Prescient AI

Need help with something?

Talk to data expert

Prescient AI

Prescient AI is a marketing measurement and media mix modeling (MMM) platform used by e-commerce and DTC brands to understand and optimize cross-channel ad spend. The platform models expected performance by campaign and channel and compares it against actual reported spend, helping marketers decide how to allocate budget across channels.
The Daton connector for Prescient AI replicates channel and model reference data alongside two performance datasets: model-recommended target metrics and actual reported spend, both broken out by campaign, channel, and date. This gives you the raw inputs needed to compare Prescient AI's modeled recommendations against observed results in your own data warehouse.

Prerequisites

  • An active Prescient AI account with API access enabled.
  • An App Token generated in Prescient AI for authenticating API requests.

How to Obtain Your Credentials

  • Now select API Keys section and click on ‘Create API Key’
  • Enter a name for easy reference like ‘Saras Analytics’ and leave Expires at and Channel scope as blank (Your can reduce the scope if needed)
  • Store the API key securely — you'll enter it during the Daton setup steps.
  • If you cannot find the section, please contact your Prescient AI Customer Success Manager.

Steps to Integrate

  1. Integration Name: Enter a name to identify this integration in Daton.
  2. Description (optional): Enter a description to help you identify this integration later.
  3. Replication Frequency: Choose how often Daton should sync data — 1h, 2h, 3h, 6h, 8h, 12h, or 24h.
  4. Replication Start Date: Select the historical start date from which Daton should begin pulling data.
  5. Lookback Window: Choose 7, 15, 30, 60, or 90 days. This determines how far back Daton re-checks recent data on every sync to capture attribution and model updates.
  6. Advanced Configuration:
    • Loading Mode: Choose Append (insert new rows) or Upsert (insert and update matching rows).
    • Loading Nested Objects: Choose how Daton unnests nested source fields into the destination — inherit warehouse default, do not unnest, up to one level, or up to two levels.
  7. Authenticate: Enter your App Token from Prescient AI.
  8. Select Tables: Choose which of the four available tables to replicate.
  9. Configure Schema: Review and adjust the columns to be created in the destination for each selected table.
  10. Submit: Complete the setup to save the integration and begin replication.

Supported Tables

Table Full Load Supported Description
channels Yes
List of available marketing channel names configured in Prescient AI.
models Yes
Model definitions, including the measurement unit used by each model.
modeled_metrics No
Model-recommended target metrics per campaign, channel, and date.
reported_metrics No
Actual observed ad spend per campaign, channel, and date.

Table Schemas

channels

Column Type Description
channelNames STRING (Primary Key)
Name of a marketing channel available in Prescient AI.
     

models

Column Type Description
name STRING (Primary Key)
Name of the measurement model.
unit STRING
Unit of measurement used by the model.

modeled_metrics

Column Type Description
metricName STRING
Name of the modeled metric.
metricValue BIGNUMERIC
Value of the modeled metric.
processDate TIMESTAMP
Timestamp when the metric was processed by Prescient AI.
reportedDate DATE (Primary Key)
Date the metric applies to.
sourceCampaignId STRING (Primary Key)
Identifier of the source campaign.
sourceCampaignName STRING
Name of the source campaign.
sourceChannelName STRING
Name of the source channel.
target STRING
Target metric being modeled.
targetChannelName STRING
Channel associated with the target metric.

reported_metrics

Column Type Description
campaignId STRING (Primary Key)
Identifier of the campaign.
campaignName STRING
Name of the campaign.
channelName STRING (Primary Key)
Name of the channel.
processDate DATE
Date the spend record was processed by Prescient AI.
reportedDate DATE (Primary Key)
Date the spend applies to.
spend NUMERIC
Actual observed ad spend amount.
foreseeing predictive

Related Articles

Table of Contents

Prerequisites How to Obtain Your Credentials Steps to Integrate Supported Tables Table Schemas channels models modeled_metrics reported_metrics