Insights on LinkeD365 Blog

Customer Service Insights

Thu, 04 Jun 2020 00:00:00 +0000

Following on my series of posts about Sales Insights, I thought I would learn more about Customer Service Insights.

Sales Insights, in my view, is all around making the sales process more effective reducing the time spent editing data, highlighting activities that a sales user should be doing and tracking activity with your customer. Sales Insights includes forecasting and predictions as well to allow a better understanding of what makes a sale and what is preventing closing deals.

Customer Service Insights seems to be a more typical “Insights” application where a service manager can get an overview of the current state of the Service team, measure the level of KPIs and drive understanding of the bottlenecks.

TL;DR

Customer Service Insights provides a great suite of reports to visualise a servicing centre. It allows AI categorisation and grouping of cases to highlight patterns in your case resolution, ensuring you can quickly reduce bottlenecks. It also highlights areas where you should create FAQs or Virtual agents for your more frequent case categories.

At £56.60 per user per month, it is not an application you would give to all your support users, but service managers would definitely benefit from the insight into their organisation it brings. With the ability to generate your own reports via Power BI, one license could open up possibilities for combining this data with other sources in your board packs or daily reporting scenarios.

Setup

Start with a trial. It is free to sign up. Head over to the Microsoft site here and select Get started.

Choose the region and select Get started.

Now you need to connect your data. Good to see Salesforce and other data sources are coming soon. I selected the only choice, Dynamics 365. This then prompts you for an instance to use, select the most appropriate.

It then goes and looks at your data. Nice animation!

Once done, you are presented with the base screen for Customer Services Insights, with a banner warning you that it is not quite ready yet.

It is that simple! Unfortunately for me, I am working with OOTB demo data. My data range is very limited. This is where I rely on Power Automate to generate some case data. I documented this in my previous post here.

Now that the data is available, time to walkthrough the application and it’s features and benefits.

The information is refreshed daily. Be patient, particularly if you are changing things. The date of the last refresh is visible in all the screens, top left.

Home

The home page has lots of useful stats on your topics and tools to help you. Top right, you can define the timeline between 3 options.

The first section is Topics to Watch. This gives you 3 topics that are high in a combination of the number of cases, negative CSAT score and an increasing volume of the number of cases over the time period you select. So, over the last 30 days, App sulfamethoxazole/trimethoprim damaged was highest overall. You can also see the 3 ratings below.

Next comes Topics Impacting resolution time. This shows the highest time to resolution for these topics, with shortest, average and longest time period show. Drill down to the topic to see who is performing well in this topic and do a bit of cross-learning for those that aren’t.

The next section highlights topics that could be automated. This means that the resolution time is quick (or simple) and the number of cases is high. Automating these cases could drastically reduce the impact on your support desk.

Next shows the impact of channels on the CSAT score. Ensure your users are trained and skilled in responding effectively in each medium. It also shows the channels that problems are surfacing on. Be careful to ensure the channel has the appropriate time to react.

Finally, the time at which cases are raised over the last month is shown. In my data, as I mentioned, the create time was not used, so shows as blank. This little graph shows you the peaks and troughs in case creation which allows you to alter shift patterns to increase the helpdesk users at those times.

Drilling into the topic presents you with Topic analytics page.

Topic Analytics

This gives you a snapshot of all scores gained from analysis of the cases and insights into how this type of case is fairing in your business. In the top right, you have the option to Automate this topic (only sends you to the website, does not bring anything through from the topic) and to rename the topic. Here you can re-label the generated topics, to ease the tracking etc. This applies throughout the application once done.

In the top zone, you can filter the results to a given time frame, product, channel, BU, and owner of the case. These lists filter by those that apply rather than all the available channels.
The second zone shows some simple stats from the case topic. It also has how this category affects the overall figures.
Essential to any service desk is the satisfaction rating an end user gives the company on the back of this case. This can come from various sources but is an immediate rating of your customer support desk and broader, to your organisation. This, broken down by product and channel and includes statistics on the resolution time and the average CSAT score for each. Also, it shows how this case has impacted the average CSAT score for all the cases.
The next section shows these cases by agent, highlighting those with the highest number of open cases, longest resolution time and lowest CSAT score. This data shows little in my created data as my user account owns everything, but you can quickly see those underperforming support users, or who needs a little training.
The final area shows the impact of cases by product and channel on the resolution time. Here you can see that some cases take a lot longer than others, which is something you may need to provide training on.

Below the final area is a graphical representation of the makeup of the cases that are in the timeframe that selected. This shows how each channel flows through to its current state.

Remember, this is Power BI so the usual tools to export the data and show the table of records is all there.

KPI Summary

The KPI summary view shows your progress against resolution and escalation SLAs and how topics, priority, channels etc affect this.

Again, the top area is to allow you to filter the data, with timeframe, channel, product, BU and team.
This area shows core stats on the cases. This includes the number of cases, number of resolved cases, number of escalations, the number of cases that are compliant in the SLAs that are associated with the case, the average time to solve a case and the average CSAT score.
This lists the case topics with a volume for each. The link takes you to the Topic analysis screen.
This set of graphs shows you some insights in cases.
1. The first graph shows you the new cases in the period and the cases that were still open when the period started.
2. This show the breakdown of the priority of your cases.
3. Finally, the case channel with a breakdown with the status.
Emerging topics shows a change in the cases data for each topic over the current period and the previous ones. Here you can see the case topics that are increasing, maybe a product role out or a new software patch will drive more calls to your contact centre.
The final area is two graphs, the first showing you the number of cases per day raised in the period as well as those that are resolved. The last chart shows the time to resolve cases and the count for each period.

These graphs can be selected, filtering the other graphs as you go around.

New Cases

The next tab focus on cases created over the period.

Again, this is a set of filters available for you to interrogate the data further.
This section groups the cases created in the timeframe into topics and shows the percentage of each, the number of each, the increase in volume since the last period, the average resolution time, percentage of cases closed and the average satisfaction rating. The link takes you to the Topic analysis page for each.
These graphs show
1. Cases by day by priority
2. Cases by day by channel
3. Case creation time over the period by channel (again my data is all created at midnight)
The final area shows cases that have an increased number over the period. This shows the volume of cases per topic, the number of new cases, the volume change, the average time to resolution, how many are closed and the average customer satisfaction

Customer Satisfaction

The next panel is a selection of statistics related to customer satisfaction and topic affect on overall CSAT score.

