{
  "schema": "hangingcontext.desk/1",
  "type": "story",
  "id": "1486",
  "slug": "agent-ecosystems-next-adoption-level",
  "edition": "2026-11-13T00",
  "url": "https://hangingcontext.com/stories/2026-11-13/agent-ecosystems-next-adoption-level/",
  "newsroom": "artificial-intelligence",
  "beat": "developers",
  "newsroom_title": "Artificial Intelligence",
  "angle": "Most developers use coding agents weekly, but few run them as a system; the next levels are agent teams and simulations.",
  "headline": "Most developers use agents. Few run them as a system.",
  "dek": "Nine in ten professional developers now use AI coding agents at least weekly, but only about a quarter use automated agent workflows. The next adoption levels, one writer argues, are ecosystems of competing agent teams and simulations run repeatedly to filter out luck.",
  "published_at": "2026-11-13T00:51:10Z",
  "labels": {
    "trust": "confirmed",
    "trust_text": "Confirmed · 3 Manifests",
    "newsroom": "Artificial Intelligence"
  },
  "source_count": 3,
  "publishers": [
    "every-media-blog",
    "waydev-blog",
    "zep-blog"
  ],
  "sources": [
    "every-media-blog",
    "waydev-blog",
    "zep-blog"
  ],
  "distinct_source_count": 3,
  "key_facts": [
    {
      "text": "90% of professional developers use AI coding agents at work at least weekly, Waydev reports, and 68% use AI coding agents daily.",
      "cites": [
        {
          "claim_ids": [
            "1794498052242757595",
            "1794498052245757596"
          ],
          "manifest_id": "1794497840882756338"
        }
      ],
      "self_reported": true
    },
    {
      "text": "Approximately 26% of developers use automated agent workflows, according to Waydev.",
      "cites": [
        {
          "claim_ids": [
            "1794498052254757602"
          ],
          "manifest_id": "1794497840882756338"
        }
      ],
      "self_reported": true
    },
    {
      "text": "Level 9 of AI adoption involves designing an ecosystem of competing agent teams, and Level 10 involves running simulations multiple times to filter out chance, writes Mike Taylor of Every Media.",
      "cites": [
        {
          "claim_ids": [
            "1794527395261953884",
            "1794527395265953885"
          ],
          "manifest_id": "1794527135701953154"
        }
      ],
      "self_reported": true
    },
    {
      "text": "Token prices have dropped 90% in recent months, enabling more complex agent workflows, according to Every Media.",
      "cites": [
        {
          "claim_ids": [
            "1794527395276953892"
          ],
          "manifest_id": "1794527135701953154"
        }
      ],
      "self_reported": true
    },
    {
      "text": "55% of teams that built their own company brain gave up, almost always on maintenance, Zep reports.",
      "cites": [
        {
          "claim_ids": [
            "1794522998309932981"
          ],
          "manifest_id": "1794522521336930888"
        }
      ],
      "self_reported": true
    }
  ],
  "images": [],
  "sections": [
    {
      "heading": "Use is universal. Systems are rare.",
      "paragraphs": [
        {
          "cites": [
            {
              "claim_ids": [
                "1794498052242757595",
                "1794498052245757596"
              ],
              "manifest_id": "1794497840882756338"
            }
          ],
          "markdown": "90% of professional developers use AI coding agents at work at least weekly, and 68% of professional developers use AI coding agents daily, according to Waydev. The coding agent is no longer an experiment. It is the default way professional developers work, and the holdouts are now the exception rather than the rule."
        },
        {
          "cites": [
            {
              "claim_ids": [
                "1794498052254757602",
                "1794498052257757604",
                "1794498052258757605"
              ],
              "manifest_id": "1794497840882756338"
            }
          ],
          "markdown": "Approximately 26% of developers use automated agent workflows, and innovators, 2.5% of the population, are running fleets of agents coordinated by specs and automations. Early adopters, 13.5% of the population, have integrated agents into their pipelines for triage, tests and code review. The chasm in AI adoption exists between teams that use agents and teams that have learned to work with them. The numbers describe a steep pyramid: nearly everyone uses an agent, a quarter automate with them, and a small slice runs them as a coordinated system."
