← Research
Research

Big picture thinking in the age of AI

The constraint that used to force a choice has gone, and what replaced it converges.

Last reviewed: 26 September 2026 · Next review due: 26 September 2027

This territory is full of frameworks and empty of argument about what artificial intelligence does to it. The altitude tradition predates AI entirely and the recent work on AI and strategy never mentions altitude. This page is the join, and the disagreement.

Question this page answersAll 996 questions this research covers

Every altitude is now instantly available, and what arrives at each one is the median view. The frameworks in this territory were written for a world where reaching another level took effort, so they diagnose a failure that is no longer the one people have.

Share this line as a card

The answer, in one line

Two things. Reaching any level is now effortless, which removes the constraint that used to force a choice between them, so the failure moves from being unable to reach a level to committing to none.

Share as a card

The standard account, stated fairly#

Big picture thinking is described in the management literature as the capacity to move between levels of detail: zoom in and zoom out, fifty feet and fifty thousand, the dance floor and the balcony. The failure named in every version is fixity. A leader gets stuck at one level, usually the operational one, and cannot see the pattern. The remedy follows from the diagnosis and is always the same: exposure, rotation, deliberate practice at the level being avoided. Woodward puts the incidence of it at around 70 per cent of senior executives, without a study behind the figure. The frameworks are set out and graded at models of big picture thinking.

None of them mentions artificial intelligence, because all of them predate it. The recent work that does address AI and strategy, in turn, never mentions altitude, zoom, balcony or levels. The two literatures do not touch, and the join is where the argument is.

What changed, first: the constraint has gone#

Reaching another altitude used to cost something. Getting to the pattern meant analysis and time. Getting to the twenty-year view meant a scenario exercise with a budget and a sponsor. That cost is what forced a choice, and the choice is what people experienced as strategic thinking.

Dashboards now render the detail continuously, forecasts produce the horizon on request, and a model will generate the pattern in seconds. Nobody is prevented from reaching any level. What has disappeared is anything that makes them settle on one, and the resulting failure is not fixity. It is altitude lock: thorough at every level, committed to none, the level of the work set by whatever output arrived last. It reads as rigour, so it is not caught.

What changed, second: the view converges#

This is the part the abundance argument misses. Romasanta, Thomas and Levina reported in March 2026 that leading models consistently recommend strategies that match current managerial trends and vocabulary rather than with the logic of the situation put to them. They named it trendslop, and their recommendation is to use these systems to expand options rather than to make choices. The wider pattern of homogenised reasoning and perspective is treated at does AI make everyone think alike.

Put the two together and the conclusion is sharper than information abundance. What is instantly available at every altitude is the median view rather than the view: the same fashionable answer at every level, and produced for your competitors at the same moment it is produced for you. A strategy any competitor's model would also have generated is not a strategy; it is a position everybody is about to hold.

Where this disagrees with the current management literature#

The 2026 work from business schools and consultancies converges on a reasonable position: analysis has become abundant, so the bottleneck moves to selection and commitment. That is correct and it is incomplete, because it treats the abundant analysis as neutral raw material. If the material converges, selection is not merely the scarce step in a familiar process. It is the only remaining source of a position that is not generic, and choosing well from a homogenised option set is a different problem from choosing well from a varied one.

It also disagrees with the altitude tradition, more politely. Exposure and rotation are the right remedy for a leader who cannot reach the other level. They are no remedy at all for a leader who reaches all three every morning before the first meeting.

What follows in practice#

What would change my mind#

Evidence that model-generated strategic advice diverges rather than converges once the situation is described in enough detail, which would make the trendslop finding an artefact of thin prompting rather than a property of the systems. Or a measured finding that leaders working with instant access to every level do in fact settle on one, which would make altitude lock a worry rather than a pattern. Neither exists at the moment in either direction, and this page is an argument rather than a finding.

What this does not establish#

No study has measured altitude lock, and no instrument distinguishes it from thoroughness. The trendslop article discloses no sample, measures or figures on the page read, so nothing here rests on a rate. The claim that convergence removes competitive advantage is an inference from what strategy is for rather than a measured effect.

Key sources

Evidence review · SS-2026-354 · Graded against the published rubric

Cite this page

Hirji, R. (2026). Big picture thinking in the age of AI. The SuperSkills evidence base, SS-2026-354. https://thesuperskills.com/research/big-picture-thinking-in-the-age-of-ai. Last reviewed 26 September 2026.

An evidence review by Rahim Hirji, not peer-reviewed research. For a material claim, cite the underlying study as well; every study here carries its own permanent link.

How citations and IDs work
Questions answered on this page

What does AI change about big picture thinking?

Two things. Reaching any level is now effortless, which removes the constraint that used to force a choice between them, so the failure moves from being unable to reach a level to committing to none. And what arrives at each level converges: researchers reported in March 2026 that leading models recommend strategies that match current managerial vocabulary rather than with the situation, which they named trendslop.

Do the existing frameworks still work?

They describe a different failure. Kanter's trained incapacity, Woodward's altitude sickness and Heifetz's dance floor all name somebody who cannot get to another level, and the remedy for that is exposure and practice. Neither applies to somebody who can reach every level instantly and does.

Is this just information overload?

No. Overload is about volume and the remedy is filtering. This is about the absence of a forcing constraint, and about what fills the space: not more views but the same view produced at every level. Filtering does not help with convergence.

What is the practical answer?

Choose the level before the work starts, write down which one and why, and deliberately ignore what the other levels would have shown. The three filters of the Altitude Lens each carry a test for that purpose, and each was written to be answerable in one sentence.

In this hub

Definitions

The terms this field uses, defined against their primary sources.

Ask the evidence
What does the evidence actually show?What should our board be asking about this?Where does Rahim disagree with the consensus?
Bring this into your organisation

If this describes something happening in your teams, say so.

Keynotes, board sessions and advisory work, drawing on research across more than 200 organisations in 30 countries. Tell me the room, the date and the shift you need. A reply within 24 hours.

Start a conversation

Topics and audiences  ·  All research

Box of Amazing

Rahim’s free weekly letter on AI and human capability

If this was useful, the weekly letter is where the thinking happens first. Most of what ends up on this site starts there. Weekly essays on AI, capability and the future of work. Read by 25,000 people, every week since 2017. Free, and one click to stop.

Opens Substack to confirm. No pitch in it, unsubscribe in one click, and nobody follows up because you read something.

Running an event, or responsible for how AI arrives in your organisation? Keynotes  ·  Advisory  ·  Boards  ·  Enquire