You’ve now connected data, made predictions, and explored relevant insights and model details.
While this may be enough for some use-cases, there’s a lot more you can do.
Namely, you can deploy your AI model as an API, as a web app, or even create integrations in Zapier.
A dead-simple deployment option is to make a public or private webpage to run your flow. This can even be embedded in external websites.
In one-click, you’ll get both a URL to your deployed site and an embed code. The web app will look something like the below, which allows visitors to input data (manually or with a data upload) to get a prediction.
Another popular deployment method is to connect your flow to Zapier, which is easier than figuring out the API, and a good no-code option.
For example, you could create a Zapier flow to automatically make predictions, via Akkio, on new data in a Google Sheets file.
Akkio comes with a built-in Salesforce integration, enabling you to integrate with the number 1 sales CRM in one click. You can now predict key outcomes, deploy to Salesforce, and build an AI-driven sales team in record time.
A more flexible option is API deployment, as APIs can be integrated virtually anywhere - not just Salesforce. You could deploy the API to make predictions in a mobile app, or through your website, regardless of what tools and technologies you used.
This also means the API is the most complex deployment option, so if you’re new to APIs, you may want to take a 101 course.