New predictive features at NAV 2017 with Cortana Intelligence

We now have the latest version of Microsoft Dynamics NAV available, and with it have come countless improvements and technological changes, which will mark a before and after in the "classic" conception we have of an ERP system.

One such improvement is “Cortana Intelligence”whose purpose is to provide predictive capabilities to our system.

Table of Contents
  1. How is it possible?
  2. How do we configure this in NAV 2017?
  3. Time series library:
  4. Conclusions:

How is it possible?

It is possible thanks to a "Machine Learning" service that Microsoft makes available to us. For those who don't know, “Machine Learning” is a branch of artificial intelligence whose goal is to develop techniques that allow our information systems to “learn” from unstructured information provided in the form of samples or examples. In the case of NAV we will obviously use our historical ERP data, from which we will be able to analyze how and what the evolution of our business is.

Therefore, if we start from the premise that “we have a method to teach our system to make future predictions based on historical data”, let's imagine the possibilities:

  • How would you like to know how much we will sell of a product based on historical sales of the same or similar products?
  • And know if a customer might purchase a product based on past purchases from customers in their profile?
  • Or make a cash forecast with historical data on receipts and payments?

What can I do immediately with Dynamics NAV 2017?

Well, I'm telling you, we'll find one "extension" in the system, which is called “Sales and inventory forecast”.


As we can deduce, thanks to this extension we will be able to make forecasts on both sales and inventories. In fact, once the extension is installed, we will be able to view in the ribbon, for example, this "Forecast" section of the product list:


As well as information in graphical mode in the form of a “fact-box” on the right side.

How do we configure this in NAV 2017?

First of all we have to go to the “Configuration of sales and inventory forecasts”, where we see that there are 2 fields not filled in:

  • API URI.
  • API key.

We must obtain this data by registering Microsoft Azure Machine Learning Studiosince as I already mentioned we will use a "Machine Learning" service in Azure and then we will make a copy of the "experiment" in our gallery to be able to use it.

  1. The first thing we do is visit the experiment website in the Cortana Intelligence collection:
  2. On the right we have a button where it tells us “Open in the Studio”.
  3. This redirects us to Microsoft Azure Machine Learning Studiowhere it asks us to log in with our account (if we have one) or register. USE: With the “free” account we can use the features we need without problems.
  4. After that we have to copy the experiment in the gallery, to conclude we do it "RUN" Yes "Deploy Web Service":

On the "Dashboard" screen, we will have something like this:


And we have to select "REQUEST ANSWER".

With this we will access the file "API URI" (Request URI)that together with “API Key”they will help us insert data into the NAV for the extension to work.

Now we need to go to NAV to complete the information in the forecast extension configuration.


And with this we have completed the configuration. As you can see, this is a fairly simple and quick process.

If something was left behind, you will receive a message like this:


Which generally corresponds to an incorrect entry of the API Key.

We can also use this same API information to Cash flow forecast en Microsoft Dynamics NAV 2017.

To do this we must run the wizard "Set cash flow forecast"that we have inside “Assisted setup”:


The wizard is very simple, we just need to select the accounts on which we want to base our forecast, the frequency of updating our “cash flow forecast” and whether we want to include “Cortana Intelligence” data.

When we continue with the process, we enter the data we already have related to the API key and API URI. To conclude with the guided procedure, we are asked to specify the data relating to the payment of taxes (frequency, payment window and account). .

And in our predictions we will now have something like this:


Well, up to this point I have told you what we can do from the first minute with NAV 2017, but what is really interesting is the range of possibilities that opens up thanks to this technology.

Time series library:

We found within the NAV 2017 a Codeunit called “Time Series Management (20000)”where a series of functions are collected that we can call from our code to exploit the predictive capabilities of Cortana Intelligence.


Inside we find various functions, which we can obviously invoke from our code.


It would be something like this, for example:


With this we could make the prediction for a product whose parameter we change.


I sincerely believe that this functionality will give great possibilities to the product, and working with it intelligently can greatly extend the current capabilities of the ERP, in many facets.

Any example I can think of is valid: companies that supply and store chemicals with expiration dates and shelf life control, companies with a lot of seasonality, etc. etc. The potential is brutal.

Well, I trust you found this brief review of the new predictive features in Microsoft Dynamics NAV 2017 interesting.

Thanks and regards, Miguel Llorca.

If you want to read other articles similar to New predictive features at NAV 2017 with Cortana Intelligence you can visit the category Power Apps.

Leave a Reply

Your email address will not be published. Required fields are marked *

Your score: Useful

Go up