Alden Morris presents LicorChat, an AI platform with 11 specialized agents and 37 tools built for liquor store owners. It connects to your POS, inventory, and customer data, then orchestrates agents in dynamic swarms to generate business plans, profit analyses, and market insights that a single LLM can't match.
Chapters
AI agents for retail liquor stores. So we're always trying to make our chat smarter. So it's not just the basic in and out. How do we differentiate from just really basic providers like ChatGPT, Anthropic, and so on? So what we're doing is not just the power of having your, your POS connected and be able to do that, but then how do we push that information further and start doing just having the LLMs work smarter? Smarter and not harder. So what we've implemented are the agentic flow. So we have, we have agents, we have 11 different agents, we have tools, all the agents have tools. So we have 37 different tools and then the ability of them to orchestrate in addition to doing swarms. So what does that mean? For instance, let's, let's, let's do an example of a just, it's work in progress, but to give you a sneak peek, make me a business plan for my liquor store. Right. So this is a complex question. You put this into just a basic model and it's going to give you just a very generic overview. So overall, it's like if you ask one person versus you ask the whole team. So what we're doing here is we're asking a whole team. And so out of our different agents, we have access to different information. So for instance, the, the within the POS inventory customer, I can do profit analysis and then also can go on the internet and find things like local trends, market trends, XYZ, then comes together to, to overall just synthesize everything. And so this is just a linear progression that we have here. Again, it's highly dynamic. It can go in this instance very linear, but overall when we're doing— we call them roundtables, but when we're doing larger swarms that are allowed to be fully dynamic, they can actually talk to each other and they're going off of confidence. So this is linear, it's going to go in progression. But if we want them to bounce back and forth, for instance, or if we— if one has a question for another, they come together into a synthesis and they're unsure about something, they may throw it back to one of the previous agents with more context. So that's what we're working on. In this case, you would, you would see what's going on. You can see exactly what these are. If you don't like it, you can start over. If you want to skip it and just get a basic answer, you can do that as well. But again, these do take longer. The really cool thing is that we have this visualized. You can see some of them working faster, some of them working longer. We have token estimates just because it goes to how our modeling works. And then you can kind of watch it go. I'm not going to go through the full thing. It's probably got another another minute in here potentially. So just want to give a quick overview into our Agentic flows, our pipelines, and our linear and dynamic orchestration, which is coming to all LicorChat users.
