Help Center

Getting Started with AI Workbench

BlueConic-AI-Machine-Learning-Marketing.pngAI Workbench combines the power of machine learning with the rich profile data in BlueConic.

See the BlueConic AI Workbench in action

How to use AI Workbench in BlueConic

BlueConic users with the “AI Workbench” permission enabled (Settings > Access management > Roles) can use AI Workbench to apply machine learning marketing models to BlueConic profiles and data. With prebuilt AI models for CLV and RFM, marketing teams can use AI for marketing without writing any code.

  1. To get started, select AI Workbench from the BlueConic navigation bar.
    How to do machine learning and AI marketing using customer data in the BlueConic CDP AI Workbench for CLV and RFM modeling and predictive analytics for marketing?
    AI Workbench UI opens, and here is where you add, delete, and manage Jupyter notebooks.
  2. Select Add notebook and choose a new notebook type, for example "API examples notebook" to get started. These examples contain sample Python code, and also a basic simple notebook with some visualizations.
    Getting Started with ai marketing, machine learning marketing, predictive analytics, and CLV and RFM modeling with BlueConic customer segments in a CDP
  3. Provide a name for your notebook at the top of the page. See Best practices for names and labels for tips on naming your AI notebooks.
    If you're familiar with Jupyter, you'll recognize the menu, toolbar commands, and code cells from the notebook interface. See the AI Workbench API page for reference documentation.
    Here's an example notebook showing an application that calculates customer lifetime in days, based on profile properties that store first and last visit dates.
    How to measure customer lifetime value (CLV) using machine learning marketing, predictive analytics, A/B testing and AI marketing with the BlueConic CDP?
    Saving and deleting notebooks: Use the Save menu to save, save as, or delete a notebook in AI Workbench.

    Adding parameter values to a model: Marketing teams can add parameter values to a notebook in the UI on the Parameters panel without having to edit or write code.

    Running notebook code: To execute the selected notebook cell, select Run on the notebook toolbar, press Shift+Enter, or use the run commands in the Schedule and run history panel. Learn more about scheduling AI Workbench notebooks.

    Kernels: The Jupyter notebooks you create inside the BlueConic AI Workbench are connected to a kernel through the Jupyter UI. A Python3 kernel executes the notebook code and returns results. The kernel and the notebook remain active while you are actively working in them. The kernel will be automatically terminated after 36 hours of inactivity, or when a BlueConic user manually terminates them. Also, a kernel will be automatically terminated when a notebook is opened without running one or more cells (once you navigate to another notebook).

    Privacy management: In the right-hand side bar, you can see a list of the privacy and consent management Objectives and other related items for this notebook. Select Add to objective to add this notebook in an existing privacy objective.

  4. Return to AI Workbench to see the list of notebooks.
    In the list of Jupyter notebooks, each notebook icon shows the AI Workbench notebook's status. 
  5. To see all the running kernels, select Running notebook kernels from the Save menu in AI Workbench.
    How do I run AI models, predictive analytics,and machine learning marketing CLV and RFM models with BlueConic AI Workbench in a CDP?
    The Running notebook kernels window appears, showing all running notebooks and the amount of memory in use and free.
    How do I run ai marketing models, predictive analytics, and CLV and RFM models with machine learning in the BlueConic CDP?
    Memory: At the bottom of the window, you can see how much memory is available and in use. You need at least 100MB of RAM available to start a new notebook and kernel.
  6. To terminate a kernel, select its name, and select Terminate kernel(s).
    Jupyter kernels that have been idle for 36 hours will automatically be terminated.

Using the prebuilt AI notebooks to calculate CLV and RFM 

BlueConic provides several prebuilt AI notebooks marketing teams can use (without writing any code) to apply the power of AI to your customer order data to calculate a user's customer lifetime value (CLV) and find their recency, frequency, monetary value score. 

To get started, open the AI Workbench page in BlueConic, choose Add notebook and select CLV notebook or RFM notebook. The CLV and RFM notebooks use Order event information you have uploaded to customers' BlueConic Timelines to calculate customer CLV and RFM metrics. You can use the Probabilistic matching notebook to employ fuzzy matching to find matching profiles. Follow the directions in the notebook editor panel for setting parameters that feed these calculations, customer scores, and predictive analytics.

How do I use BlueConic AI Workbench to calculate customer lifetime value CLV or RFM recency frequency monetary value using ai marketing models and predictive analytics for BlueConic customer segments?

