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

Frequently asked questions

Frequently asked questions on AI in companies

What people ask us most often about artificial intelligence (AI) in companies, projects, costs and security. If your question is missing, ask us.

FT Support and FT Mercati

What is FT Support's AI consulting?

It is the work we do with the people who run the company to decide where artificial intelligence should deliver a result, in what order to tackle the processes, under what rules and with what budget: a first conversation with no obligation, a day with senior management on the process map and priorities, and a step-by-step plan with the cost and expected result of each step. The plan belongs to the company, which can carry it out with us or with others.

What does FT Support do?

FT Support brings artificial intelligence (AI) into companies in Italy and across Europe: it helps the people who run the company to define the adoption strategy and builds what is needed, namely AI agents inside the ERP (the company's management software) and the CRM, assistants that answer on company know-how, and custom software. It is the company that created FT Mercati; its team has over thirty years of experience in software development and in managing financial and commodity data feeds.

What is the relationship between FT Support and FT Mercati?

They are the same company. FT Support srl, founded in 2008, is the software company that since 2010 has designed, developed and run FT Mercati, the commodity prices and data platform used every day by over a thousand companies. FT Mercati provides the data; FT Support puts it to work inside companies' processes, with software and artificial intelligence.

Where do you start when bringing AI into a company?

From a specific process, not from the technology. You choose a repetitive, measurable task, check which data is available, build a fixed-scope pilot and measure the result. The first step with us is a no-obligation conversation to identify the right process.

Does my company's data stay confidential?

Yes. We use the models through business APIs, under contract terms that exclude the use of your data to train the models, and processing takes place in the Microsoft Azure cloud, in a European region. Databases and documents can stay on the company's servers, the AI respects the same permissions as the users, and every action is logged.

Do you only work with companies in the commodities sector?

No. Adoption strategy, AI agents, assistants on company know-how and custom software are useful to any company with repetitive processes and scattered information. In manufacturing and in the world of commodities we have an extra advantage: we know the sector at first hand and we can connect the systems to FT Mercati data.

Where do you work, and in which languages?

We are based in Italy and work with companies in Italy and across Europe. Meetings are held by video call (Teams, Zoom or Google Meet), in English or Italian; proposals, plans and all written material are in English.

Projects, timescales and costs

How can artificial intelligence be used in a company?

Above all where skilled people search for, copy, check and re-enter information: preparing quotes, reading orders and documents in PDF, filling in customer and supplier records, answering recurring questions about procedures and price lists, checking contracts and invoices, summarising a customer's history. These are frequent, measurable tasks, which can be entrusted one at a time to an AI assistant or agent, leaving the decisions to people.

How long does a first artificial intelligence project take?

A fixed-scope pilot on a single process generally takes a few weeks. The move into production and the later extensions are planned in versions, with scope and cost defined each time.

How are the costs set?

Fixed scope and a price defined before we start, with no quotes by the hour. Each phase, that is, analysis, pilot, production and maintenance, has its own cost and its own expected result, so you can decide step by step whether to continue.

Are there incentives for software and artificial intelligence?

There are national and European measures for digitalisation, whose requirements and deadlines change often. We do not handle the applications ourselves: we provide the technical documentation for the project and recommend getting help from a consultant who specialises in grants and incentives.

What does the company need to provide to get started?

A process to improve, a point of contact who knows it well and can give it a few hours a week, and access to the necessary data, even read-only. No in-house technical skills are needed.

Technology and security

Which is the best artificial intelligence for companies?

There is no single best one: it depends on the job to be done. Four criteria matter: quality on the specific task, the terms on data (in a plan for organisations, your data must not be used to train the models), the ability to connect it to the company's systems, and the cost at the expected volume. That is why it pays to have an architecture in which the model can be replaced without rebuilding the system.

Which artificial intelligence models do you use?

Mainly Anthropic's Claude models, chosen for their reliability in reasoning and in the use of tools. The architecture stays independent of the provider: where needed we use other models, and switching from one to another does not mean rewriting the system.

Can AI act on its own in our systems?

Only within limits decided together. In our projects the agent has a closed list of tools, those that change data ask for confirmation, records are created as drafts and irreversible operations stay with people.

Can the system run on our servers?

Yes. Software, databases and document indexes can be installed on the company's servers or in the cloud. The choice depends on confidentiality, costs and continuity of service, and is made at the start of the project.

Do you use artificial intelligence for development too?

Yes, and we say so: we develop with AI-assisted programming tools, under the responsibility of our developers, with code review, automatic security checks and acceptance testing together with the customer. It is one of the reasons why delivery times have become shorter.

Do you comply with the AI Act?

We design our systems with the European regulation in mind: people in the decision loop, transparency about the use of AI, an action log, instructions for users. We do not develop systems classified as high-risk without a specific assessment shared with the customer.

Is your question not here?

Call us: answering is the best way we have of finding out which other questions to add to this page.