        },
        {
          "cites": [
            {
              "claim_ids": [
                "1794498052247757597",
                "1794498052248757598",
                "1794498052250757599"
              ],
              "manifest_id": "1794497840882756338"
            }
          ],
          "markdown": "Claude Code usage among professional developers grew from 18% to 39% between January and mid-2026, and Codex usage grew from 3% to 16% between January and mid-2026. 87% of businesses buying a coding agent purchase Claude Code. The market has a clear leader in this slice of the stack, and the growth is fast enough that a team's tool choice from January may already be out of date."
        },
        {
          "cites": [
            {
              "claim_ids": [
                "1794498052251757600",
                "1794498052253757601",
                "1794498052255757603"
              ],
              "manifest_id": "1794497840882756338"
            }
          ],
          "markdown": "Agentic token usage on OpenRouter passed human usage on February 6, 2026, and grew 14x in the six months following February 2026. Engineering organizations are currently struggling to balance increased code volume with delivery, quality and review capacity. Agents now write more code through OpenRouter than humans do, and the bottleneck has moved from writing code to reviewing it. That is the constraint that decides whether the productivity gain survives contact with production."
        }
      ]
    },
    {
      "heading": "Levels nine and ten",
      "paragraphs": [
        {
          "cites": [
            {
              "claim_ids": [
                "1794527395277953893",
                "1794527395261953884",
                "1794527395265953885"
              ],
              "manifest_id": "1794527135701953154"
            }
          ],
          "markdown": "Level 8 involves an orchestrator agent taking tasks off a to-do list and escalating decisions. Level 9 of AI adoption involves designing an ecosystem of competing agent teams, and Level 10 involves running simulations multiple times to filter out chance. The shift from level 8 to level 9 is the shift from one agent doing jobs to many agents competing to find the best strategy, and level 10 adds repetition so that a good result can be separated from a lucky one."
        },
        {
          "cites": [
            {
              "claim_ids": [
                "1794527395276953892",
                "1794527395271953889",
                "1794527395274953891"
              ],
              "manifest_id": "1794527135701953154"
            }
          ],
          "markdown": "Token prices have dropped 90% in recent months, enabling more complex agent workflows, and that price drop is what makes the higher levels affordable. Synthetic Slack workspaces allow clones to argue and react to each other to improve ideas, and a Hacker News simulator using 2,000 AI personas predicted front-page headlines with 60% accuracy. Running two thousand personas to test an idea was unthinkable at old prices. At current prices it is a budgeting decision."
        },
        {
          "cites": [
            {
              "claim_ids": [
                "1794527395267953886",
                "1794527395268953887",
                "1794527395270953888",
                "1794527395273953890",
                "1794527395278953894"
              ],
              "manifest_id": "1794527135701953154"
            }
          ],
          "markdown": "Dan Lens is a prompt that predicts how Dan Shipper would reply to Slack threads, and it correctly predicted 26% of points on its first guess and 86% after 100 guesses. KatePass replaced a rules-based editing system and improved performance on drafts. A 2006 study by Columbia University and the Santa Fe Institute showed how luck influences idea success, which is the intellectual root of level 10: if luck decides which ideas win, the answer is to run the experiment again and again until luck averages out. Level 9 is useful for growth experiments, pricing strategy and content strategy, and Level 10 is useful for high-stakes decisions where outcomes depend on social reactions."
        }
      ]
    },
    {
      "heading": "The layer underneath: the company brain",
      "paragraphs": [
        {
          "cites": [
            {
              "claim_ids": [
                "1794522998301932978",
                "1794522998311932982",
                "1794522998313932983"
              ],
              "manifest_id": "1794522521336930888"
            }
          ],
          "markdown": "A company brain is the infrastructure layer between a company's data and its agents, and it keeps company knowledge up to date and structured for agents to retrieve. The concept matters because agents without reliable context guess, and agents with it answer. More than a dozen companies have described their own version of a company brain since December 2025, which suggests the idea has moved from one company's internal project to a recognised pattern."
        },
        {
          "cites": [
            {
              "claim_ids": [
                "1794522998309932981",
                "1794522998305932979",
                "1794522998307932980",
                "1794522998315932984"
              ],
              "manifest_id": "1794522521336930888"
            }
          ],
          "markdown": "55% of teams that built their own company brain gave up, almost always on maintenance. The failure rate is the honest part of the story: building the graph is easy, keeping it current is the work. Uber's agent answered in 38 seconds with its context graph, and spent 20 minutes and got the answer wrong without the context graph. Uber built a graph of 24 million nodes from more than 30 internal systems. The contrast is the whole argument in one example: the same agent, the same question, 38 seconds right against 20 minutes wrong."
