Key Moments

How Amplitude Went From Skeptics to “All In” on AI

Y CombinatorY Combinator
Science & Technology6 min read45 min video
Dec 3, 2025|44,845 views|756|34
Save to Pod
TL;DR

Amplitude initially dismissed AI as "grifting" but now sees it as a necessary reinvention, launching new AI-powered features while acknowledging the challenges of integrating AI into an established company.

Key Insights

1

Amplitude was skeptical of AI until late 2024, viewing early models as "jagged" with unpredictable capabilities.

2

The company acquired Command AI and hired a new engineering leader, Wade Chambers, in October 2024 to spearhead their AI efforts.

3

Amplitude's AI Visibility product launch doubled new signups for their free plan, demonstrating the immediate impact of AI features.

4

The CEO emphasizes that AI adoption requires a "technology-first understanding of what is possible," as customers cannot articulate AI-driven needs.

5

Amplitude has undergone two major reorganizations in engineering, product, and design since early 2025 to accommodate its AI pivot.

6

For Amplitude, AI isn't replacing its existing roadmap but enhancing it, with a goal to rebuild the platform to be AI-native and easier to use.

Initial skepticism and the "AI strategy" question

Amplitude, a leading analytics platform, initially viewed AI with skepticism, particularly in 2022 and 2023. The team, including CEO Spenser Skates, perceived early AI models as "jagged" – excelling in some areas while being terrible in others. This led to frustration when faced with vague directives from executives or investors to "get our AI strategy." Skates pushed back against this top-down approach, encouraging teams to explore AI if they believed in its potential. This skepticism was partly fueled by observations of perceived "grifting" in the AI space, with many promising transformative changes without clear, demonstrable capabilities. The company's co-founder, Jeffrey, was particularly vocal about the disconnect between AI hype and practical application.

The turning point: Engineering productivity and key hires

The transformative effect of AI on software engineering, notably through tools like Cursor, became a significant catalyst for Amplitude's shift. Witnessing the productivity gains, the company decided to "go after this" in earnest around October 2024. This period marked two critical strategic moves: the acquisition of Command AI, a YC company with expertise in AI-driven user assistance, and the hiring of Wade Chambers as their new engineering leader. Chambers brought experience from the bleeding edge of AI model application, having worked on products designed to intelligently guide end-users, much like a chatbot assisting with confusion. His and Command AI's expertise provided the critical "change agents" Amplitude needed to move from skepticism to serious development.

Building an AI-native analytics platform

Amplitude's long-term vision for a "self-improving product" that dynamically responds to user feedback and behavior, once thought to be a decade away, now seemed achievable with AI advancements. This vision became a driving force for their AI integration. In the past few weeks, Amplitude launched several AI products, including AI Feedback and AI Visibility, with a major push planned for December, January, and February. A key upcoming launch is dubbed "the cursor for analytics," aiming to revolutionize how users interact with and leverage analytics data. This initiative required a significant organizational transformation, including two major reorganizations within the engineering, product, and design departments since early 2025, to align the company with its new AI-centric direction.

The challenge of AI adoption: Technology-first understanding

Unlike traditional SaaS product development, where customer feedback directly shapes the roadmap, AI requires a different approach. Skates explains that customers often cannot articulate what AI can do for them, leading to the "give me a faster horse" phenomenon. Instead, Amplitude emphasizes a "technology-first understanding of what is possible." This means familiarizing oneself with AI model capabilities and then mapping those to product functionalities. This perspective shift is crucial because AI's capabilities are "jagged," meaning direct customer requests might not align with what the technology can realistically and effectively deliver. The company is focusing on training its organization to understand and leverage these capabilities, moving beyond traditional customer-driven roadmapping.

Internal shifts and organizational restructuring

The transition to an AI-focused company involved significant internal changes. Amplitude has undertaken two reorganizations in its engineering, product, and design departments since the beginning of the year, moving out individuals who were deeply entrenched in the traditional SAS modality and not on the cutting edge of AI. This disruptive process also included acquiring several companies and integrating their teams, such as Craftful, Anari, and June, with long-term Amplitude employees. This melding of new talent and established expertise is seen as a key factor in their AI pivot. The company also ran an "AI Week" to train employees, which involved leaders demonstrating AI tools and applying them to real Amplitude workflows, including a live demonstration of a new feature that resolved a bug.

AI visibility and broader product strategy

Amplitude's AI Visibility product, which was launched quickly and given away for free, has served as a powerful lead generation tool, doubling new signups for their free plan. Skates views these "visibility businesses" as potentially commoditized, with future value lying "downstream" of the AI's core capabilities, similar to how SEO evolved. While AI enhances their existing products, Amplitude's broader strategy for the next year includes rebuilding Amplitude to be AI-native, making it easier to use, ensuring parity for their non-analytics products, and serving marketers effectively. Foundational work continues on products like session replay, with new features like "zoning" in development, demonstrating that AI is augmenting rather than entirely replacing their existing roadmap.

The mindset shift: From SAS to AI-native engineers

The transition from pre-AI SAS to AI-native product development necessitates a mindset shift, which can be influenced by age and experience but is fundamentally about adaptability. Skates draws a parallel to learning a sport or instrument: direct experience and coaching are key. He notes that engineers who thrive in the AI era are those less focused on code as an end in itself and more on solving customer problems with new tools. Conversely, AI-native teams might lack the deep understanding of established product workflows. The company encourages embracing iterative development, where failures are not immediate deal-breakers but opportunities to refine prompts and approaches, contrasting with the lower tolerance for errors in traditional SAS. This requires a "rewiring of the brain" to be patient and persistent with emerging AI capabilities.

Leadership evolution and the future of analytics

Transitioning from a founder-led startup to a public company CEO managing 800 employees has required Skates to evolve his leadership style. While founders often lead by example in the trenches, a large company executive must delegate and manage resources across various departments. This involves a disciplined approach to time management and a willingness to embrace hierarchy and established structures, which can be challenging to unlearn. The company's overall growth and resource allocation, with hundreds of millions in revenue, provide leverage for innovation. Skates remains committed to leading Amplitude through the "reinvention of analytics," aiming to create a "cursor moment" for the industry, where AI-driven analytics become the indispensable new standard, making the old ways seem obsolete.

Common Questions

Amplitude was initially skeptical about AI due to concerns about 'grifting' and the 'jagged' capabilities of early AI models, which were inconsistent. There was frustration from engineers seeing a lot of talk and not enough demonstrable action regarding AI's potential.

Topics

Mentioned in this video

More from Y Combinator

View all 576 summaries

Found this useful? Build your knowledge library

Get AI-powered summaries of any YouTube video, podcast, or article in seconds. Save them to your personal pods and access them anytime.

Try Summify free