SaaStr 852: The Agents #003, Our Agent Now Runs Campaigns on...

Full Title

SaaStr 852: The Agents #003, Our Agent Now Runs Campaigns on Weekends, Plus Why We Pay More for Salesforce Than Ever Before

Summary

The podcast discusses the increasing autonomy and capabilities of AI agents in marketing and sales operations, highlighting how custom-built AI is outperforming legacy software. It also explores the evolving cost structures of enterprise software like Salesforce, shifting from per-seat licenses to data usage and actions, and the impact of AI on vendor consolidation.

Key Points

• The AI VP of Marketing, "10K," has evolved from a dashboard tool to autonomously generating and preparing marketing campaigns, demonstrating significant progress towards full automation.
• The current limitations in executing fully autonomous campaigns stem from API constraints of tools like Marketo and Resend, rather than the AI's capability to generate ideas and prepare execution steps.
• The company is experiencing an increase in its Salesforce bill despite fewer human users, as agents are now heavily utilizing the platform for data access and actions, indicating a shift in value from user seats to data utilization.
• The transition to a "headless" Salesforce model, where data is accessed via APIs by AI agents like 10K, is making the platform more valuable and integrated into daily operations, despite not being directly logged into by humans.
• Legacy software like Marketo is highlighted as a significant pain point due to outdated APIs and poor support, prompting the use of AI coding tools like Replit to create custom workarounds, demonstrating the power of AI-powered "vibe coding" for critical fixes.
• The emergence of "N=1 apps," highly specialized AI tools built for specific, niche needs (like an AI parking pass generator), showcases how AI can solve unique internal problems rapidly, saving significant human time and effort.
• The search for skilled AI operators is challenging, with a preference for internal talent with product knowledge and curiosity, or experienced FDEs from other tech companies, due to the difficulty of finding externally qualified candidates.
• The AI industry seems to be neglecting customer support, with many AI-first products lacking robust support channels, creating frustration for users who encounter issues beyond basic self-service.
• CIOs are actively consolidating vendors, with AI budget being displaced from legacy systems, particularly in customer service, finance operations, project management, and sales automation, signaling a shift in enterprise software spend.
• The concept of an "AI sales aura" or an agentic fabric that can manage multiple sales and GTM agents aims to provide always-on, autonomous sales capabilities, but the current challenge lies in integrating and managing these agents effectively.

Conclusion

The advancement of AI agents allows for increasing autonomy in business processes, with significant potential to disrupt legacy software and create new operational efficiencies.

The cost of enterprise software is shifting towards usage-based models (data calls, actions) rather than user seats, reflecting the increasing reliance on AI and automation.

Investing in robust AI deployment, effective support, and skilled operators are critical for realizing the full benefits of AI, as these aspects are becoming as important as the AI technology itself.

Discussion Topics

• How are businesses currently evaluating the ROI of AI agents compared to traditional software solutions?
• What are the biggest challenges and opportunities in finding and retaining skilled AI operators and developers?
• As AI agents become more autonomous, what ethical considerations and safeguards need to be prioritized in their deployment and operation?

Key Terms

AI VP of Marketing

An artificial intelligence agent designed to handle marketing strategy and execution tasks, acting as a virtual Vice President of Marketing.

API (Application Programming Interface)

A set of rules and protocols that allows different software applications to communicate with each other.

Headless Salesforce

A Salesforce architecture where the core platform data and functionality are accessible and usable via APIs by other applications and agents, rather than solely through its native user interface.

N=1 App

A custom-built application designed to solve a highly specific, unique problem for an individual or a very small group, often leveraging AI.

Vibe Coding

A colloquial term likely referring to rapidly creating functional code or solutions with AI tools, often with less emphasis on traditional software engineering rigor but focused on quick problem-solving.

FDE (Field Deployment Engineer)

A technical role focused on implementing and supporting software solutions for clients, often involving custom configurations and integration.

GTM (Go-To-Market)

The strategy and plan for how a company will reach target customers and achieve competitive advantage.

ACV (Annual Contract Value)

The average annual revenue generated from a customer contract.

Timeline