Tag: analytics

Omnichannel & Sentiment Analysis

In general, it’s usually quite useful to be able to see how customers are engaging with your company, and how they’re feeling about things. If customers are disgruntled, annoyed, or complaining, it’s important to be able to understand the root cause/s of their issue/s, and resolve them as soon as possible.

One of the tools available in Omnichannel is Sentiment Analysis. What is this?

Being able to identify how customers see/interact with your brand, accurately, is vitally important. Using people to manually trawl through your data to attempt to identify this has many drawbacks:

  • Lack of consistent approach
  • Large amounts of time needed
  • Many manual touchpoints

As a natural follow-on from this, being able to identify & categorise the sentiment in customer communications through using machine learning can unlock many business use cases that can then result in immense value for your company.

Microsoft provide the ability for this through Azure Cognitive Services. It’s really quite interesting in how this actually works. You can go to https://azure.microsoft.com/en-gb/services/cognitive-services/text-analytics/, put in a sentence, and see what results come back. It can be quite amusing to see what different colours come out as!

As part of the analytics around chat (and by chat, I’m not referring to just a chat bot – anything within Omnichannel can be referred to as ‘chat’, from an agent perspective), sentiment analysis can be used.

This is quite easy to set up. To do so, open the Omnichannel Administration Hub, go to the Settings area in the left-hand menu, open ‘Sentiment Analysis’, and click to enable it. Remember to save it to apply it!

This will then result in the agent interface showing the following:

Now, this isn’t static. The sentiment will update in real time as the conversation continues, and will change based on what the customer is saying.

Now, obviously we’d expect agents to be able to judge the tone of the conversation based on what’s being said (at least I’d personally expect it). So for this, the sentiment that shows within the chat isn’t that helpful.

However, it does come into its own in a slightly different place. This is the Omnichannel Sentiments Analysis Dashboard, which is served through PowerBI.

Through this, supervisors can understand how their company is measuring up to their KPIs & necessary trends. They can also understand the overall support experience that omnichannel is having, along with tracking the sentiment of customer interactions. As a result of having this to hand, better understanding of customers can take place, resulting in improvement of the overall customer experience.

Once the dashboards have been configured within PowerBI (I’m going to do a separate post on this), it’s then possible to surface these within the Omnichannel Customer Service Hub (which users with the Supervisor role will be able to see). This means that supervisors won’t need to open a separate place to see these; it’s all available through the same interface.

There’s also a more detailed view into what’s actually happening, through the ‘Omnichannel Insights – Sentiment Analysis Report’. This displays a lot more information, drilling down & splitting the data up into agents, queues, channels & trends. Here’s an example of this:

With all of this information as the fingertips, it’s now really possible to drill down into the details. Through this, we’re able to carry out full & proper analysis on what’s actually causing customer interactions. From looking into what’s occurring, it’s then possible to review the current state of things, and see what can be improved. This will then result in more positive sentiments shown by customers, and drive their loyalty to the company!

Channel Integration Framework (II)

Last week I shared a post about the Channel Integration Framework 1.0 ( https://thecrm.ninja/omnichannel-the-channel-integration-framework-i/), where I introduced what CIF actually is, and how it was originally launched.

I was somewhat naughty – I ended the post on a slight teaser note. See, as I mentioned there, CIF 1.0 was all about allowing channel widgets from providers to integrate with what we refer to as ‘model-driven apps’. Eg Accounts, Contacts, etc, which are all ‘single session’ entities within Dynamics 365 (ie that a single channel is being used, like a phone call). It was never designed to support ‘multi session’ apps, which is what Omnichannel is all about (ie the ability to have multiple ‘conversations’ going on at the same time, regardless of the channel in which they’ve come in on).

This is why in 2019, Microsoft was working on the next version of CIF – 2.0! Imaginatively named, of course. The aim was to take the general concepts from CIF 1.0, and apply them to be able to be used with multi-session applications. Obviously at this point in time, the ONLY multi-channel application is….OMNICHANNEL!

Incidentally it’s important to note that any communication widget built on CIF 1.0 will continue to work. Microsoft is not removing the 1.0 APIs, as they’re still needed to support interfacing with single-session apps

So, what’s new (and improved) with CIF 2.0:

  • The ability to have multiple communications happening at the same time, through either the same channel, or multiple channels
  • The ability to have multiple third-party provider solutions. You could have Provider A for telephony, and right alongside it Provider B for SMS
  • Different modes for the widgets, which can be adjusted per session launched. It’s now possible to have them set to be docked in the interface, minimised to a smaller size (to give agents more screen space to work with) or hidden (where it’s running in the background). Agents are able to switch between docked & minimised modes

This is all really exciting. It will allow companies to mix and match solutions based on their actual requirements, rather than having to settle for a single solution provider that may not actually be everything that they’re wanting.

Note: With an eye to data security and things like GDPR, information, data & events for support sessions will only go to the provider for the channel that the session is coming through. They don’t have access to any other session/s that are happening

One of the other main features that comes with CIF 2.0 are the channel analytics. With CIF 1.0, all of the interactions are surfaced into the Dynamics 365 app, but are actually running on the provider’s system (in the background). Data can of course be exchanged between them, but there’s no real ability to perform analytics of what’s going on (especially as it’s only a single channel)

Black Samsung Tablet Computer

In order to get the full insight into what support agents are actually doing, along with seeing the performance of the support centre, it’s necessary to be able to see information across all of the following:

  • CRM data (itself). This covers the actual data of the customers, communications with them (eg activities and cases), their history over time, etc – all of this is stored in the underlying CDS
  • Agent Behaviour. How the agent handles the session – what they do before, during and after they’ve helped the customer
  • Communication data. How the customer has contacted the company now, how they’ve done so in the past, their experience, etc.

So how exactly does Channel Analytics help with things? Well, what it does is:

  • Provides APIs to bring the conversation data into CDS
  • Along with the APIs, it has a standardised schema for how the analytical data will be stored
  • Brings in a standardised scheme for how all of the data points shown above will be correlated together
  • Extendable – you can bring in your own KPIs and other data to best suit your needs.

With all of this now being available through CIF 2.0, it’s possible to carry out really advanced analytics, crunch the data using a tool such as PowerBI, and other things. It’s also possible, of course, to configure real-time dashboards as well as alerts to cover any issues that may come up.

Benefits of this of course include:

  • Identifying root causes of anomalies, help with audit reports, and tracking KPI’s
  • Help with predictive insights for volume of requests, and where to best focus resources
  • See customer satisfaction & average handling times to resolution, along with on-going customer sentiment

So in summary, this is really great. I’ve actually now started to speak with different companies who have ISV solutions that are going to be on CIF 2.0, and will be writing about them in the near future as well.