ChatGPT Work and GPT‑5.6- The Rare Weapon-Grade Strategy Engine Has Arrived
- Rare Writer
- 3 days ago
- 6 min read
The age of AI as an interesting assistant is ending. The age of AI as an operating capability has begun.
ChatGPT Work, powered by GPT‑5.6, is not another chatbot upgrade. It now confirms the shift we, Rare, described in How Far for Stage 2?: from episodic conversation to persistent synthetic operational intelligence—systems that gather context, reason, use tools, act, produce and remain with complex objectives for hours.
That changes the strategic question. It is no longer: Should we let staff use AI?
It is -
“Is the organisation fit to direct, govern and exploit persistent intelligence before a better-prepared competitor does (or already has)?”
Or -
“What if we do nothing or ignore it all? Pass me another huhu grub, please. I’m sure the exponential age will wait.”
Whereas Rare is working on and will release, towards end of 2026, an irrefutable model and index to enable free and self-measurement of the Huhu-Grub Strategic Drift Rate, these are extremely difficult warning signals to dismiss -
88% of organisations now use AI (many untrained), yet agent deployment remains in single digits across almost every business function. Adoption is racing ahead; operating-model readiness is not. Stanford AI Index 2026
Nearly 75% of organisations plan to deploy agentic AI within two years—but only 21% have mature agent-governance arrangements and few have determined use-cases. That is an industrial quantity of ambition resting on a very small control surface of silver-bullet syndrome. Deloitte 2026
Only 30% are redesigning important processes around AI, while 37% are using it superficially without materially changing the underlying work. In other words: new intelligence, same organisational plumbing. Deloitte 2026
Only 25% have moved 40% or more of their AI pilots into production. The remainder risk becoming highly accomplished collectors of pilots, workshops, steering groups and laminated innovation roadmaps. Deloitte 2026
51% of AI-using organisations have already experienced at least one negative consequence; nearly one-third report consequences from AI inaccuracy. Doing AI without diagnostic intelligence is not neutrality—it is accelerated exposure. McKinsey State of AI 2025
86% of employers expect AI and information processing to transform their business by 2030, while 39% of workers’ core skills are expected to change. Waiting is therefore not the conservative or prudent strategy. It is an active wager against the direction of the environment. World Economic Forum
The warning lights are already flashing
Indeed - the numbers are brutal. Stanford’s 2026 AI Index also report that 88% of surveyed organisations now use AI and 70% use generative AI in at least one business function. Generative AI reached 53% adoption in three years—faster than either the personal computer or the internet. Yet agent deployment remains in single digits across almost every function.
That is not safety. It is the starting gun.
Deloitte’s 2026 enterprise study, covering 3,235 leaders across 24 countries, found that 85% of organisations expect to customise agents and nearly three-quarters plan deployment within two years. But only 21% report mature agent governance. Only 25% have moved 40% or more of their AI pilots into production. Just 30% are redesigning key processes around AI, while 37% are using it superficially without materially changing the work.
McKinsey finds the same fault line with further intel- 88% report regular AI use, but only about one-third have begun scaling it. Fifty-one percent of AI-using organisations have experienced a negative consequence, with nearly one-third reporting consequences from inaccuracy.
What has actually arrived
Basically - a synthetic strategic and operating intelligence team working alongside you -
With Rare-Strategy and Digital-Strategy design, ChatGPT Work operates like an on-demand, multidisciplinary intelligence team across the organisation’s authorised environment. Give it an outcome and it can retrieve evidence, connect applications and files, analyse patterns, break work into controlled steps, automate suitable activity and produce finished documents, spreadsheets, presentations, dashboards, reports and interactive Sites.
Through plugins, browser and computer use, it can work across email, Teams, SharePoint, CRM, projects, operational records, local files and knowledge sources. Scheduled Tasks can monitor conditions, respond to triggers and continue working while leaders concentrate on judgement, relationships and decisions. Progress remains visible, work can be redirected and important actions require approval.
Combined with a pivot from legacy, silver bullets or application-paralysis to Rare Data and Intelligence environments, this becomes far more than productivity enhancement. Persistent identity, governed data ingestion, organisational context, modular applications and i‑Glue allow organisations to rapidly resolve, progressively decouple or retire constraining legacy systems—without simply replacing one inflexible monolith with another.
