A complete introduction to Twilio Flex Insights

What is Twilio Flex Insights?

Twilio Flex Insights is a contact centre reporting tool that enables contact centre supervisors, managers, and data analysts to create, modify, share, and consume reports and dashboards for interactions handled by Flex.

It provides historical reporting and insight into conversations between agents and customers, with the ability to drill down from top-level key performance indicators (KPIs) to individual conversations.

There are numerous ways to represent data in Flex Insights reports such as tabular data, bar charts, pie charts, and many more different options.

Configured and displayed in the right way, organisations gain a complete overview of customer behaviour and can deliver much better customer experiences as a result.

Accessing Flex Insights

Organisations must be on a paid plan to use Twilio Flex Insights. Provisioning a Flex Insights instance requires upgrading, clicking to enable, and entering details (company name, contact and timezone).

When Flex Insights has been successfully enabled, additional icons become available on the left menu.

Left: Account without Flex Insights. Right: Account with Flex Insights (additional icons).

How Flex Insights works

Flex Insights gathers individual events and maps them onto a business-orientated analytics data model to provide a 360-degree view of every conversation in a Flex contact centre.

Above: High-level overview of the pipeline.

‘Conversations’ are the backbone of Flex Insights. A ‘Conversation’ is defined simply as the communication between one customer and one or more agents.

Every conversation consists of one or more ‘Segments.’ A Segment represents an interval in time dedicated to certain phases during a Conversation.

‘Segment’ examples include:

• ‘Queue Segment’ (time from when a customer entered a queue until they left it)

• ‘Conversation Segment‘ (time from when a customer is connected to an agent until the communication ends or transfer occurs)

• ‘Missed Conversation Segment’ (an unsuccessful attempt to connect a customer to an agent i.e. the agent didn't accept the reservation and it timed out)

• ‘Rejected Conversation Segment’ (reservation explicitly rejected by an agent)

• The ‘Agent Status Segment’ (agent's activity state in a given time period)

Conversations can be segmented and filtered by the agents (those interacting with customers) who handle them. They can also be segmented and filtered by the demographic attributes of customers.

The key ‘metrics’ and ‘attributes’ in Flex Insights

Organisations can choose the metrics and attributes by which to measure.

Examples include:

• ‘Abandoned Conversations’ (the number of conversations where the agent and the customer have not talked to each other)

• ‘Abandon Time’ (the time a customer or an agent spends waiting on the other party before giving up)

• ‘Communication Channel’ (communication medium used between the agent and the customer, such as voice, chat, SMS, etc)

• ‘Conversation Attribute #5’ (in the Conversations metric, in addition to various other properties, users can also customise 5 numeric values [called measures] and 7 text values [called attributes]. Of those 7 attributes, for example, conversation_attribute_5 is auto-assigned the Task SID if not customised)

• ‘Handled Conversations’ (the number of conversations processed by an Agent)

• ‘Handling Time’ (how much time agents spend working on behalf of a customer. This includes all the work related to the conversation, including wrap up time)

• ‘Conversations’ including Abandoned’ (all incoming calls, including handled calls as well as abandoned calls)

• ‘Queue’ (queue in which the customer was waiting before reaching the agent)

• ‘Queue Time’ (how long the customer waited before reaching an agent, including both ‘Queue Time’ and ‘Ring Time’)

• ‘Segments within SLA’ (the count of inbound segments handled within the target waiting time, excluding short, abandoned segments)

• ‘Segments within SLA %’ (the ratio between segments within the SLA versus all offered calls)

• ‘Service Level Agreement’ (the key goals of a contact centre, which are the same for all customers, across various queues and channels)

• ‘Talk Time’ (time the customer spent talking with an agent)

• ‘Waiting Time’ (same as ‘Queue Time’)

• ‘Wrap Up Time’ (time agents spend completing additional tasks after a conversation with a customer)

'Available out of box' vs 'Needs to be provided by the customer'

Within Flex Insights, organisations get a huge amount of content, reports, dashboards and metrics that have been created through extensive experience of working with contact centre customers all over the world. They gain an immediate understanding of their contact centre’s performance.

Out of the box, Flex Insights captures and reports a number of different metrics including ‘Abandoned Conversations’, ‘Handled Conversations’, ‘Queue Time’, and more. Users can also add custom data Tasks and worker attributes to collect the data they need.

The Service Level Agreement (SLA) is hard-coded to 60 seconds. If the customer has a different SLA, a custom metric needs to be created which calculates the SLA using the customer’s own requirements.

