Customer Success Stories
In April 2024, Est Ensemble, an Établissement Public Territorial du Grand Paris grouping together nine towns in Seine Saint Denis and totalling over 440,000 inhabitants, decided to use generative AI for its calls for tenders, in order to automate part of the processing. The aim is to save staff time by simplifying public procurement award procedures and making tender analysis fairer and more objective.
Published: 22/05/2025
To achieve this, Est Ensemble has chosen to call on Inetum, a European leader in digital services and solutions, to develop and operate a turnkey tool that runs on the Azure OpenAI engine. This will help agents analyze files and free up their time to make the right decisions.
The tool was designed around four functionalities, the first two of which were tested in 2024. The solution devised by Inetum automatically verifies a candidate's eligibility and the completeness of his or her file, ensuring that all the required documents - approvals, certifications, declarations, insurance - have been gathered and are still valid.
In the second stage, each file is screened, and the relevance of the answers submitted is assessed exhaustively and objectively against a list of several requirements. The tool automatically compares with the criteria in the technical and administrative specifications. For each requirement, the AI extracts the part of the candidate's bid that meets expectations and proposes a score ranging from 1 to 10 according to a scale set by Est Ensemble.
This saves agents around 30% processing time per file, while avoiding tedious, time-consuming tasks with no added value. The benefits are very tangible.
Another advantage of this solution is that it guarantees applicants equal treatment, since the machine has a constant and faultless level of attention and analysis capacity, from the first file examined to the last.
Thierry Moisson-Bonnevie, Chief Information Office at Est Ensemble, adds: “The development of generative AI is a real opportunity to optimize our internal processes. With Inetum's support, we can offer an efficient public service that is accessible to all.”
These two functionalities herald the next one delivered recently: the solution developed by Inetum can now compare application files with each other and highlight discriminating elements for key requirements. Eventually, a chatbot will be able to dialogue with the tool in natural language, and query it on specific points in applications, independently of the list of requirements.
The introduction of this bid analyzer is not intended to replace the agents, but merely to facilitate their work by enabling them to concentrate on higher value-added missions. A project team made up of a member of the IT department, two people from the purchasing and legal departments, and one from the buildings department handles the largest orders, and can intervene at any stage of the process to correct any errors that are generated. At the same time, three people are working on the Inetum project.
Olivier Serfaty, Director of Inetum's Gen AI Business Unit, concludes: “This initiative represents a real opportunity for Est Ensemble to offer a modernized and more efficient public service to users, agents and stakeholders working with the intercommunal authority. In the long term, this tool could be used by other local authorities as well, to make a wider contribution to the efficiency of public action. So this is just the beginning.
