SaaStr 854: The Agents #005, Our AI is Hiring! Would You Work...

Full Title

SaaStr 854: The Agents #005, Our AI is Hiring! Would You Work for One? And Are Autonomous Agents ... Safe?

Summary

The hosts discuss the surprisingly low cost of running AI agents ($257/month for two) and how this impacts business decisions.

They explore the evolution of AI agents from simple tools to complex assistants and the growing trend of "coding in English" through AI.

Key Points

• The cost to run two AI VPs (marketing and customer success) for a month was only $257, which shocked the hosts and was significantly less than anticipated.
• This low cost is attributed to using cost-effective models like OpenAI Mini for most tasks, nominal API call costs for data retrieval, and free or low-cost data storage.
• The hosts emphasize that the cost of AI agents is not the primary constraint; soft costs like founders' time and the strategic value of not building in-house are more significant considerations.
• They discuss the debate around whether AI agents are truly replacing VPs of marketing, concluding they can replace junior marketing roles and handle a "sliver" of the VP's responsibilities, with the potential for this sliver to grow.
• The discussion highlights how AI agents are becoming integrated into workflows, with SastraAnnual.com evolving from a website to a live agent that handles tasks like email marketing and sponsor communication.
• The conversation touches on the potential for agents to interact with each other and the need for an orchestration layer, as well as the growing importance of platforms like Replit that provide robust development environments for AI agents.
• The hosts also explore the continued value of Salesforce as a backend for AI agents due to its extensibility, existing infrastructure, and the ability to manage hybrid human-AI teams, despite the cost.
• They note the increasing capability of AI agents to perform tasks like creating newsletters and handling customer queries, displacing existing point solutions and reducing operational "drama."

Conclusion

The cost of AI agents is surprisingly low, making them accessible for businesses of all sizes.

Businesses should focus on the strategic value and soft costs of AI agents rather than just the monetary expenditure.

AI agents are rapidly evolving and integrating into core business functions, potentially displacing niche SaaS tools and becoming essential components of the tech stack.

Discussion Topics

• How will the surprisingly low cost of AI agents impact startup strategies and the broader tech landscape?
• As AI agents become more sophisticated, what are the ethical considerations around their autonomy and decision-making?
• With AI agents taking over more complex tasks, how should companies rethink their human workforce and the definition of a "job"?

Key Terms

AI VP

Artificial Intelligence Vice President, a role or function performed by an AI agent that mirrors the responsibilities of a human Vice President.

OpenAI Mini

A specific, likely lower-cost version of OpenAI's language models, used for less complex AI tasks.

Claude Opus

A powerful, high-end language model from Anthropic, often used for more complex or nuanced AI tasks.

GTM Agents

Go-To-Market Agents, AI agents designed to support sales, marketing, and customer success functions.

Replit

A cloud-based integrated development environment (IDE) that allows for collaborative coding and deployment of applications, often used for building AI agents.

SaaS

Software as a Service, a software distribution model where a third-party provider hosts applications and makes them available to customers over the Internet.

Timeline