Again, this is a set of filters available for you to interrogate the data further.
This list shows topics with their average CSAT score, worst at the top. It shows the percentage volume of cases, number of closed cases, number of CSAT scores recorded, the Average CSAT score, and the impact those results have on the overall average.
These two simple numbers show the number of surveys completed (cases with a CSAT score) and the percentage collection rate.
These graphs show
1. Average CSAT score across the period
2. Breakdown of score per channel

Resolutions

The final report section of the application is Resolutions. This details insights across resolved cases and the length of resolution cases.

Again, this is a set of filters available for you to interrogate the data further.
This is a list of topics with the volume of case topics, the average resolution time per topic and the impact on the average resolution time for that topic. The detail link takes you to the Topic Analytics page.
These two graphs show
1. the agents with the longest resolution time in the period
2. the new cases logged in the period vs the resolution time
These graphs show
1. Number of open/closed cases per agent (my data is all owned by me
2. New vs resolved escalations over the time period.

Configuring your data

There are several options to configure your data, available in the customer service insights settings, top right.

Data Mapping

Data mapping allows customisation of the mapping of the fields required by CSI to your data. By default, this maps the OOTB fields on the Case entity to those expected. You could use a different entity to case in your organisation or map topic to a separate field, for instance.

Clicking on the pencil to edit the configuration firstly allows selection of an entity.

Once this is done, hit Next to map the fields that are required, on the left, to those in the entity selected.

To ensure effective use of the tool, make sure all 15 fields are mapped correctly. This is pretty straight forward if you use the OOTB Case (Incident) entity and fields.

Case Title cleaning

A lot of organisations have a set configuration for their case titles. This is typical when the title is populated with parts of the case such as subject or category selection. CSI allows you to denote these parts so that they are ignored when it builds up the topic categorisation.

By default the whole of the title is used, but you can state that you use tags before, after or both the main part.

This then allows you to highlight what tags you use to separate the topic from the tags. CSI will then ignore suffixes and prefixes to your title, anything after or before one of the tags.

Topic Granularity

CSI uses the title to group topics. It could use all the words as a unique sentence, but filler words would also be used, which would not group them correctly.

Consider this sample topic that was created using my random data. It is effectively creating a topic from the drug name I used in the centre, rather than taking the whole sentence as the topic. Filler words and their uniqueness is defined by the topic granularity. If you used the highest level of granularity, very high, each unique case title would be used as a topic, reducing the effectiveness of the insights you want to achieve.

It will be down to your data and the fields you are using to decide on the best setting for your granularity. As you change the slide value to high, the groupings become smaller and more unique, using more of the title to drive that uniqueness.

Export Data

This feature is still in preview. This feature allows an external connector to use the data, to go above and beyond the reports given in the OOTB CSI configuration.

A unique URL is generated if you select Generate URL. This is available for anyone to use if they have this URL. The data is refreshed daily. There are also limitations on the number of calls per day that can be made to this API, documented here.

It would be easy to generate your own reports using Power BI or similar tool or have a daily poll to trigger notifications of high case levels etc to supervisors with this endpoint.

To add to Power BI, select Get Data, Web, paste in the generated URL and you have access, nice and simple!

This export of data is the first step, the screen shows that there will be a Power Automate\Power Apps\Logic apps connector, though you could probably use the existing one if you can limit the interaction.

Featured Image by Icons8 Team on Unsplash

Using Power Automate to generate Data

Tue, 02 Jun 2020 00:00:00 +0000

Previously, I have used Mockaroo and Flow to generate data, but the data I required for demoing the Customer Services Insights functionality has to be linked from existing data to be valid.

In this post, I will guide you through the methods I have used to randomise and create demonstration data that fits the requirements for Customer Service Insights, just using Power Automate and an Excel file.

TL;DR

Customer Service Insights only really works with data. This is great if you have some, but most demonstrations or trials have 10s of cases rather than 1000s. Here I walk through generating data using Flow, Rand() and Excel, mainly. Also, how to correct data if you didn’t quite get it the right first time 😊.

Creating Cases

Cases in D365 are linked to Contacts and Accounts and they need a title. As a bare minimum, this will create a case. But, for Customer Service Insights to work effectively, you need a few more parts.

Origin is where the case came from, typically Twitter, Facebook, Phone, Email, Web etc. This allows for the segmentation of your customer base by source. Products link the case back to the faulty product or service that the case was about. A Priority is also needed to define the workload case for your staff or denote a superior service. Also, we need to know the satisfaction or the originator of the case with the resolution. This can be taken from a Forms Pro survey, for example.

All these fields I have taken from the various parts of the Customer Service Insights reports and wanted to generate data with this populated to allow for an impressive and realistic visualisation in CSI.

But I wanted to randomise the data, it is quite easy in Flow to create records, but to randomise them you need to choose from a selection, you could query the data in your system for each of these, but I found it easier to use a central store of the lookup data.

Excel for Random Data

In Excel, I created an OneDrive file and added a table. In this table, I added several columns with information to build up a case.

Topic Start - Combine 3 fields to make the title of the case, which allows for randomising the title but keeping some keywords to allow for Topic categorisation by CSI.

IdCol - this is the key for the table. Each valid row has a unique number so I can retrieve the row in Flow later.

Topic Main - This is the middle part of the Title field. I am actually using a list of drug names I found on generatedata.com. By using a list, I hope to be able to drive the Topic categorisation.

Topic End - The final part of the topic, just some random words I thought sounded like ends of case titles.

Contact - This is the GUID for all the active contacts in the system. I exported the list to excel, unhide the columns on the left and grabbed that list.

Channel - These are the numeric values for the optionset used for Channel. I drilled into the options set from the field on the case.

Priority - Again a numeric values for the Priority optionset.

Product - This is a list of the active Product Ids in the environment.

Create Cases Flow

The overall flow is quite lengthy and has several conditions. Let’s step through it from the top. The flow is available in my CustServInsights solution here.

First of all, trigger the flow from a button. The input parameter denotes how many cases you would like to make. Ensure you keep this as 1 for testing. Also, initialise the case counter for use in the next loop.

Next, create a Do Until loop. This will keep looping until our case counter matches or exceeds the required number of cases.

