How we created Telnyx Flow for AI workflow management

At Telnyx, innovation is at the core of what we do. Our journey to creating Telnyx Flow—a low-code, drag-and-drop interface for building and deploying AI-assisted workflows—began with a simple yet profound need: to make our own tools more flexible and user-friendly. This is the story of how we transformed our approach to AI management and created a solution that we hope will empower businesses worldwide.

The challenge: Hard coding AI for customer service

Initially, our team was hard coding a customer service chatbot using our Inference and Storage products. This chatbot was designed to assist users with various queries by leveraging AI to provide accurate, timely responses. The development process was meticulous, requiring our developers to define the chatbot’s business logic, system prompts, and function calls within the service.

While this approach allowed us to tailor the chatbot’s responses to specific needs, it had significant limitations. Every change, no matter how small, required developer intervention. For example, updating the system prompt or adjusting the chatbot’s behavior involved creating new development tickets and extensive back-and-forth between developers and non-technical team members.

A simple change request led to a big improvement

The turning point came when a member of our Sales team noticed an opportunity to enhance our workflow. The chatbot was directing users to email the Sales team for inquiries, but the preferred method was actually to use the contact form on our website. Although this seemed like a minor change, implementing it in our previous system would have required going through the development team to modify a significant amount of hard-coded logic.

This scenario highlighted the benefits of having a more agile and user-friendly solution. We needed a way to allow non-technical users to make changes to our workflows without taking up tons of development time. That’s how Telnyx Flow was born.

Creating Telnyx Flow: From vision to reality

Instead of creating new structures for ordering content and facilitating exceptions, Telnyx Flow allowed us to make the changes we needed in seconds and deploy them instantly to production. This eliminated the need for developer intervention and ticket creation, streamlining the process and improving the user experience.

Telnyx Flow provided a central place for us to control the management of our own workflows, making it easier to create smaller workflows for niche use cases that were part of a wider workflow. We could quickly iterate on existing behavior without numerous deployments to our production servers. Additionally, we could run side-by-side comparisons without rebuilding a chatbot from scratch, and perform quicker similarity search tests to identify gaps in our knowledge bases, enabling teams to focus on improving their documentation.

By leveraging Telnyx Flow, we transformed our approach to AI management, ensuring a more efficient, flexible, and user-friendly system. This proactive adaptation not only solved our immediate needs but also positioned us to better handle future changes and enhancements.

Ultimately, Telnyx Flow was born out of the desire to streamline AI workflow management and reduce dependency on developers. Our development team envisioned a platform that would allow any team member to create, manage, and modify AI workflows with ease. The goal was to democratize access to AI tools, making them accessible to non-technical users while maintaining the sophistication and control required for advanced workflows.

Flow’s low-code, drag-and-drop interface became the cornerstone of this vision. It empowers users from all departments—Sales, Marketing, Customer Service, and more—to build and deploy AI workflows without needing extensive technical knowledge. This user-friendly interface ensures that anyone can contribute to the development and optimization of AI-driven processes.

Leveraging our own products

In developing Telnyx Flow, we leveraged our existing products—Inference and Storage—integrating them into the new platform. This integration ensured that Flow could seamlessly handle the same complex tasks our previous systems managed but with greater flexibility and ease of use.

The true test of Telnyx Flow came when we migrated our own AI Assistant on [telnyx.com] to the new platform. This migration allowed us to experience firsthand the benefits of Flow’s low-code environment. We could quickly implement the change from directing users to email the sales team to using the contact form, demonstrating Flow’s ability to make real-time updates without extensive developer intervention.

The benefits of Telnyx Flow

Once we realized how beneficial our workflow builder was, we knew we had to share it with our customers. Here are some of the benefits we’ve noticed from using Flow in our own processes.

