Two sales calls with potential enterprise clients, the good and the bad

I'm in the process of launching a B2B SaaS product, called Prediction Guard, that makes it easy to integrate and keep up with state-of-the-art AI models. I have an MVP (almost) ready and I've been validating the idea with a good number of people in the target market. We have a landing page and we are gathering a waitlist.

Over the past two weeks, I "bumped into" two potential enterprise customers via Slack and email. One of these was in the healthcare space and one in the translation industry, and both had good use cases for the product. First question, should I even have pursued these enterprise contracts at this point in the process? Well, let me know your opinion.

In any event, I scheduled calls with both potential customers to understand their use cases a bit more and ultimately move towards an annual contract for use of Prediction Guard (and maybe some enterprise support level). Here's what happened:

Call #1 (the healthcare use case): This call started off great, and the discussion about their pain points was good (and at least partially validated some of my product features). They then proceeded to describe some of the constraints of their work and their desire to "train our own AI models." Our system isn't really geared towards training models from scratch, and their use case wouldn't require this anyway (in my opinion).

I decided to see where the conversation went, and I described (verbally) the Prediction Guard product with a focus on what prompt engineered and fine-tuned AI models could do (especially powered by our domain specific model selection process). However, it was clear that they went into the call wanting to find a system that would train models from scratch using their data, and they weren't convinced that "general purpose" AI models could solve their very domain specific challenges.

Although I have advised and consulted on a few projects in the healthcare space, I don't primary work in that industry. I think that showed in how I was talking about their data and in how I was asking questions about their end goals. Overall, the call was a flop, although I learned a few things (see below).

Call #2 (the translation use case): This one was right in my wheelhouse. I'm been working in Natural Language Processing (NLP) and specifically in the translation industry for the past 5 years. Moreover, I know a lot about what organizations in this space care about and what they value. They want efficiency, but not at the cost of quality. They are also nervous about the risks of machine translation (MT), because comparing MT models is difficult and reliability is a problem.

In preparation for this call (after my failed first call), I decided to tailor a demo of the system just for them. It was a long night, but I managed to throw together a demo that clearly showed the value of the system for translation service providers in a simple UI interface. The demo showed how they could quickly compare many different MT systems and de-risk the machine production of first drafts.

I started the demo, explained how the system works. They asked a few questions along the way, but I saw the light bulb go off in their head. I didn't even get through my demo before they were talking to each other about how they were going to use the system. Really excited to explore this! I'm working on a proposal/ contract, and will turn to that after writing this post.

What I learned:

• Demos are a cheat code. The one I pulled together was super simple, but it was in the domain of the potential customer and immediately connected the dots for them.
• Find early enterprise customers in domains where you have familiarity. I knew exactly how to pitch in the language technology space, but healthcare was more of a challenge. Actually, I think there are some really good uses of the product in that industry, but it was definitely a harder sell at this stage. Maybe a domain specific demo would have helped?
• There is still a wide range of understanding in terms of how "AI" can or should be implemented in businesses. That's going to be a challenge moving forward, but I guess 1/2 of people trying to integrate AI is still a lot of customers. I need to get my docs online with specific tutorials/examples from a variety of industries. Any suggestions here?

What else should I have learned from these calls? What comes to your mind? I'm sure many of you here might have avoided some of the mines I stepped on, and excited to learn from you all!