Retrieve the random data

Next, get the relevant lines from that Excel file. For each of the fields, choose a random number from 1 and the maximum number for each data set and get the corresponding row in the Excel file. This is for Priority where I have 5 rows in the Excel sheet (I duplicate Normal which has a value 2 to weight the priorities in favour of Normal), hence the Rand function is

Rand(1_, 6_)

Rand (1,6) for Priority.

The Key Column is the IdCol of the file, populated with integers

This action is repeated for each of the random columns in the Excel file.

Create a case

Next, use these values to create a Case. Use the Create a new record action and populate with the appropriate values. Title uses a combination of each of the topic parts. Customer (Contacts) uses the confusing for citizen developers syntax to populate the contact on the case. This only applies if you are using the Current Environment version of the custom connector. There is an idea to revert this back to the previous syntax here. Please vote! I also added a description, just so I can retrieve these new cases later if I need to.

Origin uses the number for the optionset returned by the Excel and converts to an integer. The product field uses the same syntax as Contact.

Priority is converted to an integer as well. Record created on is used to alter the created date of the record. It is a pseudo change, as the record keeps both what you want to show and the actual date as separate fields in the CDS. Unfortunately, there is a bug in both CDS connectors and the D365 connector that strips out the time part of the field. I raised the bug here.

The compose used as a parameter is below. Use the rand function again to subtract up to 45 days from the current date and subtract up to 6 hours from the result. Finally, format the result to a date-time acceptable by the connector

Complete the SLAs

Once the record is created, randomise the closure as well as whether it met any SLAs in place for first-contact and closure. Firstly, generate another random number which will be used to define which path our case will take. Next, choose if this case is still open. Any value greater or equal to 8, so 8, 9 and 10 or 30% of the cases, should be still open.

On the positive side, the case is still open, another condition to decide if the case was responded to in time. This effectively means 10% of the cases will have failed their First response SLA and are still open. 20% will be still open but the first response was passed. The positive side of this condition is empty, the negative side, update the SLAs to show the first response SLA is met.

The first action in the negative side is a Update Record, using the Id of the created record. Update the first response field to Yes and the SLA succeeded to Yes. This is normally populated via workflow in the application.

The next step is to retrieve all the SLA instances associated with the record. These are created when a case is created (if you have a default SLA configured) and denote when the SLA for that case will be breached. To make these appear as if the case passed its first response SLA, update the first response instance for this case. Firstly, find the record. This is a list records action, against the SLA KPI Instances entity, with a filter to retrieve only those that are for the case created. The order by condition means that the first response instance is returned first, as it’s failure time is naturally before the closure SLA failure time. Using the top count of 1 restricts the return to only one record. This should hopefully mean we retrieve 1 record only and that is the first response SLA.

Loop through (there will only ever be one) the results and call an Update Record for each, marking the status as succeeded and populate the succeeded on date. Using the warning time will ensure it is prior to the failure time.

Closing the Case

On the negative side of the first condition statement, case is going to be closed, but first randomise whether this case met it’s case SLA. If the random number is 1 (10 % of cases) the case will be closed, but both the first response and case closure SLAs were not met.

In the Yes part, use the same technique to the get the case SLA instances and update all of them to Non compliant.

On the No side, make them all Succeeded and populate the date with the warning date, same as previous. The final part is to close the case. Whilst you now have the ability to perform an unbound action, I could not get it to work. Instead, I used a custom connector.

Custom Connector to close Incident

I have numerous posts about configuring custom connectors, start with this article on LUIS to create one. The action is CloseIncident and defined like this.

The body of the request is below.

1
2
3
4
5
6
7
8
9


{
 "IncidentResolution": {
 "subject":"test postman",
 "incidentid@odata.bind": "/incidents(7b978321-e7a0-ea11-a812-000d3a7fc8be)",
 "overriddencreatedon": "2020-05-22T13:28:43Z"
 },
 "Status":-1

}

This is a post action against the CloseIncident action, the request is defined below.

Drilling into the body parameter, you can see each of the parameters that are required.

Using this in Flow is straight forward, add a new custom connector and populate the parameters. Be careful over the case Id, as it needs the odata binding populated.

Finishing the loop

The final part we do is increment the counter variable to ensure only the number of cases we specified are created.

In terms of creating cases, that is it. I created nearly a thousand cases this way in 30 minutes (took a few hours to build out the code obviously) but this is reuseable and even 1000 records in 3 hours is not achievable by hand.

Updating cases Flow

There is always a time when you forget something. In this scenario, Customer Services uses the CSAT score and whether the case is escalated, and my original flow didn’t include that information. In hindsight, add these values to the Create case flow. To correct, I created another flow.

The trigger is the same, manual flow button. The first step is to retreive all the cases that I created. Remember I added to the description “This is a test case generated on UtcNow()”? This allows me to retrieve all the cases with a description that starts the same.

Then I loop through the retrieved cases. First of all, create 2 random numbers for use later. Then check if case is closed.

If a case is closed, all you can do is update the status and a few other fields, not all of them. If you try to update the Satisfaction on a closed case, you will get an error. So, for closed cases, open it first, by updating the record status and status reason.

Then, update the record for a second time, this time adding the Satisfaction and Is Escalated flag. Satisfaction uses a formula to state if the random number, 1-6 is 5 or 6, use 5 as the rating, otherwise use the value of the random number. This allows for a weighting of the satisfaction emphasising a higher rating.

Is escalated is true or false. In this scenario, I use a random again, so if the number is 5 or 6 using a range of 1-6, the case was escalated. This means only a small proportion is escalated.

Now the case has the data required, close the case again with the custom connector used earlier.

For the open case, side, just update the record, no need to open first.

Header Photo by Alexander Sinn on Unsplash

Sales Insights: Sales accelerator

Thu, 28 May 2020 00:00:00 +0000

This is the final of a 6 part series on Sales Insights, where I attempt to walk through of the Sales AI functionality that is part of the suite. The other articles are listed here.

Basic (Free) Setup - The “free” AI capabilities that can be added to any Sales Instance

Assistant Studio - Full capabilities - The full suite of options for Insight Cards

Productivity Intelligence - Activity and Contact suggestions as well as Notes Analysis

Connection Insights - Relationship Analytics and Talking points

Predictive Models - Lead and Opportunity scoring along with premium forecasting.

This post will look at the Sales accelerator, which is still in preview.

Sales Accelerator

Playbooks have been around for quite a while - the ability to have a pre-defined set of tasks when something happens. Sequences, which are fundamental for the Sales Accelerator, are a series of prompts to do things which are your best way of completing or reacting to an event. Applying this across your opportunities and leads, it can define what the next action should be today and for which deal.

