Onboarding Drives Enterprise AI Adoption
Eric Simons, CEO of Bolt, on consumer vs. B2B vibe coding
Enterprise AI coding will be won less by the best demo and more by the team that can turn a messy pilot into a working habit inside a company. Bolt is already seeing that large customers need help plugging the product into real codebases, design systems, and internal workflows, because the hard part is not buying a license, it is getting product managers and engineers to use it enough that renewal becomes obvious. That is why enterprise services are becoming part of the product in this category.
-
Bolt says enterprise onboarding started as manual work. The team would ingest a customer codebase, map the design system, write tooling, and only then automate the workflow. That shows the product still needs hands on setup to reach production ready use inside a company.
-
This is a familiar pattern in builder software. Airtable expanded enterprise sales, support, and services past 50 people, then used implementation specialists and integration engineers to design schemas, write docs, run trainings, and keep deployments from stalling out after purchase.
-
The risk is fake enterprise traction. Bolt expects B2B to pass 50% of revenue this year and says enterprise is about 25% of revenue today, but it also warns that a paid pilot or one year contract means little if customers never build durable workflows and churn at renewal.
The next phase of AI coding in the enterprise looks more like customer success plus solutions engineering than classic self serve SaaS. The winners will be the companies that can package onboarding into a repeatable motion, get real apps into production, and turn scattered curiosity from PMs and developers into embedded, renewing usage across teams.