Identify what was counted and what it was divided by before publishing an AI-supplied statistic. A large percentage can describe a small, selected group rather than the population your Indian readers assume it represents.
Create a statistic interpretation card
For each number, record:
- Exact claim: the proposed sentence, including qualifiers.
- Numerator: the count or quantity of interest.
- Denominator: all cases against which it was measured.
- Population: who could be included and who was excluded.
- Time window: observation dates and comparison period.
- Units: people, organisations, transactions, rupees or another measure.
- Baseline: the starting value and its definition.
- Collection method: how observations entered the dataset.
- Source location: the accessible passage, figure or methodological note.
Mark missing fields as unresolved rather than asking an AI system to supply plausible values. A publication date is not automatically the measurement period.
Distinguish respondents from the wider market
Consider a hypothetical survey sent to 1,000 businesses. Two hundred respond, and 120 respondents report using an AI tool.
The result is 60% of respondents, not a demonstrated 60% of all contacted businesses or all Indian businesses. The 800 non-respondents have unknown status. Reporting 120 out of 1,000 as the market adoption rate would also substitute an unsupported assumption about those missing responses.
A bounded sentence is: “In this hypothetical survey, 120 of 200 responding businesses reported AI use; the response group may not represent the wider market.” The arithmetic and the representativeness question must remain separate.
Audit the comparison language
Suppose, hypothetically, a measured rate rises from 10% to 15%. That is an increase of five percentage points and a relative increase of 50%. Those descriptions are compatible but answer different questions.
For money figures, record the currency and whether amounts refer to a month, year, transaction or user. Ask whether the compared figures use the same cost definition. Do not quietly convert a foreign estimate into an Indian budget forecast or imply tax treatment that the source does not establish.
Decide whether the number earns publication
Publish when the denominator, period and population support the intended sentence. Narrow the wording when only a subgroup is supported. Stop if the missing denominator changes the interpretation; request the underlying definition or omit the figure.
If exploring the AI tools catalog, check access format and save terms before buying research access. Marketplace account, code or token offers are not proof of official subscription equivalence, nor do they validate statistics.
Sources and scope
Source material checked 29 October 2026; features/policies may change. The supplied Perplexity citations reference narrowly supports inspecting linked source material. Direct fetching was blocked; no statistical-checking feature is claimed. The interpretation card and numerical examples are editorial guidance.