By Dr. Aram Seddigh ·
Between an AI investment and a measurable return sits a description of the work itself. MIT Technology Review’s article on the missing step between hype and profit points to weak real-world evidence for economic returns, to the question of whether tools fit the workflows they enter, and to the need to redesign the work itself. This article explains why workflow fit is a workplace strategy step, and how to write it into the decision base before area, layout and lease commitments are decided.
What the source establishes, and where our reading begins
The source is about the economics of AI in working life. It sets out three signals: real-world evidence of economic returns remains thin, value depends on how well a tool fits the workflow it enters, and the work itself needs to be redesigned.
It does not make claims about offices, attendance or space demand. Everything from the next heading onward is our interpretation for workplace strategy practice.
Workflow fit rests on a description of work — and that is familiar ground
Whether a tool fits a workflow can only be judged against a description of the sequence it enters: who does what, in what order, with which inputs, and under which conditions. That description is the same material a workplace analysis works with when it maps the distribution of tasks across the workplace.
So the missing step is a mapping task for a workplace team, expressed in a vocabulary the team already uses. The workplace strategy framework exists precisely to hold that description steady while decisions are made around it.
Thin evidence changes what a decision base can claim
Thin evidence means the effect is not yet documented well enough to carry a fifteen-year commitment on its own. The effect may well be there. It is the documentation that is missing.
That has a direct consequence for the decision base. An assumption about future task volume, headcount or output belongs in the plan as an assumption, written out and dated, with a named point where it is revisited. Observed practice — what people do this quarter, measured — sits in a different column. An assumption that is documented, dated and tied to a review point can be tested against evidence when the evidence arrives.
The sequence: map needs, then decide area and lease
The order matters. Needs are mapped given the organisation’s conditions and its chosen ways of working, and only then do decisions about area, layout and agreements follow.
Applied to AI, that means the redesign of work comes before the workplace conclusion. If a workflow is genuinely rebuilt, the needs it places on the workplace can change with it: more sustained concentrated work in some roles, more joint problem-solving in others, different patterns of client contact. Those needs are what a workplace has to meet. Reading them straight off a productivity forecast skips the mapping.
Redesign is a design task before it is a technology task
Redesigning work means deciding which steps remain human judgement, which become review, which disappear, and where the handovers sit. That is design work, and it produces a specification: what needs the changed workflow generates, for whom, and how often.
Workplace strategy then answers a concrete question. Which of these needs is the current workplace already meeting, and which are not yet supported? Occupancy and satisfaction measurement, described in our overview of workplace insight tools, tells you where the work is done today. The redesigned workflow tells you what tomorrow asks for. A useful decision base holds both.
What stronger practice looks like
Teams that handle this well tend to do four things.
They write the workflow before they discuss the tool, in enough detail that a colleague outside the project can follow it. They keep expectation and measurement in separate columns of the same document. They set a review date on every AI-related assumption that touches area or working patterns. And they tie workplace conclusions to needs that have been mapped, with the forecast named as input.
This is what makes an AI-driven case auditable a year later.
Three questions to answer in the next strategy cycle
Use these as decision gates.
- Which workflows are we changing, and can we describe them step by step as they run today?
- What has been observed in practice, and what is still expectation about how the work will run?
- Which needs does the changed way of working place on the workplace, and how are those needs met — before area and lease are decided?
If your team can answer all three in writing, the AI assumption in the business case is visible and testable. If one is still open, you know where the next piece of analysis belongs. This is also where the role boundary becomes practical: see what a workplace strategist does when technology decisions and property decisions arrive in the same quarter.
A repeatable step, not a one-off exercise
The description that workflow fit is judged against needs maintenance. Tools change, roles change, and the description has to be kept current alongside them.
That makes it a capability question. Give the step a place in the method — the mapping template, the assumption log, the review point — so that keeping the description current becomes part of ordinary work. Write the workflow down before the next investment case is drafted, and the workplace conclusions that follow will rest on something you can check.
Source: MIT Technology Review, "The missing step between hype and profit", technologyreview.com (retrieved 29 April 2026).
Next step
build the workflow-fit step into your practice
If AI assumptions are entering your workplace decisions, the useful work is turning them into a mapped decision base your team can defend. Get in touch to talk through how workflow fit can be tested before the area decision. Get in touch if you want the step adapted to your own strategy cycle and reporting format.
FAQ
What is the missing step between AI hype and profit?
In the source, it is the step between deploying AI and seeing economic returns: whether the tool fits the workflow it enters, and whether the work has been redesigned around it. Real-world evidence of returns is still thin, which is why the step is worth naming explicitly.
Why is workflow fit a workplace strategy question?
Because judging workflow fit requires a description of how tasks are distributed and performed, and that description is what workplace needs are derived from. When a workflow is genuinely rebuilt, the needs it places on the workplace can change with it.
How should AI assumptions be handled in a space decision?
Keep them as labelled assumptions with a review date, separate from what has been measured in practice. That way the case stays readable when evidence arrives, and area and lease decisions can be revisited against something specific.
What capability does this require from a team?
The ability to describe a workflow step by step, to distinguish observation from expectation in a decision base, and to sequence the work so needs are mapped before commitments are made. It is a method that can be trained and reused.