---
type: "synthesis"
articles: ["a024", "a072", "a073", "a099"]
tags: ["agentic-ai", "autonomy", "lams", "service-as-software"]
id: "cross-agentic-trajectory"
sources: ["futures"]
sourceVaultSlug: "hbr-seg-futures"
originDay: 2
articleStem: "hbr-seg-futures"
sourceUrl: "(unified vault: 11 sources)"
sourceTitle: "HBR — Futures / Macro — where it's all going"
---
# From Generation to Action

Four articles describe the same escalator — LLMs that *generate* → agents that *act* → autonomous cross-silo systems — using different vocabularies that map cleanly onto each other.

**A073** gives the ladder: [[concept-large-action-models|LAMs]] ([[quote-llm-vs-lam|"LLMs predict what to say next; LAMs predict what should be done next"]]) → [[concept-personal-large-action-models|PLAMs]] → [[concept-corporate-large-action-models|CLAMs/GLAMs]], all inside [[concept-living-intelligence]].

**A024** gives the enterprise instantiation: [[concept-agentic-ai-systems]] — networks of specialized agents under a "digital manager" ([[entity-org-ema]], [[entity-org-anterior]]) — powered by the [[framework-five-forces]] and the [[concept-ai-driven-flywheel]], with [[concept-forward-deployed-ai-engineers]] as the go-to-market mechanism.

**A099** gives the business-model consequence: [[concept-service-as-software]] — AI "workers" that *do* the work, not tools for humans — enabled by [[concept-chain-of-reasoning]] and [[concept-recursive-algorithmic-development]] ([[quote-service-as-software]]).

**A072** gives the organizational response: [[action-modular-org-design]] — build modular structures *now* to anticipate agentic AI.

**The shared caveat:** across A073 and A024, "agent"/"LAM" labels overstate reality — much production "agentic" work is still orchestration around LLM calls plus deterministic workflows, and cross-silo autonomy remains gated by [[question-multi-agent-compliance|non-deterministic compliance]] risk. The trajectory is real; the arrival is uneven. See [[cross-beyond-llm-frontier]] and [[cross-judgment-accountability]].