To achieve smart user journey personalization, Coho AI has 2 areas of integration: Events and Actions. Events are pieces of data about your users' interactions with your app, and Actions are how you can influence your users' journeys.

Events are coming into Coho AI in real time, and are being used to build user profiles, AI segments and predictive actions. Building upon all of those, you can create different Playbooks, which will automatically run for each user and trigger the Actions you connected, to personalize each and every user journey.

The following illustration shows few examples of events and actions:

Coho AI can integrate with many of your current tools. Here's the full list, if you can't find an integration - don't hesitate to contact us.

Events are the basic building blocks of Coho AI. An event represents a users interaction with your app. If you ever used a product analytics tool or a CDP, you probably already have events generated by your app. Otherwise, it's very easy to start sending them.

At the end of this guide you will find several ways to send events to Coho AI. Before that, we'll discuss the recommended events you should be sending, as well as what information they should contain.

Coho AI has a very flexible dynamic data mapping mechanism. It allows us to process your events in whatever structure they already have, which means less work for you during implementation. Instead of changing event structure and property names, we just need to make sure the right information is there.

Each event should contain the following properties, some of it is not strictly required, but recommended:

On any event you can add other properties which are relevant to your app.

Here's an example of an event that represents a user who updated their profile:

Here's a list of event types we recommend sending from your app. You might already be sending some of these events, while others might be irrelevant for your use-case. Don't worry if you don't send all of them. It's always possible to start with what you already have, and add more events when the need arises.

To implement Coho AI effectively, you will need to address 3 areas of integration.

The first thing you need to connect is a real-time stream of product usage events. The live usage data can come either through one of our SDKs or through a third party CDP integration. Follow the relevant integration guide for each data source:

Once you connect a data source, you can view recent events by going to the Events page in the portal.

When implementing Coho AI for the first time, it's highly recommended to import a one-time dump of historical product usage events data. This data will reduce the "Cold Start" effect of our AI models. It allows us to train on real relevant data, which increases the models' accuracy.

For most applications, a period of 3-6 months of historical events is usually enough.

To import historical data, follow the relevant guide for each data source:

You can also send us Historical Usage Events using common cloud storage tools like Google Drive, OneDrive, Box. Contact us for more details.

Coho AI can trigger actions in many different tools using native integrations. It can even trigger actions in your own internal systems, using the HTTP action. If you need assistance creating a specific action, don't hesitate to contact us.