Configuration

First, the terms need to be agreed to within the Sales Insights section.

Next, the content and layout needs to be defined, which adds components to the Sales Insights forms.

Ensure you configure the Engagement settings as well as security, I left these as the default.

Let Publish and see what has changed. Veronica, my tame Sales user, now has a My Work section in the menu. This shows the Sales Accelerator option. Clicking on this displays a list of things that should be done and the record as shown here.

Without adding any sequences, the My work is pretty blank, just showing incomplete tasks.

Sequences

As we have discussed, sequences are a series of recommended tasks of your best practice to complete a sales process successfully. As an administrator, add a new sequence from the Sales Insights settings.

Choose a name and also which of the entitles you want this sequence to start on.

You are then prompted to create your first activity.

I chose Email. There are a variety of options depending on what you choose, email having the ability to use a template.

Next, add steps as your requirements dictate. Another important one is a wait time. This will mean the next activity will not be displayed until that time has passed.

Build up your sequence with other steps and delays.

Save the sequence then activate it. Activation allows a user to apply the sequence to a record but prevents you from editing it while it is active. You need to deactivate first. Any sequences currently running will be disconnected from the sequence when is deactivated. This is nonsense, in my opinion. If you have a published sequence that you need to tweak, all users would have to re-apply the sequence and start from scratch.

In a Lead or Opportunity view now, if you select one or more records you get a new option to connect a sequence.

This will display a list of the active sequences. Select one and choose Connect.

![]((images/2020/05-image-230.png?fit=1024%2C395&ssl=1)

In the My Work section now, you will see an additional list, like below listing 3 lead actions to be done today. This view also shows the other steps I created.

![]((images/2020/05-image-231.png?fit=1024%2C477&ssl=1)

The steps can then be completed or action from that central pane. In this case, we can send an email or just mark it as complete. The next action, with the associated record is displayed in the main view.

In the Up Next pane, the action depends on the type of step. Email will create a new email and populate with the template is selected. Task will use a quick create form. Phone call will attempt to use your configured dialer. I find this a problem as most sales environments won’t have an integration, and will frequently rely on a mobile to do the actual call. I would still want to log the fact that this call occurred.

On the left panel, there are options to sort the list by score and at this point, it brings in all the capabilities of Sales Insights together.

Final Thoughts

Sequences are a great tool to aid your workforce if your sequence of events is simple. I would like to see branching within the sequence with steps being defined by data on the record. Imagine if the sequence branched determined by the expected revenue or days to decision date? I added this idea here.

You can achieve both in Flow, but you need to create the records as you do if you use Playbooks. Sequences allow you to suggest record creation rather than enforce. Also, there is no customisation in sequences, except for Email Templates. Custom activities are not supported, for example. Would love some of the tasks to be automated, like an email reminder if no email has been returned from the customer. And it should interact with the email engagement part to

The deactivation process also needs to be altered. Deactivating will de-link the sequence form the record. I would suggest deactivating leaves sequences connected to records to run but prevents new connections. You should be able to clone a sequence so that your new one can be used once you have tweaked it. I added the idea here.

Whilst I think it is a good step forward, the functionality available in tools like SalesSpark is above and beyond what is available in Sales Accelerator now. I do think that Microsoft will get there though and hopefully will bring more automation and tools to the standard Sales functionality.

Sales Insights: Predictive Models

Wed, 27 May 2020 00:00:00 +0000

This is the fifth in the series on Sales Insights, were I attempt to walkthrough the Sales AI functionality that is part of the suite. The other articles are listed here.

Basic (Free) Setup - The “free” AI capabilities that can be added to any Sales Instance

Assistant Studio - Full capabilities - The full suite of options for Insight Cards

Productivity Intelligence - Activity and Contact suggestions as well as Notes Analysis

Connection Insights - Relationship Analytics and Talking points

This post will walkthrough the Predictive models functionality, which includes Lead and Opportunity scoring as well as Premium forecasting.

Lead Scoring

Predictive Lead scoring is a machine learning model which puts a score against each open lead. This allows your sales users to prioritise their interactions with leads, nurturing the right ones at the right tie to improve qualification results and reduce the time take to qualify a lead.

The model used is specific to your organisation and needs to be configured by the administrator prior to publication and usage by the salesforce.

Like Relationship analytics with similar opportunities I discussed previously, it needs some data. 30 qualified and unqualified leads. I utilised Mockaroo to create a thousand leads then used Flow to qualify and unqualify randomly. I detailed my work in the previous post, here. Once this is done, the predictive lead scoring is enabled.

Pressing Get Started now shows this screen.

I was expecting the ability to create my own model, but it seems it uses the default first. Once ready, it shows this screen.

Not really good accuracy so far…. Selecting Publish brings you to the configuration site.

Now, let’s add some fields that have been populated in the Lead to personalise the model. Select Edit Fields. The next screen allows you to define the list of fields that you think have an impact on the qualification rate of a lead. Here, the fields that were populated when I created Leads are used. Your mileage on which fields is part of the training of your model.

Once this is done, Click Retrain model for this new field selection to be used. After a period, the model is ready, and you can publish this model. Once published, the toggle for retraining automatically is available. This will, as it suggests, retrain the model regularly based on the fields selected. It will only publish the new version of the model it is better at predicting that the previous model

On the Lead form, there is a new section available, namely Predictive Lead Scoring.

This is telling the sales user that this lead is likely to be disqualified, and the reasons why. There is also a view that is made available, My Open Leads Scored which gives a visual indicator of the score for each. Great information for a Sales user to quickly identify the Leads worth progressing.

Opportunity Scoring

Similar steps can be completed for the Opportunity Scoring, with a Predictive Opportunity Scoring section added to the Opportunity form.

Top right of this section is a Chat icon, allowing a user to give feedback and provide a different score to the one given.

Premium Forecasting

Premium Forecasting is based on the standard forecasting but adds a prediction on the value of each opportunity dependent on historical information as well as data in the opportunity.

To establish forecasting, first, as an administrator, establish a forecast by choosing the Forecast configuration in the Sales App Settings section.

Add a Org chart Forecast. This presents you with a form to fill out.

Required fields are a name and the top of the hierarchy. Here, I have configured the management hierarchy as shown to reflect the data in my system.

