Set-up AI for Sales with Dynamics 365 - II

This is part II of my blogs on Dynamics 365 AI. In part-1 I wrote about how you can increase your Sales by using Dynamics 365 AI for Sales. This blog is about setting up AI for Sales. There may be some challenges in installing AI for Sales and I hope this blog helps you to simplify the process.

Follow the steps below to set-up AI for Sales in a trial environment.

1. Trial environments & licenses - I created a trial environment for Dynamics 365 which provided Office 365. Also added the following trial licenses to my Office 365 subscription.

Trial licenses

As I mentioned in my earlier blog, AI for Sales is a different application and needs a separate license. It is important to have the right licenses in place before you can do any set-up for AI for Sales. Assign the above licenses to the admin user.

2. Install Sales Insights add-on -
Once the license is assigned to the admin user you can log into Dynamics 365 as the admin user. Go to Settings > Sales AI and select Get it Now.

Get AI for Sales

You can also install from the admin portal.
In Office 365 go to Admin portal  > Dynamics 365 admin center and select your instance. Go to configure solutions as shown below. Add the sales insights add on solution. You will need to wait for the Sale insights add on install to be completed.

Add Sales Insight solution

You can check the install status of the Sales Insight add-on as shown below.

Installed

 You will need to go back to Dynamics 365 > Sales AI and provide permissions for external systems (Azure ML) to access the user data for AI for Sales App as shown below.

Permissions needed

I also got an update notification which basically required me to re-install the app again - refer below.

Update notification

App install

Once the app is installed you will see the following on Sales AI - 

Ready to configure

3. Configure Sales AI - Now that the solution is installed you are ready to configure AI for Sales. 

In Dynamics 365 go to Settings > Sales AI > Overview > Dynamics 365 AI for Sales.

Here you will be able to configure the following and also note the required data for set-up.

3.1 Relationship analytics - Requires access to data in Exchange online. 

Relationship Analytics

3.2 Predictive Lead Scoring - Requires 100 qualified and 1000 disqualified leads for model creation.

I created a sample 1100+ leads data and imported them to create the lead prediction model. (If you need this sample data for testing, do connect with me on LinkedIn or twitter @mihircrm and i can provide the same)

Lead model creation

After few hours the lead prediction model was created by AI for Sales

Lead prediction model

As you can see the model is categorizing the leads in 4 buckets (A,B,C & D) with a scale from 0-100. You can manually configure the scale if needed. If there is a new set of data that was imported you can even retrain the model. Over a period of time as new leads are generated, rejected or qualified it may be a good idea to check the veracity of the model.

3.3 Predictive Opportunity Scoring - This requires 400 qualified and 600 disqualified opportunities to create a model.

4. Error handling - As I was doing the initial testing on model creation I sometimes got the following error message. 

Error in model creation

You may stumble upon this error and there is a workaround for it. You can go to advanced find and do a search on records for LeadModelConfig entity. 

Lead Model Config

Delete the records in the LeadModelConfig and this will allow you to  re-run the model creation.

Delete

I hope this walk through help you in setting up Dynamics 365 AI for Sales. In my next blog I will write about the Dynamics 365 AI for Sales app. Do let me know about your experience with AI for Sales and keep reading.

Mihir Shah

365WithoutCode