Quick Start

System Requirements

Before you begin, make sure your system meets the following requirements:

Browser Requirements:

• Chrome 90+ (recommended)

• Firefox 88+

• Safari 14+

• Edge 90+

Data Requirements:

• Structured data in CSV, Excel, or JSON format

Create Your View AI Account

Head over to app.viewai.ca/signup and create a new ViewAI account.

Sign Up Options:

• Email: Create account with work email

• Google: Sign up with Google account

• SSO: Enterprise single sign-on (Enterprise plan)

Step 1: Connect Your First Dataset

Navigate to Data → Upload Dataset or click the Quick Actions card on your dashboard.

Direct Upload: You can upload files directly from your computer. This is done by clicking on the upload area where you can select files from your local system.

Data Connection: If your data is stored in a cloud data warehouse, you can connect by selecting this option. This will require you to enter credentials to securely access the data. You can see the supported connectors here:

1. Google BigQuery

2. Salesforce

3. SAP

If you already have your own model, View AI supports explaining blackbox models or even pipelines. Alternatively, if you are looking to build new models, View AI trains interpretable models tailored to your specific needs. Learn more about Interpretability at View AI. ​

Step 2: Train Your First Model

After uploading data, you're ready to train a model.

1. Navigate to Models → + New Model

2. Select your dataset from the dropdown

3. Review the data summary showing:

   • Total rows and columns

   • Data types

   • Missing values percentage

4. Click "Next"

Select Target Variable:

1. Choose what you want to predict (e.g., "churn", "loan_default")

2. ViewAI automatically detects problem type:

   • Classification: Predicting categories (Yes/No, High/Medium/Low)

   • Regression: Predicting numbers (price, quantity, score)

3. Click "Next"

Configure Training:

1. Select training mode:

   • Guided (Recommended): ViewAI selects best algorithm automatically

   • Advanced: Choose specific algorithms and hyperparameters

2. Enter model name: "Customer Churn Predictor"

3. Click "Start Training"

Training typically takes 15-30 minutes. You'll receive an email notification when complete.

What is a Dashboard in View AI?

Explainable AI dashboards are interactive interfaces that visualize the performance and inner workings of machine learning models. They offer insights into how the model makes decisions, how it's trained, and how it's performing on current data.The main goal of these dashboards is to promote transparency and understanding of AI systems. They provide a platform for users and stakeholders to see the impact of the model's predictions and its usefulness in solving real-world problems. You can access the:

• Performance metrics: These are numerical measures of how well the model is performing. They can include accuracy, precision, recall, F1 score, and others.

• Copilot: This helps users to see the relationships between features, explore the data, and understand how the model makes predictions using natural language queries.

• Interpretability techniques: These are methods for explaining how the model arrived at a prediction. View AI includes feature importance, decision paths, feature attributions and what-if/counterfactual analysis.

Dashboard Management

• "Dashboards in this team" displays everything your team is working on. You can see all your team's dashboards here in one place.

• Recent Boards: This is where you’ll see the dashboards you've worked on most recently, so you can pick up right where you left off without having to search for them.

• Starred Boards: If you have go-to dashboards you frequently check, star them for quick access here, kind of like how you might pin frequently used datasets or scripts in your development environment.