João Leitão's Homepage

João Leitão's Homepage : Research

SMART-CRM Research Project

Project Name: SMART-CRM: Autonomous and Intelligent CRM Systems for Public Administration
Portuguese Title: Sistemas CRM Autónomos e Inteligentes para a Administração Pública
Funding Institution: Fundação para a Ciência e Tecnologia (FCT)
Funding Programme: Inteligência Artificial, Ciência dos Dados e Cibersegurança de relevância na Administração Pública
Project Reference: 2024.07664.IACDC [SciPROJ] [DOI]
Principal Investigator: João Carlos Antunes Leitão
Leading Institution:
NOVA.ID.FCT - Associação para a Inovação e Desenvolvimento da FCT (NOVA.ID.FCT/FCT/UNL)
Research Unit:
NOVA LINCS - NOVA Laboratory for Computer Science and Informatics
NOVA SST - NOVA School of Science and Technology (Faculdade de Ciências e Tecnologia - Universidade NOVA de Lisboa)
Project Duration: 01-03-2025 to 31-01-2026
Total Funding: 124.997,81 €

Project Abstract

Customer relationship management (CRM) systems are essential in many contexts to ensure that an adequate service can be provided to customers while also improving those services and tracking their activities. While in many companies these CRM systems are essential to sustain their economic growth, by improving, tracking, and predicting the needs of clients and providing services in an efficient and trackable way, in the context of public administration, they have the potential to play a pivotal role in bridging the gap between the general public and organisations, significantly impacting the quality of life in modern societies. Public sector CRM systems can help streamline interactions, provide more personalised services, and improve the efficiency of public service delivery.

Since these systems handle multiple aspects of interactions, among a potentially high number of users, they tend to be complex and require a significant amount of expertise to operate correctly, which hinders their penetration somewhat. In this project we plan to combine knowledge from recent advances in security, self-management, and machine learning to make these systems more autonomous, easier to operate, and ensure their correct operation. In particular we will take advantage of existing anonymised data sets and recent advances in machine learning to train models, using realistic emulation of these systems in the lab to achieve a set of complementary goals:

Self-management of the CRM system: This will allow to automatically control the number of different instances of CRM components in execution at each time, handling the tension between quality of service and operational costs under dynamic workloads. Fine tuning of different components in an automatic way will be a significant innovation both to these systems and to micro service architectures in general.

Automatically manage security aspects of CRM systems: Monitoring the request patterns of users this will allow to automatically manage user accounts (e.g., temporary disabling) and control firewalls to thwart attacks and protect sensitive user data. Such efforts will leverage on advances in intrusion detection systems that can actively learn new attack patterns.

Automatic reclassification of requests: By analysing user requests we will detect errors in classifications, and correct and redirect requests accordingly, using operator feedback to adjust such corrections when they are inaccurate. This will allow the system to become more accessible and improve the response times of organisations.

Through this integrative approach with a strong foundation on applied research, we aim at taking advantage of a micro-service architecture, to develop open source components that can enrich a commercial CRM system with these new functionalities, specifically targeting their role and usability in the context of public administration, while at the same time providing tools that can be leveraged by other sectors. This project benefits from collaborations with Câmara de Almada and Oeiras Municipality.

Research Team

Name	Institution	Role
João Leitão	NOVA SST and NOVA LINCS	Principal Investigator
João Leite	NOVA SST and NOVA LINCS	Researcher
José Júlio Alferes	NOVA SST and NOVA LINCS	Researcher
João Costa Seco	NOVA SST and NOVA LINCS	Researcher
Kevin Gallagher	NOVA SST and NOVA LINCS	Researcher
João Pedro Bordalo	NOVA SST and NOVA LINCS	PhD Student
Pedro Camponês	NOVA SST and NOVA LINCS	PhD Student
Diogo Jesus	NOVA LINCS	Software Engineer
Juliana Monteiro	NOVA FCT	Research Management
Rafael Mira	NOVA SST and NOVA LINCS	Master Student
Andrés Bonilla Quiroz	Politecnico di Milano / NOVA FCT	Master Student
João Brilha	NOVA SST and NOVA LINCS	BI Licenciado
Diogo Almeida	NOVA SST and NOVA LINCS	BI Licenciado
Diogo Fona	NOVA SST and NOVA LINCS	BI Licenciado
Fábio Silva	NOVA SST and NOVA LINCS	BI Licenciado
Lourenço Vieira	NOVA SST and NOVA LINCS	Bolsa de Iniciação à Investigação
Bernardo Gracioso	NOVA SST and NOVA LINCS	Bolsa de Iniciação à Investigação
António Mendes	NOVA SST and NOVA LINCS	Bolsa de Iniciação à Investigação

