Skip to main content
Version: v5.0

AI Service use-cases

Why use the AI Service?#

The AI service provides a way to retrieve platform data using natural language inputs.

Real world use case of applying AI#

Consider the use case of detecting an anomaly in a building's Heating, Ventilation and Air Conditioning (HVAC) system's pressure sensors. A building operations manager wants to identify if any HVAC units are operating outside of safe pressure ranges. If this is the case then it could indicate some system problems.

Traditionally, platform data would be queried as follows.

Traditional method of querying HVAC pressure sensors#

The traditional method would proceed as follows:

  1. Users interact with a fixed user interface on a front-end application to manually select:​
  • Specific sensors​
  • Pre-defined safe pressure ranges​
  1. The front-end then constructs API queries for the Item Service and retrieves the pressure readings.​
  2. Basic logic on the front-end is then used to identify anomalies based on hard-coded thresholds.​

Limitations of this approach

The main limitations of this method are:

  • A fixed, rigid UI restricts user flexibility and limits the possible range of supported use cases.
  • Writing complex Item Service API queries requires technical effort.
  • Hard-coded logic makes the system less adaptable to new sensor types or anoaly criteria.

AI method of querying HVAC pressure sensors#

Utilizing AI, the following method could be used:

  1. Users simply provide natural language inputs. For example, "Check if any HVAC units have abnormal pressure levels".
  2. App developers can configure AI agents and Tools within the AI service to:
  • Understand the user's intent
  • Translate the user's input into dynamic criteria for the Item Service API.
  1. Tools execute the API calls and retrieve relevant data
  2. AI Agents can analyze the data and highlight units with abnormal readings.

Advantages of this approach

The advantages of this approach includes:

  • Uses a natural language interface
  • Facilitates dynamic query generation which supports a wide range of use cases without needing UI changes.
  • Reduces the technical complexity for front-end developers for writing queries to the Item Service API.