MedContact Center achieves 58.2% effective management with AI Agents and increases customer service capacity by 22% in six months

Managing higher volume without losing agility or control

MedContact faced the following challenges:

• High volume of requests and the need for fast, consistent service: The voice channel received thousands of calls per month, and scheduling represented a particularly high share: 73.7% of interactions involved medical appointment management. MedContact needed to provide fast, consistent service, even during peak demand, without making service quality depend solely on expanding the team.

• Processes with multiple business rules: Resolving a call in the healthcare sector usually involves much more than answering a question. It can mean confirming an appointment, freeing up an appointment slot, rescheduling a test, guiding a member, or starting an authorization that requires the correct documentation in order to move forward. Every possible scenario had to be considered.

• Redefining agent team roles: Technology had to support the team, not replace it. The goal was for advisors to work with greater focus, more context, and less repetitive workload, while maintaining their role in the moments where they were still needed.

• Proactive management limited by cost: Confirming appointments proactively was not financially feasible, but the lack of proactive management caused no-shows and unexpected rescheduling, which made schedule management more complex.

Joint work between MedContact and Inconcert

From the outset, the project was approached as a joint effort between MedContact and Inconcert. The priority was to fully understand the process, define which tasks could gain speed with AI agents, and clearly establish when the specialized team needed to intervene.

From there, both companies worked with the goal of advancing wherever technology could resolve or prepare the interaction, and routing the case with context whenever it required validation or human judgment.

1. Analysis and definition of MedContact processes

Before incorporating Inagent into the voice channel, MedContact and Inconcert dedicated an initial phase to deeply understanding each process. It was essential to review how requests were managed and define where technology could provide the greatest support.

This preliminary work made it possible to organize the processes in which MedContact intervened as a BPO from the ground up:

• The main types of interactions were mapped.

• The most frequent scenarios were identified.

• Conversation flows were defined.

• Clear routing rules to the team were established.

• The points where human validation was essential were detected.

This analysis was especially important because the process could not be approached as something closed or static. In healthcare, needs change and new tasks arise that require constant adjustments in order to function well.

That is why MedContact and Inconcert worked from the beginning with a continuous improvement approach: reviewing, personalizing, adjusting, and optimizing the process based on what happened in real conversations.

In this way, Inagent was integrated with clearly defined guardrails. The AI agents could move forward with tasks such as identifying the intent of the call, collecting initial information, validating basic data, or routing standard requests. When the case required a more specific review, mandatory confirmation, or the judgment of an advisor, the conversation was transferred to the specialized team with the necessary context to continue the process without starting from scratch.

This preliminary work made it possible to define the scope of Inagent and the processes it would manage. Ultimately, the AI agents were trained for the following use cases:

• Medical appointment scheduling.

• Appointment rescheduling.

• Appointment cancellation.

• Patient care, resolving general questions.

• Routing to the right human advisor for each inquiry.

• Authorization management.

• Collections.

2. Inagent implementation

With the processes already defined, Inagent was incorporated into the voice channel with the mission of helping each call progress as much as possible from the first contact, within the limits defined by MedContact Center and MedPlus Salud.

Inagent acts in two ways:

1. Inagent handles the early stages of the conversation: It identifies the reason for the call, collects the necessary information, performs initial validations, and manages standard requests, such as frequently asked questions or simple scheduling. This allows many interactions to be resolved more quickly and ensures that members receive a consistent response, even during peak demand.  

2. When the request requires more than a standard response, the process changes naturally: If a specific validation, mandatory confirmation, business exception, or situation requiring an advisor’s judgment arises, Inagent routes the conversation to the specialized team. It does so with the context already collected, so the person continuing the interaction does not have to start from zero.  

This last point was especially important for MedContact, because the incorporation of Inagent was not framed as a replacement for the human team. On the contrary: the goal was for advisors to focus on the tasks that truly required human experience, while the AI agent helped organize, filter, and streamline the most repetitive requests.

3. Continuous Improvement of Inagent’s AI Agents

After launch, MedContact Center and Inconcert maintain ongoing analysis and adjustment work so that Inagent continues to gain effectiveness based on real conversations. Each call provides useful information to understand what is working well, where friction appears, and what changes can help make the process more agile for members.

This monitoring makes it possible to identify new opportunities and adjust flows as clear patterns emerge. Based on this analysis, the team identifies where it is useful to intervene in order to improve the process: which transfers can be avoided, which rules need to be refined, which validations should be reviewed, and which conversations require clearer guidance from Inagent.

Instead of treating routing as an error, the analysis makes it possible to distinguish when a transfer is necessary and when the previous journey can be improved.

Results: more capacity, better experience, and measurable return

During the period analyzed, MedContact Center managed to handle higher volume with the same resources, increase service levels, and make progress in key processes such as collections.

1. More efficient operations

One of the project’s most relevant data points is the 58.2% effective management achieved by Inagent in the voice channel. To measure this, MedContact Center considers two types of interactions:

• Those that the AI agent resolves from beginning to end.

• Those in which Inagent assists with management, collects information, validates initial data, and routes the case to the specialized team when human intervention is needed.

This view provides a better understanding of the project’s real value. In healthcare, not all calls can be closed autonomously, because some processes depend on specific validations. In these cases, if the call arrives filtered, contextualized, and with part of the work already completed, the process becomes more agile.

The main operational results were:

• 44,554 calls effectively managed by Inagent during the period analyzed.

• +22% increase in the total volume of calls handled by Inagent and the human team.

• Same resource structure, with no need to expand the team to absorb the higher volume.

2. Better patient experience

The operational improvement was also reflected in the patient experience. The main advances were:

• -39% in abandonment rate, which means more patients were able to move forward with their request without unnecessary waits.

• +5 percentage points in service level, reflecting an operation with greater service capacity.

• Reduction in negative sentiment from 18.9% to 5.8% in interactions managed by Inagent.

• 87.6% resolution in interactions with positive sentiment, showing the relationship between a more fluid conversation and successful management.

In a process such as medical appointment scheduling, reducing friction helps users move forward more clearly and allows the team to work with better-prepared conversations.

3. Business impact and cost efficiency

Inagent’s impact was also reflected in indicators directly linked to the business: time savings, better use of resources, positive ROI, and progress in finance-related processes such as collections.

The most relevant results were:

• 108,774 minutes saved in the voice channel during the period analyzed.

• Positive ROI of 8.8%, resulting from the savings generated and process optimization.

• +5 percentage points improvement in collections recovery in the portion managed by Inagent.

These results show that AI agents contribute to improving indicators directly tied to business performance.

AI Agents applied to the right process, with clear limits and continuous improvement

The MedContact case shows that AI agents are especially useful when applied to a well-defined process, with clear rules and a realistic view of what each party should resolve.

Thanks to this combination of AI agents, process design, and continuous improvement, MedContact has optimized request management without putting service quality at risk. Interactions move forward more quickly, members experience a smoother journey, and specialized teams can focus on cases that require validation or human support.

The joint work between MedContact Center and Inconcert also allows the process to continue evolving. Based on the analysis of real conversations, both teams adjust flows and identify opportunities to make service increasingly efficient, consistent, and close.