Zillow Turns AI Productivity Gains into Outsize Revenue Growth
Zillow Group is emerging as a case study in how AI productivity gains can translate into outsized software growth. The company reported an 18% revenue increase to USD 708 million (approx. RM3.26 billion) in a housing market it described as essentially flat, and it directly credited AI for part of that outperformance. Engineers are shipping 40% more code on average thanks to internal AI tools, helping Zillow move features from concept to launch much faster. Beyond engineering, AI agents are embedded throughout its products: an AI-powered consumer search experience, an AI workflow engine inside Follow Up Boss for real estate teams, and AI Assist for rental leasing that automates lead management and screening. Together, these AI initiatives let Zillow deliver more value and new capabilities without expanding headcount at the same pace, improving operating leverage even in a stagnant end market.
HubSpot’s Autonomous Customer Support Agent Reaches 70% Resolution
HubSpot is showcasing what autonomous customer support can look like in production. Its Customer Agent now resolves 70% of support conversations on its own, up from 20% a year earlier and five points higher than last quarter. Some customers are already clearing 85–90% autonomous resolution, using the agent primarily for after-hours coverage and tier-one tickets so human teams can focus on complex issues. Customer Agent has surpassed 9,000 customers and consumes over half of all AI credits on the HubSpot platform, with total AI credit usage climbing 67% quarter over quarter. These numbers demonstrate that AI agents can deliver tangible AI productivity gains in frontline operations, absorbing large volumes of routine work without proportional staffing increases. Yet, despite this evidence of enterprise AI ROI, HubSpot’s shares still fell sharply after earnings, underlining that investor sentiment around AI economics remains volatile.
Amplitude Bets on AI Agents and Statsig to Power Product Growth
Amplitude is repositioning itself as a core instrumentation and experimentation layer for digital products, with AI agents at the center of its growth story. The company has spent several years consolidating adjacent capabilities—experimentation, session replay, guides, surveys, and web analytics—onto a single platform. Its agreement involving Statsig assets and customers expands that strategy, adding strength in data warehouse-based experimentation and feature flagging for environments like Databricks and Snowflake. Statsig’s technology is already used by AI-focused companies, including as an internal feedback loop for user interactions with ChatGPT, underscoring its relevance to AI-native workflows. Amplitude expects to assume about 320 customers through the deal and sees substantial opportunity in a Global 2000 base it describes as underpenetrated. By combining observability, experiment management, and AI agents, Amplitude aims to help customers rapidly test and deploy AI-powered features without dramatically scaling headcount-heavy analytics teams.
New Software Economics: More Output, Less Headcount Growth
Across Zillow, HubSpot, and Amplitude, a new pattern in software economics is emerging: AI agents are enabling companies to deliver more value without matching increases in headcount. Zillow’s 40% jump in code output illustrates how AI copilots can compress development cycles. HubSpot’s Customer Agent shows how autonomous customer support offloads repetitive work, allowing human agents to focus on high-value tasks rather than sheer volume. Amplitude is positioning its platform as the infrastructure for this shift, giving product teams the instrumentation and experimentation tools to deploy AI features faster and with more confidence. Collectively, these changes point to a structural productivity shift where AI agents, not just human teams, become core contributors to growth. The implication for the broader industry is clear: competitive advantage will hinge on how effectively companies embed AI agents into workflows to capture enterprise AI ROI, rather than simply layering AI on top of existing processes.
Investor Skepticism and the Next Phase of AI Agent Adoption
Despite strong operational metrics, market reaction suggests investors are still grappling with how to value AI agents software growth. HubSpot’s stock dropped around its Q1 results even as Customer Agent showed rapid adoption and a steep rise in autonomous resolution rates. That disconnect highlights lingering questions: Are AI agents driving durable revenue, or is usage still too early-stage and promotional? Will AI productivity gains translate into sustained margin expansion, or will competitive pressure force vendors to pass savings back to customers? Zillow and Amplitude, meanwhile, are leaning into AI narratives, emphasizing proprietary data, end-to-end workflows, and platform breadth as defensible moats against general-purpose AI tools. As more enterprises adopt autonomous customer support and AI-driven experimentation, the next phase will hinge on clearer financial proof—recurring AI-driven revenue streams, improved unit economics, and measurable AI ROI that can calm investor uncertainty and redefine how software companies are valued.