Every summer, AMC Technology opens its doors to an impressive group of ambitious interns. But 2025 isn’t business as usual. This year, our interns are diving into complex projects involving AI-enabled translation, intelligent assistant tools, and protocol-driven integrations that improve modern contact center capabilities. Their mission is to build practical solutions that improve contact center operations and transform agent-customer interactions.
Project 1: Real-Time Speech-to-Speech Translation
Team: Riley and Theodora
Imagine joining a call in English and communicating with someone speaking French, with both participants hearing real-time translations. That’s the challenge our intern team tackled using Azure Communication Services. Their prototype allows two people speaking different languages to understand each other instantly, with synthesized audio translations on both ends.
Key Features:
- Live language selection and audio synthesis
- Bidirectional speech recognition and translation
- Audio stream customization for synthesized vs. raw voice
The technology directly addresses a rising cost concern for enterprise businesses. While a human translator can charge $0.30 to $0.50 per word, AMC’s AI-driven solution provides speech-to-speech translation at an estimated $0.50 per minute. This represents a significant reduction in cost for high-volume translation use cases, such as healthcare call centers.
Audio routing presented a complex challenge, as Riley explained, citing difficulties with managing multiple inbound and outbound streams simultaneously. “There’s a lot of audio coming in, there’s a lot of audio going out, and just routing that audio to where we want it to be… has been painful, but we’re getting somewhere.”
Theodora added that getting comfortable with Microsoft’s extensive documentation was its own challenge. Despite being early in her academic journey, she chose this project because of its clear practical value: “One person can speak in one language and the other in a different language, and they’ll hear synthesized audio in their own language.”
Future improvements include automatic language detection and integration with Microsoft Teams to streamline workflows in multilingual environments.
Project 2: Agentforce Agent Assist for Salesforce
Team: Lexus and Arnav
Not every business is ready to jump into a full digital overhaul, but they still want access to smart tools. That’s the insight behind Agentforce Agent Assist. This intern team is bridging old and new by integrating AI-powered guidance into Salesforce environments, even those using traditional call systems.
What They Built:
- Custom Agentforce interface in Salesforce using Lightning Web Components
- Simulated Open CTI telephony call states (incoming, connected)
- Contextual dashboards for case lookups and account actions
Lexus explained, “We’re working on just like customer service tools in Salesforce and trying to implement Agentforce with kind of like old tech. So people who are on the old framework can actually use the new tools.”
As with any powerful platform, configuring Salesforce for unique use cases required persistence. The team encountered some unexpected behaviors while building and iterating on workflows, but these moments became valuable learning opportunities. Lexus reflected, “The small wins have been really helpful to give you that little extra assist of like, OK yes, it’s working.”
Their user-centered interface design ensures adaptability for contact centers still relying on older infrastructure, offering a clearer path toward implementing advanced tools like Agentforce.
Project 3: Enabling AI Agents Through MCP Integration
Team: Besufekad and Rishi
While many AI agents can process information, most are unable to directly interact with CRM systems. They lack essential capabilities such as accessing customer records or initiating tasks within Salesforce. This team is tackling that gap by developing the Model Context Protocol (MCP), which empowers large language model agents to access and modify Salesforce data, equipping AI tools with the functional intelligence needed to take meaningful action.
Sprint 1 Successes:
- Local integration with Claude for record lookups
- Remote MCP connection with Agentforce via Apex APIs
- Real-time authentication and session management
“Our project has a lot to do with MCP servers,” said Besufekad. “It hooks to large language models like ChatGPT and Claude and helps give those AI chatbots context to other external data sources. In this case, it was Salesforce and Agentforce.”
Unlike the more straightforward chatbot integrations documented online, their use case was uncharted territory. “Nobody has really done what we’re doing to the extent that we’re trying to,” he continued. Despite the steep learning curve, the team achieved autonomous data access within Salesforce, laying the groundwork for future automation based on incoming call metadata.
Rishi recalled the beginning of the project as confusing: “It took a lot of research and a couple of meetings with the mentors to understand the project.” Once over the initial hurdle, the team started implementing features with increasing confidence.
Key Takeaways
- AI is immediately relevant: The interns demonstrated that technologies like MCP and real-time translation can be applied today in contact centers.
- Mentorship facilitates growth: Timely guidance helped interns overcome technical and process-related obstacles.
- Adaptability is critical: Working with legacy systems and novel protocols required creativity and flexibility, which the interns consistently displayed.
- Quantitative impact matters: The translation assistant’s pricing model offers compelling cost benefits over traditional services, underscoring AI’s efficiency.
The AMC Technology 2025 intern cohort didn’t just dip their toes into emerging tech—they jumped in and started building it. And the exciting part? This is only their first sprint.
Each team is already planning their next steps. The Speech-to-Speech group is working on automatic language detection, voice customization, and integrating with Microsoft Teams. Agentforce is fine-tuning dynamic workflows while staying flexible with shifting systems. And MCP? They’re gearing up for more advanced work, like capturing call signals and expanding their toolsets.
Their projects brought measurable improvements and laid the groundwork for innovations that can elevate contact centers across industries. This year’s internship experience is already showing that meaningful progress is possible when talented individuals are empowered to create.