Clicking on the progress bar at the top (great control!), leave the user security section as default and hit Next.

Include the Prediction field in the Layout and columns section. Here, you can add a column set - I found difficulty here, as the configuration of the values was not done automatically, I had to attach a column set and then use the option to change option set selection to choose the one I actually wanted, which would then show the Auto-configure columns option.

Select Next again, then activate the forecast.

After a while, a success message is displayed.

Once this is done, sales users will be able to select the Forecasts section for themselves. As you can see from the version for Veronica, the Opportunities in the forecast include the hierarchy. The bottom panel can also be used to quickly forecast data, such as the category, revenue and close dates.

The Prediction value in the top row is meant to show the revenue, including the prediction percentage, but I have found it hard to replicate the data I need to establish a value.

Sales Insights: Connection Insights

Mon, 18 May 2020 00:00:00 +0000

This is the fourth in a series walking through the Sales Insights add-on application. The series list is below for your reference.

Basic (Free) Setup - The “free” AI capabilities that can be added to any Sales Instance

Assistant Studio - Full capabilities - The full suite of options for Insight Cards

Productivity Intelligence - Activity and Contact suggestions as well as Notes Analysis

This post will be walking through Connection Insights, which includes Relationship Analytics and Talking Points.

Relationship Analytics

Relationship Analytics uses the data stored in D365 and Exchange, whether accounts, contacts, activities and emails. It then computes numerous KPIs for this data to establish how best to interact with an individual, leading to improved closure rates and customer satisfaction. There is also an analysis of the health of relationships to improve customer retention.

The first step is to enable Relationship analytics and health by clicking the box within the Sales Insights settings app.

Hitting Save (the 2 Data Sources options were not enabled for me) you get a message stating it is being activated.

After a short period, a success message is ready to use. At this point I have not changed any settings, including not enabling the Exchange Online configuration, as I would like to see what it is like without Email data.

For the sales user, there is a new View available against opportunities, namely My Open Opportunities by Relationship. On first use, this is not very impressive.

Relationship health is done by looking at the frequency and type of interaction with the customer regarding the opportunity. By adding in some data (this is a demo system so data is not available) the view becomes a lot more usable.

In this instance, data is added to the opportunity. You will also notice that a new section to the Opportunity form is has appeared. This section is giving me a visual representation of the health and indication when the next and last interactions were.

The KPIs are not populated straight away, it states it could be a couple of hours behind, so be patient. You will see the Relationship Health State (KPI) populated with the Health Computation in Progress which indicates it has enough data to provide a value.

Once calculated, the control on the form is updated

Further, the view is giving you a visual indication across your opportunities

Unfortunately, the data I entered wasn’t enough to make the relationship anything except poor! As you add more activity, this changes. I think on a 12 hour cycle, but can not confirm.

Also, now it is a little more insightful and realistic, let’s look at the chart that comes with the solution. Click on Show Chart and select Relationship Pipeline.

This chart shows all the deals that are due to close and plots them along with their relationship health. The bigger the circle, the larger the revenue. You can also hover over a point to see those key metrics. Pretty!

Digging further into some of the tools, along with the control on the Opportunity record, there is a new tab, Relationship Analytics.

This is a great, one stop shop to look at the activity on the opportunity and how it is improving over time. Top left is the same content you seen on the first tab. Next is the count of Interactions that have occurred on this opportunity, split between us, your employees, and them, your customers.

The top right shows the time spent on the Opportunity, a count of the hours spent in appointments or tasks, again split between us and them. Us is calculated by multiplying the length of the activity by the number of employees were part of the activity. Them does not do this calculation and is just a summation of the length of the activities where they were involved.

Email Engagement relies on the Email Engagement functionality discussed in a previous post. It shows all emails that were sent to and from the customer and how many were opened etc. Response rate measures the ratio of emails we sent to the customer which they responded to and vice-versa. This also comes from the Response time, which is the next chart.

Most contacted displays the contacts that have been included on meetings, calls or emails regarding this opportunity. This is duplicated for who has sent the most emails etc in Most Contacted By.

Finally, the Relationship Activities lists when activity has taken placed, the create date for activities over the lifetime of the Opportunity.

This tab is pretty useful and will come into its own when you have the data to support it. We all know the more you speak to a client the more likely an opportunity will be successful. The likelihood of a deal being lost where the customer continues to respond frequently and quickly is pretty slim.

Fine-tuning the results

Microsoft has provided you with a way of “tweaking” the results of the KPIs by rating the relative influence of each activity type. This will depend on your business and how it operates. For a B2C business, I would suggest that email or phone calls would be normal, and a meeting would normally be rare and infer a good relationship with the customer. B2B these would be different, with meetings more common.

Relationship Analytics with similar opportunities (preview)

There is a preview addition to the Sales Insights tools, namely relationship analytics with similar opportunities. There is a warning when you enable this feature that you need to have at least 30 each of won and lost opportunities to allow the new KPIs to be available.

In that spirit, I created a Flow to generate the data. The flow asks for a number, then creates that number of opportunities for each active account with that number of activities, a random mix of emails, meetings, tasks & calls against each opportunity. It then closes the opportunity, either won or lost, with a roughly 60% spilt for won.

You can find the flow as well as the other tweaks I made to my environment in GitHub here.

By enabling the preview, you have to agree to the terms.

On the Opportunity screen now, the Relationship Analytics tab is replaced with Relationship Analytics (Preview) and is significantly improved, comparing your opportunity with the historic data.

Talking Points

Small talk is a method of embedding you with a customer or lead. If they think you are friendly and discuss more than just the product or deal, then it has been proven to increase the likelihood of winning a contract. Remembering all your client’s insignificant details, and not confusing customers is hard work. Great salespeople make notes on each client and keep them updated. Dynamics eases this pain by scanning your email and highlighting to you details about the customer which you can use to improve your relationship with them.

These talking points are using the individuals’ email, which is private to them, so any points that displayed are specific to the individual sales user.

Back in the Sales Insights settings, an administrator can enable talking points and also which categories to use.

There is one more thing you need to do to enable this, and it isn’t documented. Thankfully Priyesh Wagh steered me straight in his blog article for D365 DeMystified. Thanks, Priyesh!

The relationship analytics solution adds a control to the contact form, namely TalkingPoints_section. This needs making visible by default, the form published an then you will see in the contact form a new section. This is populated with parts of an email I sent in.