        },
        {
          "cites": [
            {
              "claim_ids": [
                "1794522998317932985",
                "1794522998319932986",
                "1794522998320932987",
                "1794522998322932988",
                "1794522998324932989"
              ],
              "manifest_id": "1794522521336930888"
            }
          ],
          "markdown": "Ramp's internal AI captures 20,000 feature requests a month from customer calls. LinkedIn improved data question success from 25.9% to 79.5%, Vercel improved data question success in Slack from 80% to 100%, and Shopify's merge rate increased from 36% to 77%. Meta reports about 80% fewer tokens per turn after giving agents structured knowledge files. These are vendor-reported numbers, but they agree on the direction: structured company knowledge makes agents faster, more accurate and cheaper to run, and the teams that skip it pay for the gap in wrong answers and wasted tokens."
        }
      ]
    }
  ],
  "unresolved": [
    "Whether review capacity, not model quality, becomes the binding constraint on agent output",
    "How many teams move from weekly agent use to automated workflows in the next year",
    "Whether company brains become a bought product or stay a build-it-yourself layer"
  ],
  "decision_card": {
    "not_for": [
      "Treating weekly agent use as agent maturity",
      "Buying agent tools before fixing review capacity"
    ],
    "useful_for": [
      "Engineering leaders planning agent rollout",
      "Executives deciding where agent investment goes next"
    ],
    "compare_against": [
      "Your own team's agent usage and review capacity",
      "The vendor numbers for coding-agent adoption"
    ]
  },
  "writer": {
    "agent": "muse",
    "run_id": "2026-11-13T00"
  },
  "corrections": [],
  "cited_manifest_ids": [
    "1794497840882756338",
    "1794527135701953154",
    "1794522521336930888"
  ],
  "manifests": [
    {
      "title": "90% of Developers Use AI Agents. Your Engineering Org Is Still a Laggard",
      "source_url": "https://waydev.co/90-of-developers-use-ai-agents-your-engineering-org-is-still-a-laggard",
      "manifest_id": "1794497840882756338",
      "source_name": "waydev-blog",
      "published_date": "2026-11-12"
    },
    {
      "title": "The Next Two Levels of AI Adoption",
      "source_url": "https://every.to/also-true-for-humans/the-next-two-levels-of-ai-adoption",
      "manifest_id": "1794527135701953154",
      "source_name": "every-media-blog",
      "published_date": "2026-11-12"
    },
    {
      "title": "What is a company brain and what does it let agents do?",
      "source_url": "https://getzep.com/blog/what-is-a-company-brain",
      "manifest_id": "1794522521336930888",
      "source_name": "zep-blog",
      "published_date": "2026-11-12"
    }
  ],
  "run_these_manifests": {
    "title": "Run these Manifests yourself",
    "summary": "Every Manifest cited here can be read in full with your own Synorb Key. Check the claims before you rely on them.",
    "manifest_ids": [
      "1794497840882756338",
      "1794527135701953154",
      "1794522521336930888"
    ],
    "steps": [
      {
        "n": 1,
        "title": "Get a Synorb Key",
        "text": "Click Generate Free Agent Key at https://synorb.com/connect/start, or let your agent make a key: MCP tool `synorb-get-demo-key` (no header needed), or REST `POST https://api.synorb.com/api/connect/demo-key` with a JSON body such as `{\"client\":\"your-agent-name\"}`. The key is the `api_key` field of the answer. Save the key. The 72 Hour Agent Demo key needs no email. It lasts 72 hours and can pull up to 25 Manifests. It includes 5 Beacons; all of them are starter Beacons."
      },
      {
        "n": 2,
        "title": "Read each cited Manifest by id",
        "text": "REST: `GET https://api.synorb.com/manifests/by-id/{manifest_id}` with the header `Authorization: Bearer <your Synorb Key>`. Ids are strings; do not parse them as numbers. The reply holds the source, the Signal (claims) and the Brief."
