Why Enterprise AI Demands More Platform Engineering, Not Less | Weave Intelligence

Why It Matters

As enterprises adopt AI and expand who can build software, robust platform engineering becomes essential to scale safely, control costs, and accelerate delivery; without it, organizations risk fragmentation, duplicated work, and slower innovation. Centralized platforms enable citizen developers while preserving operational standards and speed at large scale.

Summary

Boyan, CTO at Sixt, says the company’s product and engineering organization totals about 800 people and builds over 95% of its core software in-house across hundreds of countries. Over the past decade Sixt has invested in broad platform engineering—MLOps, developer self-service and heavy automation—to remove cognitive load from product teams, reduce fragmentation, and speed time-to-production. The platform now supports thousands of non-engineering “builders” from business units and has doubled down on AI tooling in the past year to manage rapidly evolving engineering practices. The goal: centralize plumbing, cut costs, and let product teams focus solely on customer-facing features rather than infra and deployment toil.

Original Description

What does platform engineering look like at enterprise scale when AI is rewriting every assumption about how software gets built? In this episode, Luca sits down with Boyan Ivanov, CTO at SIXT, to unpack a 10-year platform engineering journey that took one of Europe's largest mobility companies from 1–2 deployments per month to over 130,000, powering thousands of applications across hundreds of countries with 95% of core software built in-house.

Boyan shares why SIXT bet against the conventional "best tool for the job" wisdom a decade ago, how an early dogfooding approach grew into a platform serving the entire engineering org, and why the last 12 months of AI have only doubled down on that strategy. The conversation goes deep into the new realities enterprise platform teams are facing: demos that look amazing until you try to ship them at scale, fragmented data that breaks agents before they start, citizen developers building production apps without touching code, and agents quickly becoming the #1 user of the platform.

In this episode:

- The 10-year SIXT platform journey — from monolith to 130,000 deploys/month

- Why "best tool for the job" was the wrong advice for the enterprise

- Why AI is turbocharging the need for platform engineering in the enterprise

- The expanding audience: from app developers to data teams, citizen developers, and agents

- Why APIs built for deterministic systems aren't good enough for agents

- Identity, access, and observability rebuilt for an agent-first enterprise

- The coming "vulnerability apocalypse" and why security becomes the baseline concern

- Will enterprise teams shrink or scale up in the agent era?

💬 "If you have your data in 20 or 30 different silos and you just point your agent at it - yeah, it hallucinates a lot. Even a million-token context window isn't enough."

Learn more: https://weaveintelligence.io/