Replit, Gamma, Lovable, and Anthropic All Built Sales Teams Eventually — Just Later Than the Last Generation Did

Every AI-native company that grows past initial self-serve momentum still ends up building a dedicated sales organization — the difference in 2026 is timing, not whether it happens at all. That’s the argument SaaStr makes using Replit, Gamma, Lovable, and Anthropic as examples: companies that scaled on product-led growth before eventually standing up sales infrastructure later in their growth curve than prior generations of SaaS companies typically did. The piece is a direct rebuttal to a narrative that’s been circulating in 2026 AI and startup discourse — that AI-native products can scale indefinitely on pure product-led growth without ever needing outbound sales, account management, or enterprise deal teams. The counter-claim isn’t that AI reduces the need for a sales team; it’s that AI delays when the sales buildout starts, not whether it starts.

If sales teams are still inevitable, the shape of the job inside them is changing faster than the decision to build one. A recent piece on SDR skills cites data showing SDRs currently spend roughly 40% of their time on manual tasks — prospect discovery, lead scoring, CRM logging — that AI can now automate, and that 36% of SaaS companies cut SDR/BDR headcount in the past year while 28% grew AE teams over the same period. The piece cites a Gartner projection that AI agents could outnumber human sellers ten-to-one by 2028, even though fewer than 40% of sellers currently report AI actually improving their productivity. The skills it argues will define the surviving SDR role — agent orchestration, industry fluency, channel range, and product-led motion selling — describe a job that looks less like manual prospecting and more like managing a small fleet of agents toward pipeline.

Put the two pieces together and the sales-team debate resolves into two separate questions that keep getting collapsed into one: whether a company builds a sales team (yes, eventually, per SaaStr’s examples) and what that sales team spends its time doing (decreasingly manual, per the SDR-skills data). Revenue leaders planning headcount around either question in isolation are answering the wrong half of it.