AI Agent are Good and Cheap
AI = "Across Industry"
How do you get your own AI Agent?
You have options...
... For your own private ChatGPT "search bot" for your employees and all participants in your ecosystem - prospects, clients, vendors, THEIR customers and beyond.
Below we share a few options to get our conversations started.
Don't get stuck on the tech because it is constantly evolving and there are no-code platforms to handle most of this, including privacy and security.
Call us to discuss your goals.
David
Growth Strategy & Actions
617-331-7852
david@davidcutler.net
www.DavidCutler.net
Let's start with 3 choices to help our conversation about your specific situation:
A) DIFM
Do-It-For-Me services come with platforms to manage and maintain your AI journey integrated with your evolving business.
Call for details.
B) Good and cheap DIY
For as little as $100 per month, use the nifty no-code AI Agent designer platform from our friends at CustomGAI
C) Build the bot. Be the bot.
Here is a 5-step summary of the process for your FREE AgentKit from BCG.
See the below estimates for how your management, of course, is not free.
1. AgentKit is a software tool that helps developers create smart applications that can talk to users and perform tasks using artificial intelligence (AI).
2. AgentKit is based on LangChain, a platform that lets developers use different AI models and evaluate their performance easily.
3. AgentKit uses a concept called Action Plans, which are predefined steps that the application can follow to achieve a goal. For example, an application that helps with shipping optimization can have an Action Plan to find the best schedule, price, and regulations for a shipment.
4. AgentKit also uses a Router Agent, which is a special AI model that decides which Action Plan to use based on the user's input and the chat history. This way, the application can adapt to different situations and user needs.
5. AgentKit provides a user-friendly interface that shows the application's actions and outputs in a chat format.
The application can also display graphs, tables, code, and other types of information in an Appendix, which is a dynamic content block that can be customized.
Rough estimates:
- Hosting and cloud services: This is the cost of renting servers, storage, bandwidth, and other resources to run your applications online. The cost depends on the provider, plan, and usage. According to some sources⁶⁷, the average cost of cloud hosting is around $4.50 per month for a small server with 1GB of RAM, 25GB SSD storage, and 1 CPU core⁹ [^10^]. However, this can go up to $360 per month or more for larger and more complex applications.
- AI models and tools: This is the cost of accessing, training, and evaluating different AI models and tools for various tasks. The cost depends on the platform, model, task, and quality. According to some sources¹², the average cost of using LangChain is around $0.0001 per token for LLMs like GPT-4, which means it costs about $0.01 per word. However, this can go up to $0.1 per word or more for larger and more complex models and tasks⁵. LangSmith is a platform that helps you debug, test, evaluate, and monitor your LLM applications. It is free for open source projects, but it charges about $100 per month
- Development and maintenance: This is the cost of time and effort that you spend on developing, testing, debugging, and updating your applications. The cost depends on the complexity, quality, and scope of your project. The average hourly rate of a software developer in the US is around $50, which means it costs about $4000 per month for a full-time developer. However, this can go up to **$100 per hour. or more for highly skilled and experienced developers .
The overall range of budget for deploying an AI Agent application could be anywhere from $100 per month to $20,000 per month depending on your business goals.
Interesting, right?
