job specification after I asked chatGPT to be more specific and gave it examples

LinkedIn Sales Navigator search terms, developed by chatGPT using the final job description

Don’t Waste Time and Energy - It’s Better Spent Elsewhere

So, what were the benefits of this approach? Obviously, I saved time. By using the diligence scope I had already made, I could generate a passable job description. But time was saved too by being able to feed in what I refer to as “stream of consciousness” thoughts into chatGPT - I could say I want it cover benchmarks, MIT background, and look at proprietary models. Then chatGPT would turn these into sentences and put them into the correct section of the job description. Easy stuff to do yourself, but why bother? By transferring tedious work to chatGPT, it leaves so much more time and energy to write clearly on important things and think carefully through decisions.

#VentureCapital #ChatGPT #VCEfficiency #DueDiligence #AIinVC #TechInvestment #StartupDiligence #InnovationInVC #AIForVC