Empowering non-technical users

One of the most significant advantages of Telnyx Flow is its accessibility. By providing a low-code, drag-and-drop interface, Flow enables team members from all departments to create and manage AI workflows. This democratization of technology fosters a more collaborative environment, allowing non-technical users to make sophisticated changes without needing to understand complex code.

Reducing dependency on developers

Flow reduces the reliance on developers for every small change, freeing up valuable developer time for more critical tasks. This change leads to cost savings and accelerates the implementation of updates and new features.

Enhancing responsiveness

With Telnyx Flow, businesses can quickly adapt to changes in user requirements or market conditions. This enhanced responsiveness improves user satisfaction and engagement, as AI assistants can be updated in real time to provide the most accurate, relevant information.

Improving scalability

Telnyx Flow makes it easier to scale AI assistants across different departments or functions. Non-developers can configure and optimize AI tools as needed, ensuring consistent performance and relevance across the organization.

Seamless integration with AI tools

Flow’s ability to integrate with AI tools, such as system prompts and function calls, ensures that responses are accurate and contextually appropriate. This integration allows for easy modifications and real-time updates, maintaining high-quality user interactions.

Real-world application: Telnyx AI Assistant

Our own experience with the Telnyx AI Assistant illustrates the power of Telnyx Flow. The migration to Flow allowed us to implement a simple yet crucial change—directing potential leads to our contact form instead of email—quickly and efficiently. This adjustment, made within an AI Tool Node linked to a Large Language Model (LLM) Completion Node, showcased Flow’s capability to handle complex logic with ease.

By using dynamic variables, we ensured that the AI Assistant provided personalized responses, further enhancing user experience. For example, the response could dynamically include the user’s name, making the interaction feel more personalized and engaging.

What will you build with Telnyx Flow?

Telnyx Flow represents a significant shift in how businesses can manage and optimize AI workflows. By empowering non-technical users, reducing dependency on developers, and enhancing responsiveness, Flow changes the way organizations interact with AI. Our journey to developing Telnyx Flow reflects our commitment to innovation and our belief in the power of accessible, user-friendly technology.

Whether you’re looking to improve customer engagement, streamline internal processes, or gain deeper insights into your operations, Telnyx Flow provides the tools you need to build smarter, more efficient workflows.

FAQ

What is STIR/SHAKEN in Canada? STIR is a protocol that signs caller identity, and SHAKEN is the governance framework carriers use to implement it across Canadian IP voice networks. Together they authenticate caller ID to curb spoofing and restore trust in voice.

How does STIR/SHAKEN work? The originating carrier assigns an attestation level based on its confidence that the caller can use the number and then signs the call with a digital certificate. The terminating carrier verifies the signature and may display a trust indicator or treat unverifiable calls with more scrutiny.

What do attestation levels A, B, and C mean? A means the carrier knows the customer and the number, B means the customer is known but the number could not be fully verified, and C means only the gateway or network of origin is known. Higher attestation can improve trust, but it does not guarantee the call is wanted.

Does STIR/SHAKEN stop all spam and robocalls? It stops caller ID spoofing by authenticating the displayed number, but it does not judge call intent or block unwanted yet lawful campaigns. Many organizations pair authenticated voice with MMS messaging to create alternate trusted touchpoints when calls are screened.

How can I reduce spam calls in Canada? Register on the National Do Not Call List, enable your carrier’s spam blocking features, and use device-level blocking for repeat offenders. If you prefer to shift conversations to text, choosing the right messaging channel types can reduce friction caused by call screening.

*Does 69 last-call return still work in Canada? Yes, last-call return via *69 is still offered by many Canadian carriers, but availability depends on your provider and plan. STIR/SHAKEN does not affect this feature because it authenticates caller identity at the network level rather than enabling call-back services.

Does STIR/SHAKEN apply to SMS or MMS? No, STIR/SHAKEN is designed for IP voice, while SMS and MMS follow different messaging standards and spam controls. Text channels rely on carrier filtering and sender verification rather than voice attestation.