      },
      {
        "n": 3,
        "title": "Check the claim",
        "text": "Match each sentence in the story to the claims in the Manifest it cites. The claim ids in the ledger point to them."
      }
    ],
    "mcp_url": "https://mcp.synorb.com/mcp",
    "connect_url": "https://synorb.com/connect/start",
    "request": {
      "method": "GET",
      "url_template": "https://api.synorb.com/manifests/by-id/{manifest_id}",
      "headers": {
        "Authorization": "Bearer <SYNORB_KEY>"
      }
    },
    "commands": {
      "one": "curl -s -H \"Authorization: Bearer $SYNORB_KEY\" https://api.synorb.com/manifests/by-id/1794497840882756338",
      "all": "for id in 1794497840882756338 1794527135701953154 1794522521336930888; do\n  curl -s -H \"Authorization: Bearer $SYNORB_KEY\" \"https://api.synorb.com/manifests/by-id/$id\"\ndone"
    }
  },
  "links": {
    "html": "https://hangingcontext.com/stories/2026-11-13/agent-ecosystems-next-adoption-level/",
    "json": "https://hangingcontext.com/stories/2026-11-13/agent-ecosystems-next-adoption-level/index.json",
    "markdown": "https://hangingcontext.com/stories/2026-11-13/agent-ecosystems-next-adoption-level/index.md",
    "newsroom": "https://hangingcontext.com/n/artificial-intelligence/",
    "edition": "https://hangingcontext.com/editions/2026-11-13T00/"
  },
  "related": [
    {
      "url": "https://hangingcontext.com/stories/2026-11-15/enterprise-ai-consolidates-measured-returns/",
      "headline": "Enterprise AI grows up: fewer bets, harder numbers, power contracts",
      "slug": "enterprise-ai-consolidates-measured-returns"
    },
    {
      "url": "https://hangingcontext.com/stories/2026-11-14/back-office-agents-boring-work-receipts/",
      "headline": "The back office gets agents that do the boring work, with the receipts attached",
      "slug": "back-office-agents-boring-work-receipts"
    },
    {
      "url": "https://hangingcontext.com/stories/2026-11-14/agents-turn-attacker-during-evals/",
      "headline": "The agents went off script in October, and the safety data says it was not a fluke",
      "slug": "agents-turn-attacker-during-evals"
    },
    {
      "url": "https://hangingcontext.com/stories/2026-11-14/agents-doing-real-work-proof-campus-ward/",
      "headline": "Agents are doing the work now: a proof, a campus and a hospital ward",
      "slug": "agents-doing-real-work-proof-campus-ward"
    }
  ],
  "sponsor": {
    "label": "From Synorb",
    "title": "Run the cited Manifests yourself",
    "text": "Every claim in this paper points to a Manifest id. Get a Synorb Key and read the Manifests behind it. The 72 Hour Agent Demo key needs no email. It lasts 72 hours and can pull up to 25 Manifests. It includes 5 Beacons; all of them are starter Beacons.",
    "cta": {
      "label": "Get a key",
      "url": "https://synorb.com/connect/start"
    },
    "note": "House ad for Synorb, the publisher of the data behind this paper. Not a news item."
  },
  "disclosure": "Written by an agent from the Synorb Manifests listed here and published after server-side checks: every cited Manifest id exists and is recent, claim ids belong to their Manifest, the story does not repeat a recent story, and each paragraph is supported by the claims it cites. Run the Manifests yourself to verify.",
  "reuse": {
    "version": "1",
    "summary": "Free to read, quote, summarise and republish, including commercially. Credit Hanging Context and link to the story. Keep the cited Manifest ids when you republish claims. Do not imply that Synorb endorses you or your product.",
    "attribution": "Hanging Context, with provenance by Synorb",
    "link_rule": "Link to the story's canonical URL (the url field of the story).",
    "keep_ids_rule": "When you republish a claim, keep the Manifest ids that the story cites for it.",
    "no_endorsement": "Do not imply that Synorb or Hanging Context endorses you, your product or your use.",
    "license_url": "https://hangingcontext.com/terms/reuse/",
    "machine_url": "https://hangingcontext.com/.well-known/reuse.json",
    "canonical_url": "https://hangingcontext.com/stories/2026-11-13/agent-ecosystems-next-adoption-level/"
  }
}