I spent the last 3 months in SF and most startups I heard of are trying to remove humans from the equation. We decided to build around the opposite belief and want to make sure humans stay relevant in the age of AI. 🧵

AI-Native Change Management is here

Consulting firms have been lying to themselves and their customers for months. They keep charging for the pre-AI length of missions while their consultants are usually able to complete them in 1/3 of the time using AI. They have historically been selling top 1% human intelligence at a very high price but AI is already able to do most of the work their consultants do today, for a fraction of the cost and time. Their customers know that hence their doubts on the current value proposition. They will not admit it, but human consultants are the bottleneck. Most firms are trying to augment consultants instead of redesigning the work. In discovery, for example, consultants still run the interviews manually, record them, and feed the transcripts into an AI model to structure the insights and draft recommendations. That makes each consultant faster but does not change the scale of the mission. They can still only interview as many people as human time allows. Today’s models can already do a large share of the work consulting firms charge millions of dollars for. And that share will only grow. There is no reason for humans to do work that agents can already do faster, better, and at scale. One of the usual objection is domain expertise that humans and consulting firms have built over time, but that is mostly a temporary and feedback loop problem. Humans should step in at the end of the loop, let the agents do most of the work, review their outputs, correct reasoning, and feed their expertise back after every mission in order to then build deeper industry and workflow-level knowledge over time. Mission after mission, this will become a dataset no human team will ever match. We are rebuilding the AI-native version of consulting at Foaster and our customers who used to hire MBBs love it.

The softwerisation of services is here.

Foaster (@Foaster_ai) is building AI agents to replace consulting firms. They map how your company works, build your AI roadmap in days, and continuously guide adoption, upskilling, and execution. Congrats on the launch, @RaphaelDabadie and @alexcmbs17! ycombinator.com/launches/QDn…

Meet Foaster. The new way for companies to transform for AI. Our agents do the massive work no human team can do at scale: map the company, track change, and guide every employee individually. Human experts step in only where humans matter most: judgment, trust, strategy, and execution. AI transformation will not scale through human teams alone. It will scale through agents.

We are live on YC’s socials !

We are launching on social media on Wednesday 👀🍿

Generic AI training doesn’t work. We fix that. Companies need their teams to build the right skills for their role, their level, their tools, and their day-to-day workflows. That level of personalization is impossible to deliver manually at scale. But AI makes it possible. @Foaster_ai

Coming soon 🎥👀 @Foaster_ai

AI adoption ≠ AI transformation The winners will redesign their org around AI, not bolt it onto the old one. Great take by @RaphaelDabadie from @Foaster_ai

The first step towards the AI-native version of your company @jack

We believe the future of consulting won’t come from augmenting the old model, but from rebuilding it from the ground up. AI makes that possible.

We are joining @ycombinator Spring 2026 batch ! At @Foaster_ai, we are building an AI-native alternative to consulting. We think the AI era opens the door to rethinking, from the ground up, how companies understand themselves and improve. We built Foaster around one core idea: while consultants and client companies usually have too little visibility into how work actually gets done, agents can build and reason on top of a much richer map of the organization. Through AI-led interviews, we rebuild the company’s operational graph: a structured view of its workflows, bottlenecks, dependencies, handoffs, and information flows. Already today, this lets us do in a few days what used to take consulting teams months: understand how a company actually operates at scale, surface inefficiencies, and show where AI should be implemented first. In the coming months, we believe agents built on top of this graph will be able to solve more and more of the problems that were traditionally handled by consulting firms. Super happy to learn alongside @collinmathilde, @dessaigne and so many inspiring founders ! 🧵 More below