Back to Blog

Getting Started with SalesForecaster in 5 Minutes

A quick walkthrough to upload your first dataset, train a model, and generate your first AI-powered forecast.

SalesForecaster Team3 min read

Getting Started with SalesForecaster

This guide walks you through your first forecast — from uploading data to interpreting predictions. You'll be generating AI-powered forecasts in under five minutes.

Prerequisites

Before you begin, make sure you have:

  • A SalesForecaster account (sign up here)
  • A CSV or Excel file with historical sales data (minimum 12 months recommended)

Step 1: Upload Your Data

Navigate to Data > Upload in your dashboard. Drag and drop your file or click to browse.

SalesForecaster supports:

  • CSV files with headers
  • Excel (.xlsx) workbooks
  • JSON arrays of objects

Your data should include at minimum a date column and a sales/revenue column. Additional columns like product category, region, or customer segment will be automatically detected and used as features.

date,revenue,region,product_line,deals_closed
2024-01-01,125000,North America,Enterprise,12
2024-01-08,118000,North America,Enterprise,10
2024-01-15,142000,EMEA,Mid-Market,15

Step 2: Configure Your Dataset

After upload, SalesForecaster automatically detects column types and suggests configurations:

  • Target variable — The column you want to forecast (e.g., revenue)
  • Date column — Your time dimension
  • Frequency — Daily, weekly, monthly, or quarterly
  • Categorical features — Dimensions for segmented forecasts

Review the suggestions and adjust if needed, then click Save Dataset.

Step 3: Launch AI Research

This is where the magic happens. Navigate to Research > New Session and select your dataset.

The AI research agents will:

  1. Analyze your data patterns and seasonality
  2. Search for relevant external factors (market trends, economic indicators, industry events)
  3. Build a knowledge graph connecting discovered factors to your revenue drivers
  4. Generate a research report with recommendations

This typically takes 2-5 minutes depending on your industry complexity.

Step 4: Train Your Model

Go to Models > Train and select your dataset along with the research session results.

Choose from:

Model TypeBest ForTraining Time
Quick ForecastFast initial predictions~30 seconds
Standard EnsembleBalanced accuracy & speed~2 minutes
Deep EnsembleMaximum accuracy~5 minutes

We recommend starting with Standard Ensemble for your first forecast.

Step 5: Generate Forecasts

Once training completes, navigate to Forecasts and click New Forecast.

Select your trained model, set the forecast horizon (how far ahead to predict), and click Generate.

Your forecast includes:

  • Point predictions with confidence intervals
  • Feature importance showing what drives each prediction
  • Anomaly alerts for unusual patterns
  • Scenario comparisons against baseline

Next Steps


Have questions? Visit our Help Center or contact us.