Each of the icons in the bottom represents the categories that have been found in the email that you have with the contact. If you click on an icon, the relevant snippet is shown. Further you can choose to view this in carousel view, the default, or list view by clicking on the icon top right.

Lastly, by selecting the drop down arrow, the full email is displayed, where the user can click on the hyperlink to go to that email.

Sales Insights: Productivity Intelligence

Thu, 14 May 2020 00:00:00 +0000

This is the third instalment of my review/walkthrough of the Dynamics 365 Sales Insights application. Part 1 detailed the free features, with Part 2 detail the first of the premium features, Assistant Studio, including custom Insight cards and how to prioritise them.

This time, the blog is about Productivity Intelligence. This encompasses Auto capture (automatically bringing in content from Outlook and act on it, to reduce time entering data), email engagement (tracking email interaction from your customers) and Notes analysis (suggesting actions on the notes you make).

Auto capture

I have already discussed the basic version of this, so will just be going through the additions the Premium version brings.

Auto capture is not enabled by default when you enable the other Sales Insights features, toggle the switch

You have to agree to the terms and conditions as this is still a preview feature

Here you now get to choose a security role or allow all users to enrol. This is on top of the license requirement. I checked all the content to be captured for my trial.

As All security roles have been selected, the warning that basic features will be displayed is shown.

Once complete, each user will have to give individual consent as well, which appears in the timeline or in the Assistant like below.

After you select Allow access, the central panel now has an Activity Suggestions link.

Activity Suggestions

If an email comes into the users inbox now, a notification is displayed if the email address is associated with the account or contact.

At the bottom is a Show all link, which displays a simple popup before delivering the user to the Activity Suggestions page.

This is also available via the menu on the left

Back in the timeline frame, a couple of options are available to the user, highlighted below. The first dismisses the suggestion, removing it from the list of possible suggestions.

The next allows you to edit the activity, namely set the regarding to a different record

Hitting Save on the Edit screen shows this screen before the activity becomes part of the timeline.

The third option is just Save, which provides a prompt before it is created.

A small prompt is also shown when saving activities in this manner

Contact Suggestions

With the advanced features, you also get Contact Suggestions. This reduces the effort that syncing contacts etc. can be for your workforce. If an email is received from an email that is not within D365, a contact suggestion is created. An insight suggestion appears on the dashboard, which takes you to the full list on the left.

On the menu, there is also a Contact Suggestions menu allowing the user to go to their full list.

These contacts are ones that are not present in my instance and as such are prime to be brought in. Selecting a contact or the hyperlinked full name takes you to a quick create form, prepopulated with the information it has captured.

Save and Close will create a contact in the system. Hitting cancel returns you to the list, but the contact suggestion is removed. I have asked the question on the forums, as I don’t think this is correct behaviour, whether intentional or not. Will update if I get a response.

UPDATE: Had confirmation that this is a bug and it will be fixed by the end of the month! See, if you raise problems to Microsoft, they get resolved!

On the list, you also have options to either Save as contact, which will create using the default population or Edit and Save which will show the quick create form. Delete will remove this contact from the suggestions. If you select more than one contact suggestion, only Save as contact and Delete are available.

Notes Analysis

The final part of Productivity Intelligence is Notes Analysis. This uses sentiment analysis to capture information held in the notes against contacts, accounts or opportunities. From this information, suggestions are made with data retrieved from the notes.

Notes Analysis is enabled within the Sales Insights application once again. Toggle the switch.

So now, in the activity pane against accounts, contacts, case, opportunity & lead you will be offered suggestions to create other records from the content in the notes. This is by providing a hyperlink in the note itself.

Notice how the notes are read in direct notes and emails. Clicking on the link provides the user with a suggestion.

As you can see from this suggestion, not only has it determined that I should have a meeting planned for Tuesday, but automated the date entry from my “next tuesday”.

If I hit Edit and Create, a Quick Create form is displayed, prepopulated with the detail.

The only issue I have here is that the note, when I return to the Timeline, still appears like I have not done the previous step and created the activity. Some suggestions allow a quicker create process, if you don’t need to change any of the detail

Edit and Create displays the prepopulated Quick Create form. Create just accepts the defaults and creates the record.

Note Analysis is a great tool to the arsenal of the busy sales person. All of these tools are trying to reduce the time spent doing administrative tasks and push the sales person to be more active with their customers.

Sales Insights: Assistant (full capabilities)

Mon, 11 May 2020 00:00:00 +0000

In my previous post, I walked through the configuration of the standard, “free” features of the Sales Insights offering for D365 Sales. This time, I am going to start on the premium features covered by the £37.70 per month license fee.

If you look at the Sales Insights administration screen, the list of the capabilities is quite hefty. For this post, I will concentrate on the Assistant, Sales Insights cards and prioritising them.

To get going, hit the Try Sales Insights button. This launches the screen below. Read and agree to the terms and hit continue.

Once done, you are presented with the install screen.

Once installation is complete, a message appears and the Go to Configuration is now available. This button takes the user to the Sales Insights Settings portion of the application.

As you can see there is a lot more settings available to the administrator now. Lets start with the Insight Cards available in the Assistant Studio.

Assistant Studio

The advanced Assistant Studio is a separate display than the “free” version I mentioned on previous post.

The first thing you will notice is the ability to make your own. This utilises Power Automate (and from there, you could rule the world) to define a schedule or trigger to run and produce a card for display.

Each Insight card has a security role selection now.

Also, on the left is an ability to tweak the ranking of the cards, which ones appear higher in the view. This is tweaking the inbuilt sorting in addition to being able to make the card a priority. It is pretty limited, only 4 ranks available, but I would think that this is more than enough for most scenarios.

Creating a Custom Insight Card

Though the most common scenario would be using Power Automate to create a custom Insight, it is possible to go further by using cards created as XML and import them into your environment as a solution. The card is triggered by using the Web API. This will require a blog post on its own to get to the details, but for now, here is the link the Microsoft tutorial on the subject. As you are calling the Web API, when and if these cards get created is open to a whole manner of scenarios - alerting from a finance system or bespoke service offering, or even telemetry from your own app or licensing solution is possible.

