Shipping Exceptional AI Support: Inside Comm100's Workflow

Shipping Exceptional AI Support: Inside Comm100's Workflow

About Comm100

Comm100 is a leading omnichannel customer engagement platform that helps organizations enhance customer loyalty through faster, more effective service. It serves businesses across education, government, and commercial sectors with a comprehensive suite including Live Chat, Ticketing, and AI-powered chatbots.

As one of the early pioneers in live chat technology, Comm100 has evolved to lead the charge in AI-powered customer conversations. Their platform enables businesses to deliver personalized customer experiences at scale, combining intelligent automation with human expertise to resolve customer inquiries efficiently across multiple channels.

Comm100's AI-Powered Customer Experience

Comm100's AI-first approach powers multiple customer experience innovations across their platform:

  • AI Agent: Their AI Agent handles routine inquiries, bookings, and transactions across live chat, SMS, email, and messaging platforms, enabling 24/7 customer support while reducing live human agent load.
  • AI Copilot: The AI copilot analyzes incoming queries in real-time, surfacing suggested replies and solutions to human agents directly within their console, improving accuracy, reducing handle time, and enabling faster onboarding.
  • Predictive Insights & Analytics: AI-powered dashboards surface key trends, risky conversations, and intent signals–helping leaders proactively optimize team performance and customer outcomes.

Together, these GenAI capabilities allow Comm100 to deliver faster, smarter, and more scalable customer experiences – while equipping teams with the tools they need to drive continuous improvement.

Challenges

As Comm100 expanded their AI capabilities to handle increasingly sophisticated customer interactions, maintaining quality and reliability across these diverse AI workflows became a critical challenge.

Their existing processes presented several key areas for optimization:

  • Teams released on extensive manual processes to test the accuracy and reliability of AI responses across customer scenarios.
  • Prompt experimentation and refinement needed engineering support for each iteration, creating dependencies.
  • Debugging AI interactions when responses didn't meet quality standards involved extensive manual investigation to trace issues back to specific components.
  • Limited Visibility into how their AI models performed across the platform, made it difficult to optimize performance systematically.
  • Collaboration between product and engineering teams on AI improvements needed more visibility and speed.
  • Scaling quality assurance as they added more AI capabilities meant their existing manual testing approach wasn't sustainable.

These challenges slowed their development process and made it harder to rapidly iterate on new AI features that could enhance customer experiences.

Accelerating AI Workflow Iteration with Maxim AI

To keep up with rapidly evolving customer expectations, the Comm100 team needed a way to rapidly iterate, validate ideas, and scale new workflows across all channels, without slowing down development.Using Maxim, Comm100 has built a structured process that lets their team experiment with AI behavior, evaluate outcomes, and deploy better customer support agents–faster:

Structure Prompt Experimentation

Comm100 uses Maxim’s Prompt Playground as a structured workspace to test and refine AI behavior in a low-friction environment, enabling fast iteration without deep engineering involvement.

  • Prompt Versioning: Teams track changes and manage variants across use cases, ensuring consistent quality across channels.
  • Context-aware debugging: Unexpected behavior is quickly isolated and fixed in the Playground–removing guesswork from experimentation.
  • Organized Prompt Folders: Prompts are structured and managed within Maxim’s Prompt CMS, using shared folders that make it easy for product and AI teams to collaborate, test ideas, and maintain visibility across experiments.

Dataset-Based Evaluation

Comm100 maintains datasets within Maxim that mirror real support conversations and use cases from across Comm100’s channels – chat, email, and voice.

  • Scenario-Based Testing: Teams test prompts against real-world scenarios to surface edge cases early.
  • Ongoing Refinement: These datasets are continuously updated to capture evolving customer behavior and emerging intents.
  • Reusable test sets make regression testing consistent and repeatable.

Custom Evaluators for Domain Specific Checks

Going beyond off the shelf evals, Comm100 leverages custom evaluators inside Maxim tailored to their own measures of quality

  • Evaluator Prototyping: Teams design and test evaluators directly within the Playground using sample outputs.
  • LLM-as-a-Judge: The team defines custom evaluators with domain-specific instructions, scoring scales, and clear pass/fail criteria.

No-Code Workflow Prototyping

Beyond single prompt testing, Maxim’s no-code builder helps Comm100 model and test multi-step agent logic without backend dependencies.

  • Prompt Chains: Logic is modeled visually, enabling rapid iterations.
  • Step-by-Step Debugging: Each step is validated before deployment, reducing breakpoints and guesswork. 
  • PMs can own and test full agent flows end-to-end.

Maxim helps Comm100 rapidly test prompts and workflows, evaluate them with custom metrics, and ship smarter agents – all in one place.

Impact 

Maxim has helped Comm100 transform their AI development workflow – reducing reliance on engineering, accelerating feedback loops, and enabling faster deployment of high-quality support experiences.

With the no-code builder and prompt Playground, product managers can now prototype and validate complex agent workflows end-to-end, surpassing engineering support. This has transformed how ideas move from experimentation to execution.

“Before using Maxim, I had to create scripts for prompt testing and evaluation, which usually took 1–2 days to build and integrate with our existing AI Agent. Now, with Maxim, it takes less than an hour to build the prompt, the integration works seamlessly, and it also provides downloadable testing reports without any coding–saving me a lot of time.”

– Karas Shi, Senior Product Manager, Comm100

The ability to test prompts against real-world datasets and run evaluations directly in-platform gives the team confidence in the quality and consistency of their AI agents across channels. Custom evaluators ensure each iteration meets internal benchmarks for tone, accuracy, and helpfulness.

“Maxim has substantially boosted my productivity, especially with its no-code prompt chaining. As a product manager, I can rapidly prototype, test, and refine new features without relying on engineering resources.”

–  Karas Shi, Senior Product Manager, Comm100

By bringing experimentation, evaluation, and workflow design into one tool, Comm100 is able to move faster while keeping quality at the core of their AI support platform.

Conclusion

For AI teams building customer-facing applications, Comm100's experience demonstrates the value of investing in robust evaluation infrastructure early in the development process.

Their advice to other AI engineering teams: "Embrace tools that streamline your workflow and lower the barrier to experimentation – rapid iteration with minimal coding makes a big difference. Focus on solutions that allow easy testing, debugging, and prompt refinement, as this will accelerate development and improve the reliability of your AI applications."

As Comm100 continues to enhance their AI-powered customer engagement platform, Maxim plays a key role in enabling them to innovate much faster while ensuring that they deliver experiences that their users love. By investing in setting foundational evals infrastructure, Comm100 is positioned to deliver increasingly sophisticated AI experiences that enhance customer satisfaction and agent productivity.

At Maxim we are obsessed with helping AI teams move faster on their product roadmap while keeping user trust at the core of their product. If you need help with your evals infrastructure, let's connect!

Learn more about Maxim here: https://www.getmaxim.ai/
Learn more about Comm100 here: https://www.comm100.com/