The forgotten operating system
What's unlocking the latest agent capabilities? Kanban boards. Explicit handoff protocols. Delegation with scope and boundaries. Root-cause analysis. Memory and learning loops. Specs and skill files as durable operating protocols.
None of this is new. Frontier firms are rediscovering and applying late-20th-century management practices that many organisations let atrophy.
Of course, lean and agile weren't invented to manage agents. They were invented to manage work across people and machines under uncertainty, where production was faster than coordination could keep up; the very conditions agent-infused work now operates in. Visible work-in-progress on a board. Root-cause analysis when something fails. Memory that survives the task that generated it. With an operational substrate, agent throughput compounds into capability that can connect across the horizontal work of an organisation.
Will "dusting off" lean and agile practices fix the absorption gap caused by the exponential capability of agents? Here's where the evidence is stacking up.
-
Sarah Friar, OpenAI·illustration
OpenAI's CFO rebuilt finance by mapping each decision back to its data, tools and approvals, keeping validation and final sign-off with humans.
Sarah Friar describes building finance around AI by mapping each consequential decision back through the data, tools, approvals and handoffs it requires, then assigning parts to agents while finance keeps validation and final authorisation. The operating discipline is explicit: defined data access, approval thresholds, escalation rules, every output tied to a source, and any baseline change requiring finance sign-off. It binds to Layer 3's claim that management as infrastructure, not raw model capability, is what turns agent throughput into deployable capacity, with the durable protocol located in the workflow redesign and the human control points rather than the model.
-
Fouad Bousetouane, arXiv cs.MA·evidence
Across regulated deployments, governance and context evidence decided which agents were production-ready; raw capability did not.
A governance-readiness index scoring agent deployments on evaluation, context, compliance, and whether the organisation can authorise, monitor, audit, and control the agent was validated across healthcare and finance, where it separated higher-risk from lower-risk configurations. Capability improved how the agents behaved but did not determine whether a deployment was ready to ship, and context engineering was the factor that moved reliability. This binds to the Layer 3 claim that management as infrastructure, not raw capability, is what turns agent throughput into safely deployable capacity, locating the decisive factor in context and auditable control rather than in the volume of governance added on top.
-
Liao et al. (Meta)·evidence
Storing an agent's exploration history as a queryable database, not disposable logs, reached target speedups 10x faster.
The paper treats the artifacts, tool outputs, rewards, and causal lineage an agent generates over hundreds of steps as governed, queryable database state rather than disposable session logs. Grounded in a production accelerator-kernel optimiser at Meta, cross-session reuse of that stored experience reached a target speedup about 10x faster and at lower token cost. It is a quantified production instance of the Layer 3 claim that memory which survives the task compounds agent throughput into durable capability.
-
Noah Zweben, Anthropic·illustration
Anthropic's production agents act under their own scoped service accounts, so each action logs against the agent rather than a borrowed human credential.
Claude Tag issues each agent a service account in every connected system, scoped to a compartment such as a Slack channel or workspace, so the access question becomes what an agent may do within its boundary rather than what a user may do.
-
Google Cloud·illustration
Google publishes the Open Knowledge Format, a vendor-neutral spec for knowledge as a directory of frontmatter-tagged markdown an agent can query and edit.
It is the second hyperscaler to formalise markdown directories as the operating substrate for agent-readable knowledge.
-
Anthropic·evidence
Anthropic's analytics agents answer ~95% of internal business queries at ~95% accuracy; without skills, accuracy never exceeded 21%.
-
Kieran Klaassen, Every·illustration
With a good plan and the right context, the build phase becomes routine and human effort shifts to the two ends of the loop, ideation and polish.
-
OpenAI Developers Cookbook·illustration
Codex's Goals feature replaces ask-and-wait prompting with a persistent contract the agent works until a named verification surface confirms completion.