Thankfully, creating cards via Power Automate is a lot simpler. I will walk through a scenario that is applicable to a Sales team. Everyone has competitors. Competitors all have blog rolls where they publish information on their products and views. It would be good to highlight these new posts to your Sales team, so they are aware of the latest developments prior to discussions with their customers. You can have a trigger for when an RSS feed changes, but this would mean you would have a flow for each blog. Adding a competitor blog would be time-consuming and in-effective.

My design would be to on a daily schedule, loop through each of the competitors stored in the D365 environment with a blog, read any articles that have been posted since the day before and create an Insight card for each.

I created a new field on the Competitor entity to enter the blog address.

I have put it on the Competitor form and filled it with my blog for testing.

My trigger is a schedule, defined to run every day at 7am.

Then, get all the competitors that have a blog listed.

Each Insight Card that you create needs to have a user that you will show the insight for - something that flummoxed me for a long while. I would have thought that some cards would be applicable to the whole organisation, or a group of users. You can assign a security group for a card, which I will show later, so would assume it would create an active card for each member of the security group at the least. (I created an idea here).

To combat this, retrieve the Team based on a name (to allow for moving between environments)

Put this into a variable for use in the next query without a loop action.

Then, query the team memberships table. The user id is the field needed so this is easier than drilling down to the user table.

Convert this to an array so the flow can quickly use the values when creating the Insight card.

For each of the blogs, get the posts that have been created since yesterday. If there is any, loop through the articles then for each user create an Insight card.

With this information, we can post our card, firstly find the Sales Insights connector, and choose Create card for assistant V2 (preview)

When you add this for the first time, you are prompted to log in and consent for the connector. Lots of permissions needed.

Finally, ready for creating the card. Here, properties from the blog roll and the Competitor are used.

There are other properties that could be applicable to your Sales Insight, such as Card behaviour after click (whether to dismiss\snooze\do nothing to the card when the action is clicked), which entity and record to display the card on (remember this is only for core sales entities so far, namely Account, Contact, Lead and Opportunity) and display dates for the Insight.

The Button type is an interesting selection. I have chosen open URL, but there is also Custom Action (CRM Process), Launch playbook, Open Record and Open URL. I used Open URL for my demo, will describe the others later in the post.

The Show for is populated with the following. This gets the current user Id from the teams membership array.

1

items('Loop_users')?['userId']

The URL is populated with a link to the article, but as this is an array in itself, take the first one using the code below

1

first(items('Loop_new_articles')?['links'])

Once this code is run, you should create your first custom insight card, which appears on the dashboard like below

Drilling down into the note gives you more detail and allows the action to be taken.

Button Types

As described earlier, there are 5 actions available in the button types.

Custom Action (CRM Process)

Custom Action is a relatively new one. It relies on a pre-created action in the Flow that defines the action. This functionality allows the administrator to configure triggering of a workflow or plugin, as long as it is against an entity.

For example, here the action is defined to use the Opportunity custom action CloseAllOpportunityQuotes, define which record it needs to run against and also the quote Status.

The parameters are defined by the action within the D365 customisation.

This new action is configured in the Sales Insight card creation step.

Launch Playbook

Playbooks are a relatively new feature within the Sales app, which allows a sales manager or administrator to configure a pre-defined set of activities, or a play, to define the perfect way an interaction with a customer should work. You can define them on one or more of the core sales entities.

On a Sales Insight card, establish which Playbook template you want to use, define which entity to apply it to and which record.

Open Record

This is a simple one, launch a record when the user hits the button. Define the entity and the record Id.

Open URL

Another simple one, I used it in the walkthrough above. You need to define the URL to open, here I link back to the first article link in the RSS feed.

REST

This button type allows interaction with an API endpoint. Define the endpoint, which method, though I am not sure what use a GET would be. Also, define the Headers and body. This allows the user to push content or start activity in external systems, logging a case with ServiceNow or creating an activity in your own bespoke system for example. You could also use this to trigger a Flow and create an interaction via that route.

Sales Insights: Basic set up

Tue, 05 May 2020 00:00:00 +0000

I have been working for quite a while on Customer Insights, running workshops and generally getting to know it, but I realised that there is a multitude of products in the Insights suite, Customer Insights, Customer Service Insights, Product Insights, Marketing Insights and Sales Insights. These tools each have their own niche, and I don’t know them at all well. So I thought I would educate myself and hopefully someone else by walking through configuring each one and writing down my experiences.

The Business Case

Dynamics 365 Sales Insights to use its full name is designed to help your sales managers improve their team’s performance. It does this by continuously analysing the interactions a salesperson has with a customer, whether in D365 Sales or Exchange. Each business relationship is evaluated to understand the activities that are completed and how previous interactions have affected sales.

Do you know how often a salesperson contacts a potential lead? Do they always use canned responses, or is there a form of personality on the interaction? Have you provided your sales team with actionable insights about their opportunities?

Dynamics 365 Sales Insights has a selection of Free and Advanced features which improve the view of the customer, enabling rich customer experiences in the interactions with your sales team. It uses AI modelling to constantly monitor these interactions and provide insights to your salespeople at the appropriate time, increase lead closure and customer satisfaction.

Free vs Advanced

As already described, there are several free tools to start your journey, which are part of the standard Dynamics 365 Sales offering, namely Assistant (was Relationship Assistant), Auto Capture and email engagement. This first part of the series will walk through configuring the free features.

Advanced features is a £37.70 per month per user add-on, which is a hefty cost, but with it, you get the full suite of products - Assitant with Studio, Relationship Analytics, Predictive lead and opportunity scoring, Notes analysis, Talking points, Who knows whom, Sales accelerator, Assistant for teams and Conversation Intelligence. This is a lot of functionality. Over the next few posts, I will try and educate myself and step through the experience.

Configuring Sales Insights Free features

I started with a brand new trial, again, the best way to do this is to follow along with Chris Huntingford here. I chose all apps, as usual, then launched the standard Sales Hub. Down the bottom right, is a new area Sales Insights Settings.

Assistant Configuration

The first free feature is Assistant. Clicking on the Manage link takes us to the Assistant Studio to manage the cards that are available to the end-user.

You are pretty limited to what you can do on this screen, as it is the Free version. All you can do is enable/disable specific cards & make them a high priority. Each card description is vague, but there is a reference from Microsoft here.

For instance, I have enabled Recent Meeting and Recent Meeting (Exchange). I have then added a meeting both within D365 and within Outlook for Veronica Quek. When she opens her sales dashboard, she gets this view.

This is different than the view if you have not got the 2020 Wave 1 functionality enabled.

