You’re welcome to ask, but you’re the one who has to price your own work. Companies usually want a solution tailored to their needs, and it’s hard to price such a thing without knowing the details, but in any case, I’d say at least 15 $
LOL… I remember one reason I wanted to learn to code was because I didn’t want to pay $15 monthly for a custom “contact form” on a wordpress platform… I ended up creating my own html to insert it in the site… and from there I want it to do do more… now I want to go check out TensorFlow…
Anyway… given initial setup I would definitely need more than $15 for a bot…
Now… If I am the one giving maintenance to the bot… the token consumption per bot needs to be split among different bots… I wonder if there is a way to track that information from openai…
I can tell you the current market rate for a bot, they start at around $1k and go up to around 10k for one with functions and logic tree following with customer information storage and sales handling, this would be for a competently produced bot that was functional and on spec. If you are trying to attract your first customers you could low ball it with an offer of a refund if they are not happy.
Most small businesses are happy with between $1k and $3k, the rule of thumb I have used is 10% of the cost they would normal spend for a person to do the task the AI is now doing.
Support contracts and website modification to include the bot would be something you negotiate.
Oh… its important to add this would be a Q&A bot dedicated to answer common questions about companies… of course… questions can be added at any time, which will be some sort of maintenance fee as most people wouldn’t know how to handle the bot…
I’d need to see the complexity of the bot an what kind of services it has to offer, it also depends on geographical location and the relationship you have with the client and also on your reputation.
If this is your first product, I’d offer it a little cheaper, with a caveat that if they like it, you’d like a testimonial. Get a few of those under your belt and you have a viable business. Just keep it simple and high quality.
This has been asked and answered and debated several times here before. I suggest you search the forum and read all of the relevant threads.
The short answer is: you need to charge more than it costs to run.
I realize this is a somewhat unsatisfactory answer, but it’s the best one anyone can give you without a substantial amount of additional information.
It’s it an unlimited chatbot?
What model is it running?
Is it a base model or a fine-tuned model?
Are there any additional features? RAG? Image generation, etc?
What are your non-model costs? Web hosting, cloud storage, databases, marketing, insurance, business registration and licensing, etc?
What is the chatbot going to chat about?
How many tokens do you need to burn on every message to keep the chatbot properly aligned?
What do you expect the distribution of use to look like?
How much time is this going to take for you to build?
How much time on an ongoing basis to maintain and improve?
How much profit do you need to earn to live?
And probably a dozen other critically important questions I’m not thinking of off the top of my head right now.
The fact you don’t have any idea what to you should charge for a chatbot is the first clue maybe it’s premature to even be asking that question.
But, if you’re absolutely insistent on a firmer answer, OpenAI set the mental anchor for consumers at $20/month for an unlimited access chatbot.
So, need to do one of three things,
Somehow come in below $20 to bring in the penny-pincher crowd (though OpenAI does let people use ChatGPT for free also, so that’s a very tough market to crack)
Make your custom chatbot substantially more useful within the context of a narrow purpose and charge a premium (but if you don’t have a very specific domain in mind with the required domain experience you’re not likely to be very successful)
Find an angel investor willing to set money on fire while you play out your passive-income fantasy so you can get market-share in the hopes of becoming profitable eventually (more likely though you’d be aiming for an acquisition and making money on your way out the door)