Replit's CEO said it had dropped Anthropic over cost and performance. The question that tests whether a go-to-market story holds.
On October 16, 2026, TechCrunch reported that Replit is no longer working with Anthropic, according to CEO and co-founder Amjad Masad, who said so Thursday onstage at TechCrunch Disrupt. Masad framed it as a measured decision. "I think ChatGPT is better now," he said, citing an internal benchmark and adding: "We publish a public benchmark as well, and consistently, Sol and Astra perform better on building vibe-coded applications from scratch. It was a decision based on both cost, partnership, and performance. And for all these reasons, we just felt like they were a better partner for now." The account was disputed. One person familiar with the partnership, who asked not to be named, told TechCrunch it was still in place, and the outlet pointed to a Replit blog post from late September citing Anthropic's Claude Fable 5 and Claude Fable 5.1 as models running in Replit Agent's production environment. TechCrunch said it had reached back out to Replit and would update the story.
Set the dispute aside, because the more useful thing happened in the same appearance. Masad described how the business was moving: "enterprise was like 2% of revenue last year," he said, and "This year, I would say like 15% of our revenue is enterprise contracts, and more than 50% of our revenue is B2B." Put the two halves together and what a founder did onstage was tell a go-to-market story under live audit. He named the criteria behind a platform decision, named the instrument that measured it, and described the customer shifting underneath the product. Within the same news cycle, the account was being read against the company's own blog and social posts. That is the behaviour a go-to-market story has to survive, and it is why an interviewer asks for one. The question is never really about the launch. It is about whether the candidate's account of cause and effect holds when someone checks.
Why they ask it
Tell me about how you brought a product to market is broad on purpose. Breadth is the test. The interviewer wants to see what the candidate puts at the centre of the story when nothing forces the choice. Strong candidates centre a decision: who the customer was, what the product was not going to be, how success was measured, and when the measurement first said something. Weak candidates centre the artefact: features, a launch date, a press cycle. The question also tests ownership of sequencing. Bringing a product to market is mostly ordering, which segment first, which integration first, which number to watch before spending on the next. A candidate who can explain the order and what it cost has done the job. A candidate who can only list what shipped has watched it.
The trap
The first trap is the launch tour. The answer walks the roadmap chronologically, lands on a launch date, and closes by saying it went well. Nothing in it can be wrong, because nothing in it is a claim. The second trap is borrowed causality: the market moved, the product was present, and the story quietly credits the plan. Interviewers have heard the pattern often enough to probe it on instinct. The third trap is the soft number. Onstage, "like 15%" is a fair estimate from a founder who has said he is not yet sharing ARR numbers. In an interview the same hedge reads differently. It suggests the candidate did not own the metric closely enough to know it, and the interviewer starts to wonder what else in the story is approximate. The strong answer names criteria and a measure, and it is already reconciled with the record, because someone will check.
Applying STAR-T
Take a hypothetical: a usage-based billing API that had run as an internal pilot and needed to reach paying external customers.
Situation. The API had three design-partner customers, no pricing, and a sales team asking for a launch date so they could start quoting it. Enterprise prospects wanted it too, but each wanted a different integration first.
Task. Own general availability end to end: choose the first segment, set pricing, decide the integration order, and define the metric that would say whether the launch was working before revenue could.
Action. The candidate ran a four-week cohort with the design partners and found that mid-market customers reached first value in under two weeks while enterprise customers stalled in procurement. On that evidence the candidate chose mid-market first, shipped the one integration all three partners shared, cut a reporting feature from the launch scope, and set activation within fourteen days as the leading metric, with a written threshold below which pricing would be revisited.
Result. By the end of the first quarter the product had cleared the activation threshold, the pricing held without discounting, and the first enterprise deals closed the following quarter on the integration roadmap the launch data had justified. Each of those is a sentence the candidate can defend with a number if asked.
Trade-off. Choosing mid-market first delayed the enterprise pipeline by roughly two quarters and frustrated the salespeople who had those deals in hand. The candidate names that cost without being asked, because a go-to-market story with no cost in it is a story in which no decision was made.
The follow-up that breaks weak answers
What did the numbers look like the month before you decided, and who else had seen them? That is the interviewer doing what the news cycle did to Replit: comparing the account to the record. A candidate who made the decision answers immediately, because the number was the reason. A candidate who retrofitted the decision to the outcome has to reconstruct it, and the pause is audible. The second-best follow-up is What would have made you reverse it? Masad's phrase "a better partner for now" is a real answer to that question, whatever else the day disputed. It states that the decision is conditional and implies the conditions. A candidate who can say the same about their own launch, with the condition named, has shown the interviewer the thing the question was built to find.
Score your answer against the director’s bar
Q: Tell me about how you brought a product to market.
Bank the go-to-market story you would tell under that follow-up, and rehearse it until the dates and numbers come out the same way twice. Try it free →