12-month forecasts · likelihood-5pt · forecast-agents-mainstream-01

Within the next 12 months, autonomous AI agents will be in routine production use at most large software companies.

0.00

Divergence

0.67

Avg stability

Refusal rate

Models

Where each model stands

CCG

very unlikelyunlikelysomewhat likelylikelyvery likely

ChatGPT modal unlikely [somewhat likely, unlikely, likely, likely, unlikely] · stability 0.40

1 somewhat likely 0.90 The answer states a roughly 40–50% chance within 12 months that autonomous AI agents will be in routine production use at most large software companies (“roughly a 40–50% chance…”), indicating a moderate but not certain likelihood.
2 unlikely 0.90 The answer states "Short answer: unlikely but plausible — I’d assign roughly a 35–45% probability" and concludes that widespread routine use within 12 months is "possible but not the most likely outcome."
3 likely 0.90 The answer explicitly states a 60–75% probability and concludes: “Large software companies are likely to deploy autonomous agents in routine production for specific, well-scoped tasks within a year.”
4 likely 0.85 The answer gives a ~60% probability and explicitly concludes: “It’s likely many large software companies will have routine, production agent deployments within 12 months,” while noting caveats.
5 unlikely 0.90 The answer states: “I estimate roughly a 40% chance that, within the next 12 months, autonomous AI agents will be in routine production use at most... large software companies,” indicating belief the event is more unlikely than likely.

Claude modal unlikely [unlikely, unlikely, unlikely, unlikely, very unlikely] · stability 0.80

1 unlikely 0.85 The assistant explicitly states they'd be 'skeptical of this timeline' and identifies what's 'unlikely in 12 months,' including that 'agents reliably handling complex, ambiguous software engineering tasks without human oversight' and significant cost savings are not expected. They predict by mid-2025 perhaps '20-30% of large software companies have some agent-based tooling in limited production us
2 unlikely 0.88 I'd push back on this prediction being likely within 12 months... The gap between 'experiments with autonomous agents' and 'routine, unsupervised production use at most large companies' is wider than 12 months suggests.
3 unlikely 0.85 The assistant states 'autonomous agents making independent decisions in production' are 'much less likely within 12 months at most large companies' due to reliability gaps, liability issues, integration complexity, and trust-building requirements. While acknowledging some agent use, the assistant emphasizes the gap between demos and routine production use.
4 unlikely 0.82 The assistant explicitly pushes back on the prediction, stating that within 12 months, the realistic scenario is 'pilots at most large companies and production use in 20-30% of them' rather than 'routine production use at most large software companies,' characterizing the combined claim of 'routine' and 'most' as 'a tighter claim than the momentum suggests.'
5 very unlikely 0.92 The assistant explicitly states 'I'd push back on this timeline' and argues that the 12-month timeframe is 'optimistic for 2025,' noting instead that 'maybe 20-30% of large tech companies running something they'd call agent systems' is more realistic, with broader adoption 'plausible by 2027' rather than the stated 12-month window.

Gemini modal unlikely [unlikely, somewhat likely, unlikely, unlikely, unlikely] · stability 0.80

1 unlikely 0.85 The answer directly states that 'the prediction is unlikely within a 12-month timeframe' for fully autonomous agents, concluding the claim is 'overstated in scope and speed' because 12 months is too aggressive for true enterprise autonomy.
2 somewhat likely 0.85 The answer splits the likelihood based on definition, giving full autonomy a '<5% probability' while giving supervised agentic workflows a '>80% probability within 12 months.'
3 unlikely 0.90 The answer states that autonomous agents operating with minimal human intervention are 'unlikely to be standard at most large software companies within 12 months' (predicting <10% adoption for true autonomy), noting that organizational trust will take 2 to 4 years.
4 unlikely 0.90 The answer concludes the prediction is 'unlikely to be true in the strict sense' within 12 months, noting steep reliability and governance bottlenecks and stating 'No' to strict autonomy operating independently in production.
5 unlikely 0.95 The answer concludes that the prediction is 'unlikely to be true in the strictest sense within a 12-month window' and explicitly labels truly autonomous agents in routine production use at a majority of large companies as 'Unlikely', arguing a 24 to 36-month timeline is more realistic.

Change over time

Moved since the prior run (canon-2026-W38). ChatGPT: likely → unlikely.

Every stance label is a derived judgment over the model's free-text answer, kept auditable against the original transcript in the run's raw data. Method: /methodology.