‘Interactive Voice Response (IVR) Time’ is not generated out of the box. Flex Insights reports rely on interactions for which a Task gets created. If organisations have calls that are abandoned or get resolved in the IVR, the default reporting solution does not report on it.

IVR Options selected are not generated out of the box. However, visibility on how customers interact with an IVR is critical for many. A Twilio guide that shows how to use Flex Insights to consume IVR interactions for inbound calls and provides tips on building custom reports can be found here. We’re also very happy to help.

A customer identifier, i.e. customer number, is not automatically generated. Flex Insights references each conversation with a customer based on a phone number or other contact information. Additional task attributes can be set to provide details about the customer.

Skills are not auto-configured to route the interaction. Basic skill-based conversation routing to multiple queues in a Twilio Flex based call centre can be set-up.

Dashboards

The core of Flex Insights is the dashboard. In the Dashboard function, organisations can access pre-built sets of dashboards and customise or create their own reports. They also have drill down capabilities such as listening to conversations or reading transcripts.

The dashboard can display a collection of reports, providing a bird's eye view into different aspects of a contact centre. Users are able to customise how data is pulled and displayed.

Flex Insights provides ‘canned dashboards’ to give customers inspiration. Each dashboard has a large amount of filters.

For example:

• The queues dashboard
  
  Shows how traffic is being handled by queues. Users can see the SLAs, the average speed of answer and the abandon rates. They can see these metrics broken down by date, by queues, by external contacts, and get an immediate understanding of the SLA metric values, speed of answer, abandoned times and rates, abandoned volume, as well as minimum and maximum wait times.

Every dashboard can be customised to create a custom version. Dashboards and reports can be built easily using drag-and-drop tools. No coding skills are required.

Because Flex Insights is tightly integrated with Flex, any metadata and any attributes that are attached to each task can be displayed in the reports.

Speech Essentials

With customer permission, audio recordings are processed to provide ‘speech essentials’ to provide a better understanding of agent and customer interactions.

For example, Flex Insights can tell organisations how much crosstalk time and silence time there was in a call.

The ‘Handling Time Dashboard’ shows the high-level numbers of handling time broken into days and a combination of ‘Talk Time’ and ‘Wrap Up Time’. Digging deeper into the dashboard reveals individual queues or teams as well as agents’ handling time. The team leader can review individual calls, organising tables as needed. For example, looking at calls with significant long talk times, or long silences, which they may wish to understand more about.

Whenever viewing a list of conversations between agents and customers, users can drill down into the recordings by opening the Flex Player. The agent speaking is marked as one colour, while the customer speaking is marked as another. The times when agents speak over the customer are highlighted as are extensive silence times. These are the type of things team leaders may like to know more about. They can also comment on the conversations, providing notes and feedback.

Building your Flex Insights

There are many ways to represent data in Flex Insights, and lots of out of the box, pre-built dashboards to provide inspiration. Once organisations nail their analytics, they gain unparalleled insight into their contact centre(s). From topline core business KPIs all the way down to evaluating individual conversations, which facilitates the coaching and continually development of agents.

Because custom data Tasks and worker attributes can be added, the platform can be set-up to provide insight on whatever is needed.

Flex Insights is a solution that’s made for the omnichannel environment. Like Flex itself, it brings everything together in one place providing invaluable context across operations, rather than siloed analytics.

Of huge value is establishing metrics that give a sense of the overall impression customers have of a company’s customer experience. These differ between organisations but are often a mix such as queue time, call time, resolution rate, and more.

For example, one organisation has used Flex Insights to measure improved CX in the form of First Contact Resolution – from 2.2 times per topic down to 1.2/1.3. When such metrics are established, organisations have a baseline from which to measure the performance and effectiveness of their overall contact centre services.

‘Hero’ insights

We call insights that can be acted upon to make big operational efficiency gains, ‘Hero Insights.’ These can be both real time triggers, such as long queue times triggering an alternative flow whereby more information is gathered in the Interactive Voice Response (IVR), and strategic insights such as abandonment rates for AI compared to human interactions, which reveal attitudes to, and performance of, AI.

Zing support

If you need assistance with Flex Insights, whether it’s the initial set-up or refinement, we’d be delighted to help. More than just builders, Flex Insights allows us to be ‘enablers of discovery,’ helping our clients listen to what the data is telling them and iterate to excellence. Get in touch!