Monetizing AI — Alvaro Morales, Orb

Summary

In this insightful talk, Alvo, CEO and co-founder of Orb, delves into the complexities of AI pricing strategies. He explores the rapid changes in AI models, the challenges of managing margins, and the importance of measuring ROI for customers. Alvo discusses frameworks for deciding whether to monetize AI directly or indirectly, the significance of selecting the right value metrics, and the need for continuous experimentation in pricing. He also introduces Orb Simulations, a tool to help AI product teams test and optimize their pricing strategies, ensuring a data-driven approach to maximize revenue and market fit.

Takeaways

• 😀 AI pricing is challenging due to rapid changes in models, inference costs, and customer ROI expectations.
• 😀 Pricing AI products on 'vibes' may lead to revenue loss, as seen in the example of ChatGPT Pro’s pricing mishap.
• 😀 There are two main monetization strategies for AI products: direct (standalone products or add-ons) and indirect (driving behavior via upsells or free bundling).
• 😀 GitHub Copilot is an example of direct monetization, launching as an add-on for existing products.
• 😀 Notion AI shifted from direct monetization to bundling its AI features into higher-tier plans to encourage adoption.
• 😀 Expedia launched a free AI feature to turn Instagram reels into bookable trips, exemplifying indirect monetization.
• 😀 Selecting the right value metric is crucial, with options ranging from token-based pricing to task-based pricing and even outcome-based pricing.
• 😀 Task-based pricing, like Zapier's, focuses on the number of tasks or actions performed by AI in a workflow.
• 😀 Outcome-based pricing, though still in early stages, charges based on measurable results, such as Intercom's Finn charging per resolved customer support ticket.
• 😀 AI companies should continuously experiment with pricing models, as AI technology costs can fluctuate rapidly, requiring ongoing adjustments to pricing strategies.
• 😀 Orb’s simulation tool helps companies back-test pricing strategies, using real-time usage data to explore different pricing scenarios before launching a product.

Q & A

Why is AI pricing considered particularly challenging compared to traditional software pricing?

AI pricing is challenging because it needs to keep up with rapidly changing model and inference costs, the pressure on margins, and the need for customers to understand the ROI of AI technologies. The fast pace of innovation and unexpected adoption trends further complicate pricing decisions.

What are the three main challenges in AI pricing identified in the script?

The three main challenges in AI pricing are: 1) Rapidly changing model and inference costs, 2) Pressure on margins and cost of goods sold (COGS), and 3) Customers seeking clarity on the ROI of AI technologies.

What does the speaker suggest about how the AI industry should approach pricing?

The speaker suggests that the AI industry should avoid pricing 'on vibes' and instead use data-driven strategies, experimentation, and frameworks to find more effective and sustainable pricing models.

How does Orb help companies with their AI pricing strategies?

Orb helps companies by providing a platform that integrates real-time usage data and enables teams to simulate different pricing scenarios. This helps companies make data-informed decisions about how to price their AI products.

What is the significance of value metrics in AI pricing according to the script?

Value metrics are essential in AI pricing because they ensure that the pricing structure aligns with the actual value being provided to customers. The choice of value metric influences how customers perceive and are willing to pay for the AI product.

Can you give an example of a direct monetization strategy for AI mentioned in the talk?

One example of a direct monetization strategy for AI is GitHub Copilot, which was launched as a separable, monetizable add-on to the base GitHub seat. This model charges customers specifically for the added AI capabilities.

What are the challenges with outcome-based pricing in AI, as discussed in the script?

Outcome-based pricing in AI is difficult because it requires aligning both the customer and vendor on a clear definition of 'outcome' and having the ability to measure it objectively. This works well in clear-cut cases like customer support but is harder to apply in more complex AI applications.

Why is experimentation in AI pricing considered critical?

Experimentation is critical because the AI landscape changes rapidly, and companies need to continually evolve their pricing models to adapt to new technologies, customer expectations, and business needs. Regular experimentation allows companies to stay competitive and responsive to market shifts.

What is Orb's simulation tool and how does it assist companies with pricing decisions?

Orb's simulation tool allows companies to test different pricing models based on actual product usage data. By running these simulations, companies can evaluate how different pricing strategies might impact revenue, customer behavior, and overall business outcomes.

How do companies typically use Orb's platform during closed beta periods?

During closed beta periods, companies use Orb's platform to test pricing structures without actually billing their customers. This helps them gather feedback and adjust pricing strategies before launching to the broader market.