Many questions. Even more data. AI agents connecting it all into answers.
Transport questions almost never live inside one dataset. To answer something like “where should Nairobi build a new corridor?” you need traffic counts, satellite imagery, road networks, census data, climate exposure, and the policy documents that already shape the city. They live in different places, in different formats, maintained by different organisations. None of them were built to talk to each other.
TGIS·AI does that connecting work for you. You ask a question; it goes and gets the relevant pieces from wherever they live, lines them up so they actually compare, and assembles an answer right there — with maps, tables, and charts generated on the spot from the data it just pulled.
The kinds of data it pulls together
Most tools handle one shape of data. This one handles them side by side, in the same answer:
Numbers
Statistics from World Bank, OECD, transport ministries, the Asian Transport Outlook, and similar sources. Indicators by country, year, mode.
Maps & places
Road networks, ports, rail lines, administrative boundaries, satellite imagery, flood and climate exposure layers.
Documents
Policy commitments, project evaluations, research papers, NDC pledges — searchable as evidence alongside the numbers.
Live feeds
Operational data where it’s available: port throughput, freight movements, fuel prices — pulled at the moment you ask.
What happens when you press enter
- Discover
It finds the data sources that could actually answer your question — and shows you which ones it picked, and which ones it ruled out.
- Analyze
It pulls the pieces it needs, lines up units and geographies so they’re comparable, and runs the analysis. Where two sources disagree, it shows both rather than picking silently.
- Synthesize
It composes the answer: a short narrative, the map or chart that actually supports it, and the underlying tables. Every number stays linked back to where it came from.
What you get back
Not a wall of text. The output is whatever the question calls for:
- A map drawn from the actual layers — corridors, exposure zones, accessibility surfaces — not a screenshot of one.
- Tables and charts built on the fly from the data it pulled, with the columns and breakdowns the question implies.
- A written summary in plain language, with every figure traceable to the source it came from.
- A record of the work — which sources were used, which were skipped, where the data is thin or out of date. You can audit the answer, not just read it.
About this demo
This is a walkthrough, not the live system. The data shown is real in shape but pre-prepared for two illustrative questions (Nairobi accessibility, Ukraine freight corridors) so you can see how the interface flows end to end. The wider TGIS·AI work is in build — see tgis.ai/concept for the full picture.