-
Gottweis et al., Nature·evidence
A Gemini-based multi-agent system produced hypotheses that outranked the curating experts' best guesses on the hardest goals, validated in three wet labs.
-
Brandon Gell & Willie Williams, Every·evidence
Every rolled out a personal agent per employee, abandoned it within weeks, rebuilt around shared coworkers with explicit roles, group access, and group memory.
-
John Cutler, The Beautiful Mess·evidence
TBM 422: Exception, Presence, Delegation as the three-motion working repertoire of management as infrastructure.
-
Mike Krieger, Anthropic CPO·evidence
"Anthropic is shipping our harness strategy, not a product."
-
Helm, Priebe, Duderstadt, arXiv cs.MA·evidence
Adaptive control charts extend lean/SPC to multi-agent LLMs, exposing a fundamental learn-versus-monitor tradeoff.
-
Li et al., arXiv cs.MA·evidence
Constraint drift formalised: safety-critical constraints degrade through memory, delegation, communication, audit, and optimisation.
-
Mario Hayashi·illustration
Developer applies Jidoka, Poka yoke, Andon, Muda, and Five Whys by name to AI agent failures.
-
David Rice, People Managing People·evidence
"AI is the reward for operational discipline, but it won't create it."
-
Wongchamcharoen et al., CentaurBench (arXiv)·evidence
On three of seven real-world tasks, a weaker worker left unaided beat every condition where a stronger model assisted it.
CentaurBench scores models separately on automating a task and on assisting a weaker worker through it, and the two rankings barely track each other, so the strongest automator is rarely the strongest assistant. Adding a capable assistant to a work system is therefore unreliable and depends on the task, an independent reproduction of the Assign-All result where more intervention can degrade output. It complicates Layer 3's claim that management as infrastructure turns agent throughput into capability, and joins the counter-evidence showing the right operational substrate depends on the shape of the work rather than its quantity.
-
John Cutler, The Beautiful Mess·illustration
AI clears the blockers built from workflow friction and accelerates the ones built from dysfunction.
Cutler splits organisational blockers in two. Where a practice is held back by poor signal visibility, workflow friction, or missing scaffolding, AI removes the drag and can sustain something a team wanted but could never afford; where it is held back by low voice safety, misaligned incentives, or fear, the same tooling makes the avoidance and performance theatre cheaper to automate. This scope-limits Layer 3: rediscovering lean and agile practices absorbs agent capability only when the organisational substrate already favours improvement, and the same infrastructure accelerates dysfunction when it does not.
-
Rohit Krishnan, Strange Loop Cannon·evidence
Blended AI councils kept only a quarter of the good ideas a single model raised.
Krishnan decomposed model answers into atomic idea cards, clustered them, and had two blind judges rate which survived under different council structures. A peer-review round behaved as a consensus detector, promoting ideas several models shared while dropping about three quarters of the useful ideas only one model had raised, and the structure that recovered them stored and ranked each contribution before any synthesis. This is the fourth multi-agent result, after Assign-All, Too Many Specialists, and CEAD, showing that coordination structure must be designed and evaluated per problem, and it points the remedy at memory that survives the task rather than at more deliberation.
-
SkillSafetyBench authors, arXiv cs.AI / cs.MA·evidence
Even with benign user requests, task-relevant skill materials can steer agents toward unsafe actions.
That is direct counter-evidence to the framework's assumption that markdown skill files are neutral, durable carriers: the properties that make them durable (text, addressable, transferable, consumed downstream) are what make them a transferable attack surface.
-
John deVadoss, arXiv cs.MA·evidence
CEAD: design quality dominates governance in agent architecture outcomes across a 10K-task benchmark.
Capability-aligned design beats control-heavy / design-poor architectures (70.6% vs 50.8% safe success); pulls the framework toward design first, governance second.
-
Masters & Albrecht et al., DeepFlow / arXiv·evidence
Assign-All baseline outperforms chain-of-thought management 0.502 to 0.313 across twenty enterprise workflows.