For the Reminder, as it is just passed, I get an option to add some notes or open the appointment.

Also, clicking on the Assistant button top right gives a drop-down for the activities I have set up.

You will notice that the user has the option to tell the AI whether the alert is useful or not, delay or cancel a notification. The valuable data is available to the administrator (and Microsoft) appearing alongside the detail of the card in the admin section.

Some cards have options for the administrator to set, such as the Close date coming soon card. This allows an admin to specify how many days before a close date will the owner is notified via a card on the assistant.

This pops up in the assistant like below.

All these action cards are great to bring outstanding tasks or stuff to the attention of your salespeople. Some of them will be reliant on having a full license.

Auto capture (free)

The next free component of Sales Insights is Auto Capture. To enable this, you have to agree to the Terms of service, which will enable Auto capture.

Clicking on Manage opens the Auto Capture configuration.

You can see that the Premium is now tantalising close, but let’s limit our selves to the free version. This requires Exchange Online sync set up and mailboxes configured. Plenty of articles out there stepping you through that. The insights from this don’t seem to come through straight away, I set up on a Friday and did not see any insights, by the Monday, it was all visible. Maybe that is just a glitch in the system, can not see any reference to this.

I sent 2 emails to my demo account, from an email address already belonging to a contact in D365.

These are emails I would expect a salesperson to action, the contact is explicitly asking for something related to a possible sale. In Sales Insights dashboard, there is a new notification.

If I click on the first part, it opens a preview of the email where you can open the email direct in Outlook online.

On the contact record you see the insights again and also the email that is untracked is highlighted. In the Timeline, the highlighted button allows you to track the email straight away without returning to Outlook.

What cards appear depend on the insight cards you enable on the Assistant studio, discussed earlier. Anything that has a mention on your inbox is reliant on the auto-capture functionality being configured.

Email Engagement (Free)

The final “free” Sales Insight is Email Engagement. Check the Microsoft page on it here for full details. This tool allows you to be notified when the recipient opens an email, tr

Firstly, we need to grant permission by selecting the grant permission link in the Sales Insight Admin.

This brings up the usual Permissions Requested approval box.

Once accepted, you should be presented with this page.

Now, back in the Sales Insights configuration page, the Email Engagement (free) has changed to a Set up link.

Clicking on Set up presents you with a page where you enable the functionality.

If you want to track attachment opening etc. you also need to configure OneDrive integration for your environment. In effect, any attachments are stored in OneDrive, and a link to that is sent to the user rather than the actual file. This allows the logging of attachment interactions. Navigate to admin.powerplatform.com, open your environment and Settings. Click on Integration / Document Management Settings, which will open this page.

To enable OneDrive, you need SharePoint enabled, so click on Enable Server-Based SharePoint Integration. Walkthrough the SharePoint Integration dialog, choosing a site to host your D365 files. Then click Enable OneDrive for Business back in the Document Management Settings page.

When I did this, I came across the issue that I needed a default SharePoint Site, but to set the default you need to change a setting which is not on the form. Workaround here. Once that is done, you get a new part of the Document Management Settings.

There is also a bit they don’t tell you about in the config, namely to enable Email as a document storage entity. If you go into SharePoint Document Locations under the Document Management Settings, you will be presented with this screen.

Make sure Email is ticked (along with the others), enter a path to the SharePoint and select Next. I decided to link it back to the Account so selected that as a based on entity then hit Next again. After a little churning, this is the result.

When you add an email now, you get a new tab (with the Enhanced Email Wave 1 enabled).

In the old way, you get a section available on the right

There are a few additional features enabled with this, firstly Recipient Activity

Recipient Activity

To record email activity, a small 1-pixel gif is placed in the email as you send it. This will, when the email is read, notify you that the email has been read. You can toggle this on or off per email using the Do Not Follow functionality in the Email Engagement section. There is also an override per Contact in the contact preferences section.

This is a snapshot of the HTML email that I sent, you can see the embedded gif which is a link to the Microsoft tracking site.

In D365, a notification appears as a new Insight card against the contact and in the dashboard view. It also appears in the timeline for the contact that was sent the mail. The stats and alerts are not fool-proof, each email seems to be opened multiple times, which is not my interactions. I assume that this is due to various services such as spam filters etc. checking the links, but it does give you confidence that there has been an interaction of some sort.

Attachment Activity

In addition to opening the email, you can also track them opening any attachments you send too. Firstly, you need to save the email before you can add attachments. Then add an attachment. You should get a new Follow button once you have added the attachment.

Selecting the Follow button tags the attachment for following and it appear in the grid of attachments like below.

Now, when the recipient receives the attachment, it is a link to the OneDrive file like below.

Clicking on the link drives the recipient to a OneDrive page where they can download the file. Please note, that these files are not secured, you don’t need to log on to get at these files, only the link. It would be pretty random that you could determine the link, but it is just a warning. Normal sharing would be preferred for contracts etc. but it is great for marketing/sales documents.

In D365, once the file is opened, you see a Sales Insights notification and if you open the email in D365, activity gets logged like below.

The attachments are stored on the users OneDrive, which you can see below for Veronica’s OneDrive.

Link Activity

Another feature in Email Engagement is link tracking. Any link you put into your email gets replaced with a link back to the D365 server. Each time the user clicks on the link, it is logged in D365, before the recipient is redirected to the original URL.

For instance, if I link to my blog in an email

The recipient receives the email like below, but if you hover over the link it shows the full link that the user gets.

If I click on the link, the user gets directed to my awesome blog. More importantly, the link tracking appears in D365.

Delayed Delivery

When you send an email, with Email Engagement on, you can delay it. Click on Delay Send option presents you with this screen

Hit send this time does nothing, with D365 in the background sending the email out when the time arrives. A nice little feature to pre-plan engagements.

Follow-up Reminder

Another nice little quality of life feature is the ability to add a follow-up reminder. This can be done at any time after the email has been sent.

You can choose the condition and the time frame as well as a comment. This appears as an Assistant card once the date has reached.

Email Templates

Another feature enhanced with Email Engagement is templates. You can select templates by selecting Insert Template in the Email form. As you build up tracking on these emails and they are interacted with, statistics are built upon the suitability of each template - which ones ended up with an email that was opened for example.

Next time

As you can see this set of tools, including in the subscription cost of a Sales license is pretty cool. In my next post I will delve deeper into the premium features.