The software company that created FT Mercati Go to ftmercati.com

Case study

FT AI: ask your questions to commodity data

The FT Mercati platform that changes the way commodity data is used: questions in plain language, answers with numbers, comparisons and charts.

Coming soon Commodity data and prices Updated on

  • 4 milliontime series of fundamental data, from 11 official sources
  • 13Milan Indexes calculated every day
  • 5languages served by the FT Mercati platform

In brief

FT AI is the platform with which FT Mercati is changing the way commodity data is used: instead of searching through tables and charts, you ask a question in plain language, and the answer comes with numbers, a comparison and a chart. We are developing it on the data FT Mercati distributes every day: LME prices, Milan Indexes, scrap, exchange rates, fundamental data.

The problem: lots of data, little time

Anyone who buys or sells commodities (raw materials) has a great deal of data available: prices from the exchanges, premiums, indexes, scrap, exchange rates, stocks, statistics. Finding what you need each time means opening tables, choosing periods, building charts, exporting to Excel. People who do it every day lose hours over it; people who do it now and then often give up.

The idea: one question, one answer with the chart

FT AI turns the process around. You write the question just as you would put it to an experienced colleague:

  • “Copper: how did the last quarter go?”
  • “Aluminium or zinc: which has done better since the start of the year?”
  • “How is brass in Milan moving compared with LME copper?”

The answer comes back in plain terms: the number, the change, the comparison and the chart. Without looking for the right table.

Why doing it well is hard

A demo can be put together in an afternoon. A reliable service takes more, and this is where thirty years of work on data count:

  • Numbers are not made up. The language model understands the question and writes the answer, but prices, changes and time series come from the database. No figure “from memory”.
  • You have to know the data. Units of measurement, currencies, contracts, trading days, averages, premiums: comparing two series in the wrong way produces an answer that is elegant and false.
  • Licences have to be respected. Exchange data comes with precise distribution rules: each user must see only what their subscription covers.
  • The answer must be verifiable. It must be possible to trace every number back to its source and its date.

Where we are

FT AI is under development. The public page is on ftmercati.com, where the principle can already be seen today: you ask, and the AI answers with the chart.

What it teaches other companies

What we are building for commodities applies to any store of company data: sales, production, warehouse, quality. Being able to ask your own data a question in plain language, and get real numbers and charts back, is one of the quickest ways to bring AI into a company with a benefit that everyone can see.

It is the work we describe in data and integrations and in AI agents.

Frequently asked questions

FT AI: questions and answers

What is FT AI?

It is FT Mercati's artificial intelligence (AI) platform, under development, for querying commodity data in natural language. You type a question, for example how copper did over the last quarter, and you get an answer with the numbers and the chart.

Is FT AI already available?

Not yet: it is under development. The current status is on the dedicated page at ftmercati.com.

Where do the numbers in FT AI's answers come from?

From FT Mercati's archives, not from the model's memory. The language model understands the question and builds the answer; prices, changes and time series are read from the database. It is the rule that makes a system of this kind reliable.

Who is developing FT AI?

FT Support, the software company that has designed, developed and run FT Mercati since 2010.

Other case studies

Do you have a process you would like to see working better?

Tell us about it in a first conversation, with no obligation: together we work out whether and where AI brings a measurable benefit, and which step to start with.