The organisation can then climb a deliberate intelligence staircase -
Transactions → Drivers → Patterns → Forecasts → Interventions → Adaptive Intelligence
This advances organisational function from -
Reactive (or incremental demise) → Preventative → Predictive → Counter-anticipatory → Emergent strategy
Instead of merely recording what happened, the environment explains why it happened, detects what is forming and enables intervention before pressures converge into failure. At counter-anticipatory maturity, the organisation does not simply forecast the next problem—it redesigns the conditions so the problem loses oxygen before it can fully emerge.
Human leaders retain purpose, authority and accountability. Rare provides strategic judgement and architectural control. ChatGPT Work supplies the persistent analytical team capable of investigating, connecting, monitoring and producing alongside them at machine speed.
Leaders can engage with greater [career] confidence—or listen with strategic intent—supported by governed intelligence that reduces uncertainty, strengthens situational awareness and helps de-risk decisions before, during and after critical conversations or decisions.
GPT‑5.6 provides the engine- Sol for the hardest most serious professional reasoning, Terra for balanced everyday work and Luna for high-speed, lower-cost execution. It strengthens long-horizon analysis, tool and computer use, coding, design judgement and production of editable, template-faithful slides, documents and financial models. Sol scored 92.2% on OpenAI’s reported BrowseComp evaluation and 62.6% on OSWorld 2.0. In the wider GPT‑5.6 platform, ultra coordinates four agents by default across parallel workstreams for demanding problems.
This is not merely faster drafting. It is analytical reach, operational persistence and production capacity converging in one controlled workspace. Triple loop strategy, on steriods. Welcome folks, to the exponential age.
Why this is weapon-grade for Rare Strategy
Rare Strategy determines true strategic state - where an organisation stands, where it is drifting, what is exposed and what pathway must now be controlled. GPT‑5.6 Work radically expands Rare's underpinning IP, where the evidence that can be examined, challenged, connected and converted into decision-grade material. The formula, we say, is formidable, but deliberately human-led -
Rare supplies purpose, judgement, pattern recognition and moral authority. GPT‑5.6 supplies immense analytical bandwidth, structured challenge, execution discipline and production capacity.
Applied through the Rare Strategy logic, the capability becomes a strategy engine:
Discover - assemble signals from systems, documents, workflows, vendors, risk records and operational evidence.
Analyse - test fit, detect contradiction, compare options, quantify drift and challenge assumptions at previously uneconomic scale.
Identify - expose the real decision, the defensible pathway and the sequence of moves required.
Protect - surface security, governance, data, dependency and execution risks before automation amplifies them.
Redesign - map human-agent workflows, decision rights, controls and operating models.
Assure - monitor progress, refresh evidence, challenge vendors and keep strategy connected to changing reality.
Rare i‑Glue becomes even more important—not less. It binds people, systems, workflows, data, risks, vendors, governance, agents and outcomes into one executive strategy-control map. Without that connective tissue, AI accelerates fragments. With it, synthetic intelligence can be directed toward coherent organisational purpose.
The cost of waiting is comparative decline
Buying licences is not transformation. Attaching agents to broken processes, poor data, unclear accountability and unmanaged vendors is simply automated confusion at greater speed. The organisations left behind may be those that adopt AI shallowly (at best); scattered copilots, disconnected pilots, no diagnostic baseline, workflow redesign, permission model, assurance or controlled strategic path.
Competitors will use today’s capability to compress research from weeks to hours, scan risk, redesign services, challenge cost, improve decisions and learn faster. Once that advantage compounds, catching up becomes progressively harder.
The answer is diagnostic-intelligent strategy: investigate the real state before prescribing tools; map work before automating it; define human authority before delegating action; establish evidence, controls and stop-points before scale; and measure strategic drift continuously.
ChatGPT Work and GPT‑5.6 do not remove the need for serious strategy. They make serious strategy faster, deeper, more adaptive, more precise, more agile and vastly more executable. The weapon is not AI alone. It is governed synthetic strategic and operative intelligence fused with experienced human judgement.
You do not have to stay stuck. But delay is now a strategic choice—and increasingly, a Rare choice to be left behind.