Five Suggestions for Building a Generative AI Strategy for your Tech Startup's Operations

Resource: Using an AI generative LLM chat tool as a force multiplier

Sometimes I play a little game. When I get into any type of business or technology discussion, I keep track of how long it takes to get to the topic of Artificial Intelligence (AI). It usually doesn’t take long, and that is because this quickly evolving technology is disrupting everything. And the startup world is no exception. According to Gemini’s summary, 50% of all global venture funding went to AI-focused startups. Which is great if your startup is AI-focused, but kind of sucks if not.

But have no fear, AI can still be useful because you can leverage the technology to meet your goals faster and with less cost. Not necessarily to replace team members, but to make your core team more effective. We tend to think of AI for big things and transformative changes to your company, but I’ve seen the biggest benefits in day-to-day operations.

Ask anyone who has used AI, which is pretty much everyone these days, how to get your startup’s goals faster, and they will give you different answers. Often, they will dig into examples and stories of how it produced jaw-dropping results or how it was wrong. But that is not super helpful. So, I’m not going to do that.

Also of note, there are lots of specific tools that use AI to help make companies and their employees more effective. I’m focusing on chat interfaces like ChatGPT, Claude, Gemini, and MS Copilot.

To write this article, made a list of about a dozen different ways I’ve seen startup companies leverage this flavor of AI for things like HR, IT, marketing, sales, quality systems, documentation, and that sort of day-to-day operational thing, and then picked the five that have little to no upfront cost and can deliver big benefits:

1: Research with greater depth and speed

If you are still using search to do research, you are doing it wrong. Ask your LLM chat buddy to tell you about something, and keep asking. I like to think about it as if I’m on a flight sitting next to the world’s expert on a topic I need to know about, and I can keep asking them questions. It could be as simple as, “How does the R&D tax credit work in the US?” to as technically in-depth as, “Is there a way to approximate a quantum electromagnetic excitation as a classical electromagnetic force?”

Most LLM tools have “research modes” where they can dig deeper and give you more references. You can also guide it based on what you need. Type in some background information or provide a file with detailed constraints you want for the research. One great example: if you see a company as a prospect for your product, don’t just ask your chat buddy to tell you about it. Have a file that describes your product, its applications, and its differentiators, and then say, “Using the information about our product in the file, tell me if Acme Screwdriver would be a good prospect and if so, why.”

Ask, ask, ask. And then keep asking.

2: Target marketing messaging

If you did your homework early on in your startup, you should have customer personas. LLMs are great at taking an existing lengthy message and rewriting it to apply to one of those personas. Let’s say you sell to schools; the message you give teachers is different from what you send to district administrators.