Mission Research and Strategic Intelligence

Mission research deals in maps of a reality too complex to measure perfectly.

Joshua Project publishes detailed information about people groups, languages, Christian presence, Scripture status, and population. The precision is useful. Joshua Project also warns that some percentages are estimates, displayed zeros have several meanings, and field reality can differ from database values.¹

Artificial intelligence is unusually useful in such an environment because it can combine large quantities of heterogeneous information.

It is also dangerous because it can make uncertain information look more certain.

Mission Data Are Maps

Terms such as unreached people group depend on definitions. Joshua Project uses specific Christian-adherent and evangelical thresholds within its methodology. Other organizations use different counting systems and categories.

These frameworks are strategic tools, not divinely revealed taxonomies. The biblical language of peoples and nations should not be collapsed into one modern database schema.

A dashboard may display an evangelical percentage to two decimal places. The underlying estimate may come from older reports, partial census data, denominational information, or field judgment.

AI can preserve the decimal while losing the epistemic status. A strong mission-intelligence system should therefore expose uncertainty: source age, disagreement, missing values, confidence, and need for local confirmation.

Different mission datasets answer different questions. People-group counts vary according to whether groups are counted within national borders, how ethnolinguistic boundaries are defined, and which Christian-presence thresholds are used.

AI can reconcile labels and highlight differences, but it should not average incompatible methodologies into one seemingly more accurate number. Sometimes the correct output is a table of disagreement.

What AI Can Do With the Data

AI can reconcile datasets, summarize field reports, flag contradictions, detect missing values, compare languages, identify migration patterns, and generate questions for researchers.

These are substantial gains. The system should assist analysis, not become a missionary oracle. Mission geography changes as people move. Hundreds of millions live outside their country of birth. A language community historically associated with one region may now be present in European, Gulf, North American, or Asian cities.

AI can help connect migration, language, church location, demography, and Scripture access. A congregation may discover that a community it imagines as distant now lives nearby.

Serious demographic work combines censuses, surveys, estimates, and indirect procedures because complete global data does not exist.

That is not a failure. It is a reason for humility. A model can analyze incomplete data faster. It cannot make missing information stop being missing.

Mission intelligence is not only statistics. Field reports contain qualitative knowledge about openness, conflict, leadership, language, migration, and church health.

AI can summarize hundreds of reports and identify themes. Sensitive reports require careful data governance, and synthesis should preserve minority observations rather than flattening everything into majority patterns.

Ranking produces an appearance of objectivity. Any ranking requires weights. Should population matter more than Scripture access? Should existing local church strength reduce or increase priority? How should persecution affect the score?

These are strategic and theological judgments, not neutral mathematical facts. AI can calculate once humans decide the model. It should not hide the values inside the ranking.

Local Validation Changes the Picture

Global data should return to local churches and field workers. A database may show no church where a fellowship began last year. It may show meaningful Christian presence where nominal affiliation tells little about discipleship. It may classify two groups separately where migration has blurred the boundary.

Local knowledge is therefore a formal validation layer, not an anecdotal afterthought. The final mission-data pipeline is: Data → Analysis → Local Validation → Discernment → Decision With local validation feeding corrections back into the data. Mission conditions change. Churches are planted. Migration shifts populations. Governments change policy. Scripture translation progresses.

AI systems built on cached datasets should expose the date of the underlying information. Current-seeming language can hide old data.

Figure 24.1. Mission Data Decision Pipeline

DATA	→	ANALYSIS	→	LOCAL VALIDATION	→	DISCERNMENT	→	DECISION

Local validation can correct both the data and the assumptions behind it.

Strategy Still Requires Discernment

A model can rank populations according to need indicators. It cannot turn a ranking into missionary vocation.

Sending also involves church partnership, gifting, security, invitation, history, opportunity, and discernment. Data can expose neglected need. It should not become an algorithmic command. The strongest mission-intelligence use of AI is improved sight: helping the Church notice underserved languages, stale assumptions, contradictory data, diaspora opportunities, and areas where resources remain scarce.

Seeing more clearly matters. Mission still happens among people. Good mission research helps churches see realities they would otherwise miss. It should not create a technocratic elite whose models overrule local Christians.

The strongest research relationship is reciprocal: global analysis informs local discernment, and local knowledge corrects global analysis.

PART V

Footnotes

1. Joshua Project, “People Groups: Counts.”; Joshua Project, Joshua Project Overview 2026 ↩