Skip to content

Execution intelligence

Execution intelligence is the work of establishing whether a company can do what its plan says, finding where it will get stuck, and testing what to change, with the evidence behind every conclusion. Its unit of analysis is the fit between a plan and the organisation asked to carry it.

Managers report that the commitments a plan runs through across units hold all the time in fewer than one case in ten.

Why it matters when the plan changes

Finance, market and strategy are modelled before a major decision. The organisation expected to deliver the plan is usually understood afterwards, from the outcome. That asymmetry is the gap the term names: work that treats the organisation as something you can read, forecast and test before you commit, as a financial model is read before capital moves. That gap is measurable: fewer than one in ten managers report cross-unit commitments as reliable all the time, per Sull, Homkes and Sull.

The tension is that the name is useful and the ownership claim is not. Naming a category is how an idea becomes discussable, and a second company using the words is a sign the idea is taking, not a threat. Claiming to have created or to own it would be a claim without evidence, which collides with the standard the work sets: pairing human judgement with a model is not automatically better than either alone, so process matters as much as the model.

In practice

A board approves a market entry with a financial model defended line by line and a single slide on organisational readiness. Six months later the entry is behind, and the reason is an unowned interface between two functions. The financial question had a discipline behind it. The organisational question had an opinion.

Evidence

What it cannot tell you

Execution intelligence tells you whether an organisation, as it stands, can carry a defined plan; it does not judge whether the plan itself is the right one, and it cannot guarantee an outcome. It is bounded to a stated horizon and a stated arrangement, so a change to either requires the reading to be redone rather than trusted forward.

How Atlas reads it

Atlas names and defines execution intelligence in public, and does not claim to have created it or to own it. In practice the category is the plain question underneath: can this organisation deliver this plan, what may get in the way, and what should be addressed first. The category label belongs in articles rather than in an opening line.

Questions

The fit between a defined plan and the organisation asked to carry it. That means the plan itself, the roles and decision rights around it, the dependencies it runs through, observed operating patterns, and behavioural evidence read against the demand rather than in the abstract.

People analytics starts with the employee and describes the workforce. Execution intelligence starts with the plan and asks whether this arrangement can carry it. The difference decides what counts as a finding: a workforce pattern is only interesting here if the plan depends on it.

No. Atlas names and defines it and builds in public, and makes no claim to have created it or to hold market recognition for it. A second company using the words is evidence the idea is taking hold, which is useful at this stage rather than threatening.

Before commitment, when the cost of being wrong is still recoverable, and again whenever the plan changes materially. The stakes are real: Sull, Homkes and Charles Sull found in Harvard Business Review (2015) that fewer than one in ten managers trust cross-unit commitments fully, the exact layer a plan depends on.

It does not claim to predict a person's future, rank people, or replace the accountable leader. Vaccaro, Almaatouq and Malone (2024) found human-model combinations can perform significantly worse than either alone, which is why every claim carries its confidence, source and counter-evidence rather than a single automated score.