Consip, 60% faster tender times with AI

With the massive use of Artificial Intelligence, Consip aims to reduce the average time needed to manage procurements from seven and a half to three months. But this acceleration is only the most immediate aspect of a wide-ranging programme, which in addition to cutting the timeframe wants to change all the company's activities: with a two-stage, tight plan that aims to make the Treasury's public procurement company a sort of hub for artificial intelligence in the public administration, first in Consip and then in the other administrations. All this without AI ending up reducing jobs, which are indeed expected to grow from 457 at present to 490 by the end of the year.

Consip's journey along the path of artificial intelligence, after all, starts with the people. Who are currently receiving invitations to take part in a maxi training programme, comprising some 4,000 hours this year alone and aimed at 100% of staff. In the project developed by the company led by CEO Marco Reggiani, artificial intelligence is not conceived as a lever reserved for a restricted group of 'technological' skills, but as a working companion for everyone. An 'anthropomorphic' companion, because the AI destined to enter every office is obviously the agentic one (and it seems that someone has already decided on the first name to be given to their virtual colleague). "We will be the first to bring the systematic use of artificial intelligence into the life cycle of public contracts in order to strengthen timeliness, quality and transparency in public procurement," Reggiani summarises, emphasising the investment "in skills and new professionalism".

On this basis, the 'intelligent agents' will be deployed in 38 activity strands, which run along the entire lifecycle of public contracts, from the gestation based on the analysis of the needs of the various PAs up to the finish line of contract execution, which must be thoroughly monitored because it is there that the real quality of public spending is decided.
Along this route, Consip's AI will, for example, take care of tendering strategies, examining case by case what is the best set-up of lots and contract parameters, offer clarifications requested by potential suppliers a bit like what happens now in the customer care departments of some large companies, and support the analysis of the technical and economic offers underlying the tenders. Here we encounter another of the exponential accelerations made possible by AI, which in the experiments of recent months has shown that it can compress comparative evaluations that today take weeks into a few minutes, without damaging the quality of the results. Because the key point is always in the definition of the relationship between technology and human choices: which is not the dystopian one imagined by some, whereby AI replaces man, but is an integration between technology that multiplies the capacity to analyse data and information and the human component that makes decisions on that basis. This is starting to happen in the financial administration, for example, which entrusts AI with the profiling of the risk shown by different taxpayers but leaves it to officials and managers to decide where to direct assessments.

In short, the aim is to ask this technology to do what it does best: to automate routine and repetitive activities and optimise the management of a boundless wealth of information such as that of Consip, which in over 25 years of activity has signed more than 10 million contracts, catalogued 11 million items and published more than 5 thousand tenders, receiving more than 14 thousand offers. In the first four months of this year alone, according to the latest updated monitoring, 80 tender lots worth 9 billion were tendered, and public expenditure also brokered through the digital marketplace reached 11.2 billion.
And the same is trying to happen in many Italian public administrations, which here and there are increasingly adopting AI systems. Consip intends to give this spontaneous blossoming a structured character with two moves expected in the second half of the year: in the autumn the initiative for the first framework agreement for the supply of AI solutions for public administrations will start, while in the last quarter the Consip master course on public procurement organised with the Milan Polytechnic will start.