Has anyone else noticed that the hardest part of building agent workflows isn't the model anymore?
Summary
A developer observes that the biggest challenge in building agent workflows has shifted from models and prompts to maintaining context, business rules, and external data synchronization, and asks the community about their primary bottleneck.
Similar Articles
Most of our “agent” problems turned out to be workflow/state problems
A developer recounts how many challenges in building AI agents actually stem from workflow and state management issues, not model intelligence, emphasizing the need for robust state handling and observability.
The hardest part of building AI agents has nothing to do with AI
A practitioner argues that the real difficulty of building AI agents is not the model but data quality, failure-path design, process discovery, scope discipline, and handoff — sharing concrete checks and rules for deciding what to automate.
I think the AI agent conversation is about to move beyond frameworks
The author argues that building AI agents is no longer the hard part; the real challenges are deployment, testing, version control, and operational management, which remain fragmented in the ecosystem.
AI agents are starting to expose how broken most workflows already were
The article argues that AI agents are revealing how unstructured and chaotic many corporate workflows actually are, suggesting that successful automation depends more on clean systems and documentation than on advanced models.
Anyone else talking clients OUT of building an agent more often than into it?
The author discusses how clients often request AI agents without having clear underlying processes, and emphasizes the importance of documenting workflows before implementing automation to avoid blame when issues arise.