Agentic Mesh: Towards Enterprise-Grade Agents | by Eric Broda | January, 2025
Enterprise-grade capabilities will make or break the future of agent acquisition. Without them, agents will remain “science experiments” and will never get out of the lab and into production.
The first wave of agent adoption is clearly upon us. However, when I talk to employees who work with agents I hear the same themes: businesses “die the death of 1000 agents proof of concepts”, or I hear the implementation of an agent is called a “scientific experiment”, it never leaves the lab and enters. production.
I think POCs and testing are the first steps in building the technology, however, to get businesses to adopt agents, two things have to happen. The first is simple: clear business value needs to be identified. If you work with agents (and can make them work reliably) then you probably see great potential. This is a solvable problem.
Which leads to the second thing – the hard thing! You can't see the value from agents until you put them into production. Unfortunately, today's crop of agents is not suitable for production. And this turns out to be a major obstacle in getting agents to be productive and, thus, deliver real value.
