Graphing Capability Sprawl - API Resilience Podcast with Daniel Kocot

In this episode of the API Resilience podcast, guest Daniel Kocot (adorsys) joins host Kristof Van Tomme (Pronovix) and they sit down for an informal, peer-to-peer brainstorming session about the messy realities of enterprise API architecture. Fresh off their discussions at API Days Munich, they "think out loud" about how to navigate API sprawl without falling into the common trap of chasing dogmatic "silver bullet" platforms. Instead of proposing rigid industry mandates, they explore how the conceptual alignment of a business capability and its integration boundaries can help shape better systems. They discuss the vital distinction between a technical feature and a business affordance: productized affordances are actually co-created with users over time, as reflected on mature developer portals, they challenge preset marketing narratives around the "API as a Product" concept. The second half of their conversation tackles how to visualize these complex, sprawling ecosystems. Daniel describes their recent PoC that programmatically pulls metadata from the API gateway together with observability data into a graph database. In a broader reflection on the graph database and the semantic web, they discuss how modern LLMs are reviving these mature concepts by automating semantic encoding, which previously required tedious manual ontology creation. Finally, they speculate on how a knowledge graph can enable intuitive semantic search by translating natural language questions into deterministic database queries to reduce AI hallucinations. They talk about why relational databases struggle with highly diverse API catalogs, how graphs handle flexible properties, and candidly discuss the current scaling limits of graph-based architectures.