Publications

International Conferences

Goose: Optimistic Search in the IPFS Network.
Rafael Sequeira, Pedro Camponês, Pedro Ákos Costa, and João Leitão.
Proceedings of the 44th International Symposium on Reliable Distributed Systems (SRDS 2025), 2025.
[Publisher Website]
This work addresses the performance and resilience of search in decentralized peer-to-peer systems, with implications for the distributed infrastructure underpinning CRM system components and their content discovery mechanisms.

National Conferences

Um Protocolo Descentralizado para Transmissão Contínua de Dados Robusta e Eficiente.
Diogo Fona and João Leitão.
Actas do décimo quinto Simpósio de Informática (INForum 2025), 2025.
[PDF]
Proposes decentralised data streaming techniques relevant to the real-time telemetry and continuous data dissemination capabilities required by the CRM system's self-management and monitoring mechanisms.

Technical Reports

Uma Ferramenta para a Deteção e Teste de Vulnerabilidades em Arquiteturas de Micro-serviços.
António Neves Mendes, Kevin Gallagher, and João Leitão.
Technical Report. NOVA LINCS, 2025.
[PDF]
Directly addresses security testing of microservice architectures, providing automated vulnerability detection tools for REST APIs — a core concern for the SMART-CRM security management layer (Task 4).

Theses

Andrés Bonilla Quiroz
Politecnico di Milano (co-advised with NOVA School of Science and Technology).
MSc Thesis (2025): Improving the Resilience and Flexibility of Microservice Architecture
Proposes a broker-mediated communication substrate for microservices with adaptive self-scaling across multi-region deployments, directly aligned with Task 3 (self-management of CRM microservice lifecycle) and Task 6 (microservice architecture design).

Diogo Almeida
Faculdade de Ciências e Tecnologia, Universidade NOVA de Lisboa.
MSc Thesis (2025): Byzantine Fault Tolerance in Large Scale Edge Replicated Systems
Investigates Byzantine fault tolerance mechanisms for large-scale edge replicated systems, contributing techniques for ensuring the integrity and resilience of distributed architectures that underpin the CRM system's operation in untrusted environments (Task 4).

Diogo Fona
Faculdade de Ciências e Tecnologia, Universidade NOVA de Lisboa.
MSc Thesis (2025): "Large-Scale Decentralised Application-Level Streaming
Develops decentralised streaming protocols for large-scale data dissemination, relevant to the real-time telemetry and continuous monitoring capabilities required by the CRM system's self-management layer (Task 3). Note: the thesis PDF is pending an update to include the project acknowledgement.

Rafael Mira
Faculdade de Ciências e Tecnologia, Universidade NOVA de Lisboa.
MSc Thesis (in preparation, expected 2026): Autonomic Management of Microservice-Based Web Applications
Directly tackles Task 5 (automatic reclassification of requests and workflow management), developing LLM-based agents for autonomous classification and routing of CRM user requests.

Prototypes and Software

SMART-CRM Demo Prototype

A working prototype of the SMART-CRM system is available for demonstration purposes. The prototype integrates the base CRM system with the autonomic modules being developed in the project. The prototype can be accessed at https://smartcrm.di.fct.unl.pt/.

Source code for the developed components will be made publicly available after the publication of the main research papers associated with the project. All results of the project will be released as open source after the publications are accepted.

Collaborations

The project benefits from active collaborations with public administration organisations that inform the requirements and validate the solutions being developed:

Câmara Municipal de Almada — The municipality of Almada is collaborating with the project to explore the use of the SMART-CRM platform for conducting citizen inquiries and automating the processing of responses. A prototype for this specific use case is currently under active development.

NOVA FCT (Faculdade de Ciências e Tecnologia) — As a public institution with its own client base (students), NOVA FCT serves as an additional requirements source and potential pilot deployment site.

This work was supported by National Funds through the Portuguese funding agency, FCT — Fundação para a Ciência e a Tecnologia, through project 2024.07664.IACDC (SMART-CRM).