Patrick Kemble

Migrating sprawling legacy systems, like 300,000-character stored procedures or thousands of queries running on decade-old databases, has long been a blocker to modernization. CHG Healthcare proved there’s another way—saving 15 months moving to dbt Cloud + Snowflake with AI-powered migration. We’ve distilled lessons like these into the definitive guide to dbt Labs migrations so data teams can modernize with confidence and avoid the pitfalls that stall projects. Get the guide here (it’s free). Link in comments.

16

2 Comments

How are you thinking about enterprise architecture for AI agents? Do you believe we’re heading toward a standardized agent infrastructure stack, similar to what LAMP (Linux, Apache, MySQL, PHP) did for the early web? I believe the next wave of enterprise software isn’t apps. It’s autonomous AI agents working together. But here’s the uncomfortable truth many leaders are starting to realize: Most enterprise architectures aren’t ready for it. As organizations rush to deploy AI agents, they’re running into a new kind of complexity—dozens (or hundreds) of agents connecting to APIs, databases, SaaS platforms, and internal systems. The result? Fragmented governance, security blind spots, and an integration nightmare. The 2026 Enterprise Agent Architecture Report explores one of the most important emerging ideas in enterprise AI infrastructure: the Agentic LAMP Stack and the role of MCP Gateways in making agent ecosystems scalable, secure, and manageable. Some key themes that stood out to me: • AI value is no longer limited by models — it's limited by architecture • Enterprises are moving from simple chatbots to multi-agent workflows • Integration complexity is becoming the #1 barrier to production AI systems • A new stack is emerging to standardize how agents reason, orchestrate, connect, and store knowledge If you're thinking about the future of enterprise AI, this report is worth your time. It provides a clear framework for how organizations will structure production-grade AI agent systems over the next few years. For a deeper dive into the architecture behind it, read more here: https://lnkd.in/e3Agb4_W We’re still in the early days of the agent economy. Maxim AI WunderGraph Jens Neuse Dustin Deus Björn Schwenzer Vendia Tim Wagner Shruthi Rao OpenClaw F5 Amazon Web Services (AWS) IBM Microsoft OpenAI Google Greg Anderson Sharath Rajasekar Cillian Kieran Alexandre de Vigan Klue

15

A small pod can ship what big teams keep delaying. The reason technology leaders use Tangonet Solutions LLC is not “cheaper engineers.” It’s speed to a real outcome when the internal team is already maxed out. This is reflected in our case studies: -modernizing and hardening cloud deployments -building production-ready data and release pipelines -shipping SaaS features without turning releases into events (the bad kind) There’s a common pattern: Constraint: A core team with ownership, but not enough capacity for a major initiative. Moves: A nearshore pod integrated into the same tools and cadence Clear ownership and a single point of contact Delivery discipline (CI/CD, environments in code, observability) Result: Roadmap items ship instead of rolling over sprint after sprint. If you want examples, our case studies are here: https://lnkd.in/eSUyFK7V If you are trying to ship something meaningful this quarter without over-hiring, DM me with the initiative and the deadline.

2