Oncoclínicas&Co delivers fast and connected service integrations with IBM watsonx
Scaling personalized communication is a challenge for any large healthcare network, and Oncoclínicas&Co, Latin America’s largest oncology-focused group, was no exception.
Operating more than 140 units and serving hundreds of thousands of patients annually, the organization faced growing complexity in managing frequent, time-sensitive interactions. Patients had come to expect the immediacy and personal attention of in-clinic care, but digital channels didn’t deliver the same experience. Fragmented systems and manual processes created delays and high disconnection rates, making it harder to maintain the seamless, empathetic service patients expected.
To preserve the human touch and their reputation for accessibility, Oncoclínicas needed to unify communications, automate routine tasks and enable faster, more reliable connections—without adding operational complexity.
“At Oncoclínicas&Co, our mission goes beyond technical treatment. I deeply believe that every patient deserves not only clinical excellence but also humanized, agile and attentive care,” says Paula Soares, CX Manager at Oncoclínicas&Co.
Oncoclínicas engaged with IBM and IBM Business Partner PROA.AI—a Brazilian leader in conversational AI and intelligent automation—to redesign their digital engagement model and deliver fast and consistent support across every channel. With the help of PROA.AI, Oncoclínicas integrated generative AI (gen AI) and automation capabilities into their digital channels. PROA.AI also helped them consolidate the IBM watsonx® portfolio of products and the IBM Cloud Pak® technology stack under a unified governance framework.
The solution leveraged IBM® watsonx.ai® studio for natural language processing and gen AI. The implementation also included IBM Cloud Pak for Business Automation with Robotic Process Automation (RPA) for workflow automation, and IBM Cloud Pak for Integration to unify WhatsApp, chat and email under a single, secured framework. Oncoclínicas also implemented a governance approach to help address regulatory requirements such as Brazil’s LGPD and to reinforce trust.
The solution’s architecture enabled real-time responses, automated scheduling and proactive updates, which helped streamline processes and improve engagement at scale. Beyond efficiency gains, the initiative helped restore time and confidence to interactions, supporting better patient engagement through technology. “Thanks to our team’s effort and the implementation of innovative solutions, we optimized processes and ensured patients feel supported at every step,” adds Soares.
The solution delivered measurable impact across every interaction for Oncoclínicas.
Delivered in partnership with IBM and PROA.AI, this transformation shows how AI and automation can scale accessibility and responsiveness, bringing the immediacy and reassurance of in-person interactions to every digital touchpoint. “Thanks to the dedication of our team and the implementation of innovative solutions, we were able not only to optimize processes but also to ensure that our patients feel truly welcomed and cared for at every stage of their treatment,” concludes Soares.
Founded in 2010, Oncoclínicas&Co is Latin America’s largest integrated oncology network, operating more than 140 units across 40 cities in Brazil. The company performs nearly 700,000 treatments annually and is recognized for combining clinical excellence with innovation and patient-centered care.
Founded in 2016 in São Paulo, PROA.AI is a Brazilian leader in conversational AI and intelligent automation. Their solutions for customer service and process orchestration are powered by IBM technology, delivering secured, scalable and humanized digital experiences.
© Copyright IBM Corporation 2025. IBM, the IBM logo, IBM Cloud Pak, IBM watsonx, and watsonx.ai are trademarks of IBM Corp., registered in many jurisdictions worldwide.
Examples presented as illustrative only. Actual results will vary based on client configurations and conditions and, therefore, generally expected results cannot be provided.