Navigating the Future of AI: Insights from Sam Altman's Dev Day and Pay Per Result Models

The world of artificial intelligence (AI) is changing rapidly. Events like Sam Altman's Dev Day during SF Tech Week play a vital role in bringing together enthusiasts and experts to discuss what's new. This year, the spotlight was on exciting updates about AI agents and how ChatGPT actions are transforming user experiences.

In this post, we explore what these changes mean for the future of AI, especially the shift toward "pay per result" business models that could change how AI services are priced and valued.

Innovative Features from Dev Day

The Dev Day event showcased the latest advancements in AI agents, highlighting how developers can maximize the use of these tools to build more interactive applications. Sam Altman emphasized the need for customization and control through the OpenAI API. This flexibility allows developers to configure AI features to suit their unique needs, improving the overall user experience. For example, a customer service app using AI can be tailored to provide instant responses that enhance customer satisfaction.

One standout feature discussed was the ability to turn custom GPTs into proactive agents that can carry out various tasks. This means instead of just answering questions, AI can process bookings, send reminders, and engage with users in real time. Imagine an e-commerce app where AI not only chats with customers but also recommends products based on their browsing history and completes purchases seamlessly.

Choosing the Right Tools for Development

When implementing these features, developers face an important decision: use the OpenAI API directly or opt for platforms like Make.com. The OpenAI API offers more customization options, making it perfect for developers looking to build complex systems.

In contrast, Make.com simplifies the task for those with less coding experience. This visual workflow builder allows users to create functional applications without needing to write extensive code.

The Pay Per Result Model

Many data scraping providers and marketplaces available within AI agent building platforms like Make.com offer "pay-per-result" models, where you are charged only for the successful data points or "results" extracted. This approach is ideal for projects with uncertain data volumes or for cost-conscious users, as you only pay for what you receive.

Recent reports indicate it is still a small, emerging model, though it is gaining traction in specific, measurable use cases. The most common models for AI companies remain seat-based, usage-based, or a hybrid of both.

Reasons for the limited adoption

• Difficulty defining success

  For "pay per result" to work, both the vendor and the customer must agree on what constitutes a successful, measurable outcome. This can be complex for general-purpose AI tools where results are hard to quantify.

  
  

• Challenges measuring value

  An analysis of 40 AI companies found that most rely on traditional SaaS pricing because they have a difficult time quantifying the value of their solution. As a result, usage-based pricing remains rare and pure outcome-based models are even rarer.

  
  

• Lack of billing infrastructure

  Many companies' existing billing systems are not equipped to handle the complexity of outcome-based pricing, which can lead to operational headaches and lost revenue.

Model Adoption

To stay competitive in this evolving landscape, businesses must investigate different models to ensure profitability as they develop their strategies.

The "pay per result" model is most successful for AI applications with clear, quantifiable outcomes. Examples include:

• Customer support bots. Companies like Intercom and Zendesk charge per successfully resolved customer support interaction, where success is defined as the AI agent resolving the issue without human intervention.

• Sales agents. Zoom, after acquiring Exceed.ai, used a performance-based model where clients paid per qualified lead delivered by the AI.

• Marketing platforms. Using the "pay per result" model for Google Cloud-powered AI platform that maximizes return on ad spend for small-to-medium enterprises.

As AI technology progresses, its impact across industries will be significant. In healthcare, AI agents can streamline patient engagement, lowering response times by more than 30%. In finance, AI can analyze vast datasets to detect fraud, improving detection rates significantly and saving firms millions each year. Organizations that invest in understanding these trends can maximize the advantages offered by AI technologies.

The Path Ahead

Sam Altman’s Dev Day provided valuable insights into the trajectory of AI, focusing on the emergence of AI agents and the potential of performance-based pricing. As developers and businesses navigate this fast-changing landscape, adaptability will be crucial for success. By testing these new trends, organizations can improve their AI capabilities, foster meaningful interactions with their users and do it in a sustainable way.