Building generative AI chatbots

Thank you for tuning in to this audio only podcast presentation. This is week 141 of The Lindahl Letter publication. A new edition arrives every Friday. This week the topic under consideration for The Lindahl Letter is, “Building generative AI chatbots.”

You can feel the winds of change blowing and the potential of people building out election expert opinion chatbots. Maybe you want to know what they are probably going to use to underpin that sort of effort. If you were going out to build some generative AI chatbots for you might very well use one of the 5 systems we are going to dig into today.

  • Voiceflow – This system may very well be the most prominent of the quick to market AI agent building platforms [1]. I have chatbots deployed to both Civic Honors and my main weblog powered by Voiceflow.
  • LangFlow – You are going to need to join the waitlist for this one to get going [2]. I’m still on the waitlist for this one… 
  • Botpress – Like Voiceflow this system lets you pretty quickly jump into the building process of actual chatbot workflows [3]. To be fair with this one I was not able to build and deploy something into production within minutes, but you could do it pretty darn quickly if you had a sense of what you were trying to accomplish. I built something on Botpress and it was pretty easy to use. After login I clicked answer questions from websites to create a bot. I added both Civic Honors and my main Nels Lindahl domain. They just jumped in and advised me that the knowledge upload was complete. Publishing the bot is not as low friction as the Voiceflow embedding launch point, but it was not super hard to work with after you find the share button.
  • FloWiseAI – You will find this is the first system on the list that will require you to get out of your web browser, stretch a bit, and open the command line to get this one installed with a rather simple “npm install -g flowise” command [4]. I watched some YouTube videos on how to install this one and it almost got me to flip over into Ubuntu Studio. Instead of switching operating systems I elected to just follow the regular Windows installation steps.
  • Stack AI – With this one you are right back into the browser and you are going to see a lot of options to start building new projects [5].

All of these chatbots built using a variety of generative AI models are generally working within the same theory of building. The conversation is being crafted with a user and some type of exchange with a knowledge base. For the most part the underlying LLM is being used to facilitate the conversational part of the equation while some type of knowledge base is being used to gate, control, and drive the conversation based on something deeper than what the LLM would output alone. It’s an interesting building technique and one that would not have been possible just a couple of years ago, but the times have changed and here we are in this brave new world where people can build, deploy, and be running a generative AI chatbot in a few minutes. It requires some planning about what is being built, you need some type of knowledgebase, and the willingness to learn the building parameters. None of that is a very high bar to pass. This is a low friction and somewhat high reward space for creating conversational interactions. 

Messing around with all these different chatbot development systems made me think a little bit more about how LangChain is being used and what the underlying technology is ultimately capable of facilitating [6]. To that end I signed up for the LangSmith beta they are building [7]. Sadly enough “LangSmith is still in closed beta” so I’m waiting on access to that one as well. 

During the course of this last week I have been learning more and more about how to build and deploy chatbots that take advantage of LLMs and other generative AI technologies. I’m pretty sure that the development of agency to machine learn models is going to strap rocket boosters to the next stage of technological deployment. Maybe you are thinking that is hyperbole… don’t worry or panic, but you are very soon going to be able to ask these agents to do something and they will be able to execute more and more complex actions. That is the essence of agency within the deployment of these chatbots. It’s a very big deal in terms of people doing basic task automation and it may very well introduce a distinct change to how business is conducted by radically increasing productivity. 









What’s next for The Lindahl Letter? 

  • Week 142: Learning LangChain
  • Week 143: Social media analysis
  • Week 144: Knowledge graphs vs. vector databases
  • Week 145: Delphi method & door to door canvasing
  • Week 146: Election simulations

If you enjoyed this content, then please take a moment and share it with a friend. If you are new to The Lindahl Letter, then please consider subscribing. New editions arrive every Friday. Thank you and enjoy the week ahead.

Leave a Reply