Communicate with external databases or services
API actions are the primary method used to communicate with external databases or services from our AI system. Configuring these actions may require involvement from your company’s IT staff, but our team is always available to assist during the setup process.
There are two types of API actions:
Actions with a response – These involve direct interaction with the user. For example, checking the availability of a product or appointment. The AI can provide a response based on the API’s return data.
Silent actions – These perform background operations without generating a visible reply. Examples include adding customer data to a CRM, updating a mailing list, or triggering internal workflows.
Each action can be standalone or part of a more complex flow. For instance, in a booking flow, you might configure one action to check availability and a second to save the reservation.
It is essential to instruct the AI clearly on when and how to use a specific API. Instructions should be clear and concise, starting with phrases like:
“If the customer asks…”
“When the customer provides…”
Since API execution depends heavily on the AI’s understanding of context and instructions, functionality cannot always be guaranteed, especially in ambiguous or unexpected situations.
Our AI can connect to almost any external service, giving you the power to automate tasks, retrieve important information, and provide fast, accurate responses to your customers.
When setting up a connection, you can configure how the AI interacts with the service and what it does with the information it receives.
For example, the AI can respond directly to a user— confirming an appointment, checking product availability, or showing order status — or perform tasks quietly in the background, such as updating customer records, sending notifications, or syncing information across systems.
You also have full control over how the AI communicates securely with your services. Many APIs require some form of authentication, like a secret key or token, which acts as a “password” so only your AI can access the data. You can configure this securely, without exposing sensitive information.
The AI can send information to your systems when needed. For example, it can post form data, update a record, or trigger an action in another platform. You decide what data to send, when to send it, and how it should be formatted — so the AI can automatically handle tasks that would normally require manual input.
You can also define specific rules and triggers for when the AI should use a connection. It can act only when a customer provides certain information or asks a particular question. You can guide the AI on how to interpret responses, ensuring it delivers clear, useful answers every time.
All of this happens securely, so your AI can interact with your systems without exposing private or sensitive data. This flexibility allows you to create tailored, intelligent workflows that save time, reduce manual work, and enhance the customer experience.
Unlike traditional platforms where every API call must be manually coded, chained together, and maintained with scripts, our system relies on natural-language instructions that tell the AI when and how to use each action. This means integrations are fast to configure, easier to scale, and don’t require a dedicated development team.
While other solutions may appear more flexible by allowing deep scripting, in practice they often become far more complex — requiring significant technical resources, longer deployment times, and ongoing maintenance. With our AI-driven approach, the complexity disappears: what would normally take weeks (or months) of coding can be achieved with just a few simple instructions to the AI.
Pro plans let you install ready-made Apps. These Apps automatically generate the necessary API actions, minimizing the need for technical expertise. Each App may include one or more actions, depending on its functionality, and we are regularly adding new Apps to make it even easier for your AI to connect to a growing range of services.
API actions should be well-structured and clearly instructed as AI execution relies on context and instructions. Silent actions allow for background data operations while response actions support direct user interaction.