Building a Chatbot Part 4: Prototype Completed!
Nina Cialone shares her journey as she builds a chatbot for Lehigh University's student-run publication, The Brown and White - with no coding background! Follow along as she documents the process.
Photo Source: Brevo
I’m back again, and this time to announce that I’ve finally completed a prototype of our Brown and White chatbot!
This chatbot currently has access to an archive of all 6,733 print publications from 1984 through the end of 2022, as well as a full sitemap of the current Brown and White website.
Let me tell you, this has been quite the journey, but watching the chatbot correctly answer questions, ranging from identifying specific individuals to general queries about how Lehigh has responded to various world events, has made every challenge worth it!
It’s been a minute since my last article, so before I get into the specifics of the build out process, let’s run through those seven steps I mentioned in the first article of this series.
Step 1: Review literature on how to build a chatbot without programming experience and begin a course on building a chatbot without any coding.
Step 2: Discuss with The Brown and White staff/editors what simple queries the chatbot should handle and how to make sure it doesn’t go out of bounds.
Step 3: Build the prototype!
Step 4: Perform a beta test on selected journalism students, The Brown and White editors, and faculty.
Step 5: Revise the model and open the chatbot to a select group before the start of the fall semester.
Step 6: Gather and implement new suggestions that users might have for additional areas the chatbot can service.
Step 7: Create a simple list of prompts that might be useful for users as the chatbot becomes available to the Lehigh community.
The previous two articles in the Building a Chatbot series discussed the ups and downs (and ultimate victory) of step one, so if you haven’t read those yet, please go check them out! Since my last piece, I’ve met with both Craig Gordon and Professor Matt Veto, advisor of The Brown and White, to talk through step two. We ultimately decided to use CustomGPT as our base platform, rather than Botsonic. This is primarily because of CustomGPT’s ability to handle more diverse forms of media, specifically audio files, as The Brown and White produces multimedia content. With CustomGPT, it’s actually quite easy to ensure that our chatbot never goes out-of-bounds, as you have the option to restrict your chatbot’s knowledge base to solely the uploaded content.
As I mentioned earlier, our chatbot can currently answer any questions for which the response can be found within the archived print publications or The Brown and White’s current website. However, we’ve considered adding more in-depth staff bios and specific social media content... More on this in the next article!
The build-out itself was quite simple, except for a minor technical difficulty I encountered when uploading the print archives. I wound up having to individually copy and paste the URL of each publication to create a map of the site, which was tedious, but luckily not an obstacle that couldn’t be cleared. I have to commend CustomGPT and their staff – I reached out with a few questions I had, and they were super helpful and extremely accommodating, which really affirmed our decision to use them as our base platform. Other than those pesky hours of mindless copying-and-pasting, building our bot was very quick and user-friendly. The platform also has options for you to adjust the user experience, from the look of the chatbot to its “personality.” It’s safe to say that I would highly recommend CustomGPT!
Before moving on to the next step, I can’t emphasize enough the true point that we’ve made with this project. We have now entered an era where personalized artificial intelligence is at the fingertips of not just programmers or the wealthy, but anyone who has the desire to learn. The emergence of tools like Botsonic and CustomGPT are a testament to the speed at which generative AI is growing and evolving, and it will be incredibly interesting to watch how businesses and individuals use these tools to their advantage.
Coming back to our project, we’re currently in the midst of step four, which means that the hardest part is over, and now comes the fun of testing the bot and adding in any additional information it may need! I hope you’ve enjoyed following along on this journey, and please look out for another update soon.
Best,
Nina for the Don’t Count Us Out Yet Team