Software
The expert constraint in software teams is already well understood. It's the senior developer. The architect. The person whose judgement gates everything downstream. ViAGO has spent 20+ years building the systems that manage this constraint properly. PACE is the formal methodology. It's documented, productised, and proven at scale.
Book a CallThe Constraint
Agents dramatically increase the volume and speed of work moving through earlier stages of development. Generating code, writing tests, drafting specs, handling tier-one support. All of that work ultimately hits the same expert bottleneck for review, architectural decisions, and problem resolution. Throughput earlier in the pipeline goes up. Expert capacity stays flat. The constraint gets hit harder, not lighter.
There's a second problem most software leaders haven't seen clearly yet: work-in-progress explodes. When agents can spin up work faster than humans can review and direct it, unfinished work accumulates faster than it ever did with human teams. High WIP is the enemy of flow and quality. Firms that invest heavily in AI tooling without the constraint management layer will find their best people buried under an avalanche of agent-generated work that nobody has the bandwidth to properly oversee.
The Training Trap
Writing boilerplate code. Setting up test scaffolding. Handling routine tickets. Juniors who are currently learning these tasks are being trained for exactly the work that agents will take over.
The remaining human work is different: architectural decisions, resolving ambiguous requirements, diagnosing why an agent produced the wrong output, managing a system of AI tools rather than executing tasks within it. That requires systems thinking. Almost no software education or training develops it.
The System
PACE is ViAGO's TOC-based flow management system for software teams. Three rule sets: Bottleneck Rules that protect and offload the expert constraint. Variation Management Rules that protect the constraint, make time visible, and build disciplined responses. Flow Management Rules that keep work moving in small chunks, limit work-in-progress, and stop the pressure to start new work from killing your ability to finish existing work.
PACE is documented and productised. Peter Cronin wrote the book on it. It is what makes human teams capable of managing alongside AI agents rather than being overwhelmed by them.
Read PACE on AmazonHow We Help
Before investing in AI tooling, identify where the real constraint is and what automation will actually do to it. The most common failure mode in software AI adoption: automating the wrong things, or automating in ways that flood the bottleneck with more work than it can absorb. Low-cost, high-value diagnostic that creates the roadmap for everything else.
The flow management system for your team. Bottleneck rules, variation management, flow management. Not a methodology you read about and try to remember. A working system your team operates daily, with visible signals and pre-committed responses. Can be implemented alongside existing workflows without tearing up what already works.
The capability layer that makes your people effective at every stage. Systems thinking, cause-and-effect reasoning, conflict resolution. In a software context, the specific use case is diagnosing why an agent produced the wrong output and adjusting the system. Also the prerequisite for anyone expected to manage a team that now includes AI agents alongside human developers.
The Treadmill
The constraint doesn't go away by adding headcount, and it doesn't go away by adding agents. It goes away by managing it properly. If your senior developers are perpetually overloaded, if sprints keep missing with no clear systemic diagnosis, if more tooling hasn't translated to more throughput, the problem isn't effort or automation. It's flow.
Book a Call