What separates the AI winners from everyone else? | First Pass Ep. 3 with Ara Kharazian (Ramp)

Summary

Ara Kharazian breaks down Ramp’s AI Impact on Jobs report, which matches Ramp firm-level AI spend data with Revelio Labs workforce data to study what happens after companies adopt AI. The core finding is counterintuitive: the strongest adopters did not cut headcount, but grew it, with gains concentrated among firms using AI in advanced ways rather than casual enterprise chat. He says the effect appears with a 6 to 12 month lag and is tied to multiple models, APIs, coding agents, and workflow integration. The headline result is about 10% headcount growth over two years after adoption and 12% growth in entry-level hiring, which he interprets as evidence that AI reconfigures tasks and expands firms rather than simply eliminating jobs. He also argues the job market and policymakers should focus on adoption intensity, AI-native talent, and sector-level laggards like construction, manufacturing, and retail.

Notes

• Ramp’s AI jobs report is built on firm-level AI spend matched to workforce records.
• The study asks whether jobs rise or fall after firms buy AI.
• The key unit of analysis is the firm, not the individual employee.
• Intensive AI users see the gains; casual users generally do not.
• The headline headcount effect is 10% growth over two years after adoption.
• Entry-level hiring grows 12% over the same two-year window.
• Advanced AI use is the real threshold, not just exposure to a chatbot.
• Ara says AI adoption does not immediately show up in outcomes.
• The measured latency period is roughly 6 to 12 months after adoption.
• There is also a threshold effect: more advanced use is required before gains appear.
• Advanced users are described as using multiple models, APIs, coding agents, and integrated workflows.
• Enterprise chat subscriptions are treated as relatively weak AI usage.
• Best practices are concentrated in frontier firms and not yet widespread.
• Firms using AI tend to accelerate along their existing growth paths.
• AI changes the mix of jobs over time rather than instantly destroying roles.
• Job-function effects appear broad rather than confined to one or two departments.
• The report links higher entry-level hiring to AI-native young workers.
• Recent grads may be advantaged because firms want people who already know how to use models well.
• The hiring signal may be a practical search for AI-native talent.
• Ara argues AI-driven price deflation is not automatically a good outcome.
• He connects that view to the Federal Reserve’s positive inflation target.
• He uses the Fed’s 2% target to argue against celebrating deflation from automation.
• From a labor perspective, AI appears to reallocate work rather than simply eliminate it.
• The job market does not yet expose which firms are likely to grow faster from AI adoption.
• If two firms are otherwise similar, the AI user is likely the faster-growing employer.
• The result has direct implications for job seekers choosing among employers.
• Ara says policymakers should push AI adoption beyond tech.
• The policy focus should be on spreading frontier practices to non-tech firms.
• Outside-tech firms are not exposed to the same AI adoption bubbles as frontier companies.
• Construction, manufacturing, and retail are highlighted as lagging adopters.
• Retail employs more people in the United States than tech.
• Lagging sectors could still see outsized benefits if they move from low-intensity to high-intensity AI use.
• The report’s core message is that adoption quality matters more than adoption quantity.
• AI likely acts as an accelerator of firm growth rather than a pure cost-cutting tool.
• The vast majority of AI-using companies do not see gains, according to Ara.
• The practical takeaway for workers is to look for employers with deeper AI integration.
• The practical takeaway for companies is that advanced AI adoption can support broader hiring.
• The practical takeaway for the economy is that broader diffusion, not just frontier labs, matters.