A forecast is a range, not a number
Any forecast presented as a single certain figure is being misrepresented. How to read one, and what to ask for.
Read the analysisZENKAIQ insights
Clear positions on enterprise intelligence.
Short, evidence-led analysis for leaders trying to explain the whole business, measure AI value, govern metrics, and use prediction honestly.
Published deliberately
New analysis is released every three days. Cluster D and all Deltek-specific material are on hold.
Your forecast should have to beat a naive baseline
Last period repeated is free and often surprisingly good. If a model cannot beat it, the model is not adding value — and most forecasts are never tested against it.
Read the analysisSource of truth and system of record are not the same thing
One is where data is created. The other is where you have agreed to look. Conflating them is why reconciliation never ends.
Read the analysisWhy two teams get different numbers for the same thing
It is almost never a data quality problem. Four causes, in the order you should check them.
Read the analysisIntegrated data is not an integrated explanation
Two tables can join technically and remain incomparable operationally. Why moving everything into one platform does not produce a whole-business view.
Read the analysisWhat a certified metric actually means
Five attributes, one named owner. Without the owner, a metric is a suggestion — and disputes about it have no resolution path.
Read the analysisThe full lifecycle cost of an AI initiative
Licence cost is the smallest line. Six categories most business cases omit, and why the omission changes the conclusion.
Read the analysisNatural language interfaces need governed metrics more, not less
Asking a question in plain English against an ungoverned semantic model returns a confident wrong answer faster than a dashboard would. Why AI-BI raises the stakes on definitions.
Read the analysisAI activity is not AI value
Licences, logins and prompt counts measure adoption. Value is a measured change in throughput, cycle time, quality, cost or capacity. How to tell them apart.
Read the analysisDecision latency is an operating cost
A management question that takes two weeks to answer has a cost, even though no one books it to an account. How to size it.
Read the analysisWhen a local win is an enterprise loss
A function can improve its KPI and make the company worse. The metric is not wrong — the boundary around it is too small.
Read the analysisReporting every part is not explaining the whole
A company can report finance, workforce, delivery and sales accurately and still be unable to explain its own performance. The gap is structural, not technical.
Read the analysisWhy does margin fall when revenue grows?
Revenue up, headcount up, delivery up — and margin down. The answer usually sits between the functions, not inside any one of them. A worked example.
Read the analysis