Using the prebuilt A/B Notebook for A/B testing

Does BlueConic offer A/B testing with AI marketing?AI Workbench offers a prebuilt notebook marketing teams can use out of the box to optimize A/B testing with BlueConic, without writing any code. 

To get started, open the AI Workbench page in BlueConic, choose Add notebook and select A/B notebook. In the Parameters panel of AI Workbench, select the relevant dialogues in the parameter settings. Note that the A/B notebook requires that you have a control group dialogue. See the Dialogue Optimization page to learn more about optimization settings for dialogue control groups.

Run the notebook to see how dialogues and variants are performing. (Reminder: To run A/B tests in BlueConic, you set up dialogues and variants of those dialogues, using the dialogue optimization feature, deliver the variants to different audiences, and compare results.) Learn more about the Advanced A/B testing notebook.

How do I use machine learning and AI for advanced A/B testing with the BlueConic customer data platform and AI Workbench?

Notebooks for predicting and analyzing customer churn

AI Workbench offers two companion notebooks for predicting and analyzing customer attrition rates, or churn, for a segment or segments of customer profiles in BlueConic. You can choose from several machine learning models to calculate and compare customer churn rates:

How to predict and analyze customer churn or customer attrition using AI marketing models in BlueConic CDP

How do you predict customer churn using AI marketing models in BlueConicPredict Propensity to Churn notebook: Marketing teams can use this notebook to predict the probability that customers in a certain customer segment will churn. You can also use this AI model to predict the propensity to churn for other profiles in a segment based on a particular profile property. For example, if you select 'gender' as a 'categorical profile property,' the notebook can create separate models for men and women.
Start by selecting a specific segment of customers in the parameter tab. Then choose a specific profile property which will be used to make propensity to churn models for the different categories of the profile property. When you run this notebook in AI Workbench, the model saves the churn prediction to a profile property you select in the AI Workbench Parameters tab. In a few minutes, the results will be visible in your BlueConic profiles. Learn more about the Predict propensity to churn notebook.

How do use AI models to predict customer attrition or customer churn with the BlueConic customer data platformAnalyze Propensity to Churn notebook: Use the churn analysis notebook to analyze and compare churn rates for different segments of customers. Using contract or subscription data, the model calculates subscription status and how long customers have subscribed or held a contract. For a given segment of customers, the model calculates churn risk per unit of time, and displays churn risk as a function of time. 

Using the Parameters tab in AI Workbench, you can select a specific propensity to churn model and also customize colors used in your insight. Learn more about the Analyze propensity to churn notebook.

Performing identity resolution with fuzzy or probabilisitic matching

BlueConic provides a Probabilistic (or fuzzy) matching notebook in AI Workbench that you can use to find “fuzzy” matches in your customer profile database. Fuzzy matches are profiles that likely belong to the same person even though not all fields in these profiles have the exact same values.

A simple, but common, example of a fuzzy match is where two profiles have identical first names and surnames, but their phone numbers differ by one digit, possibly because of a typo. The notebook determines a match by finding common typos, misspellings, and the deliberate replacement of a character or digit by a person with multiple profiles. This happens in a probabilistic way, instead of through exact matching. It is important to note that the notebook also detects exact matches for the profile properties it examines.

The Probabilistic matching notebook uses the symspell algorithm to find near matches and measures similarity between values based on the Damerau-Levenshtein measure of edit-distance. Contact your Customer Success Manager to learn more about using this notebook.

AI Workbench resource usage

Existing BlueConic customers can use AI Workbench free of charge as long as your total data operations remains smaller than the amount that is allotted to your subscription. Choose Settings > General to see your BlueConic data usage statistics.

Contact your Customer Success Manager at if you intend to use this feature, to ensure your subscription can accommodate your intended usage. We'll discuss your requirements and upgrade your subscription as necessary.

Best practices

Hear from BlueConic's resident expert on Best Practices for Optimizing Your AI-Powered First-Party Data Strategy.

See our tips for making the most of ChatGPT and other generative AI tools with your CDP: Use Generative AI with BlueConic: How tools like ChatGPT can save you time.

Next steps

Share your results using dashboards and insights. See the Notebook - All cells insight and Notebook - Single cell insight to visualize your AI Workbench results.


Was this article helpful?
0 out of 0 found this helpful