Decision automation is widely considered to be the engine room for Intelligent Automation.  It involves the automation of corporate decision making, policies and customer interactions. This capability can drive complex customer facing processes such as selling and technical support. Decision engineering is the process of digitally capturing knowledge that may be intuitive but not necessarily accessible to the business expert, and automatically generating complex decision flows. Decision Engineering can be applied across both front and back office, and cover expertise such as technical support, need analysis, selling, risk assessment, compliance, customer negotiations, and retention.

Decision engineering involves three main types of tasks:

  1. Decision mapping

Where there are documented policy/compliance rules, troubleshooting/diagnostic flows, or standard operating procedures then these can be mapped directly into the Decision Engine using the Decision Tables, Decision Trees, and no-code formulas capabilities.

Below is an example of mapping into a decision tree the documented regulations for determining if planning permission is required by local authorities for a household installing a satellite dish.

    

  1. Decision mining of tacit expertise

Almost all decision automation platforms on the market support Decision Mapping only. This is a real constraint as much of the expertise in areas such as risk assessment, selections, diagnostics etc. are tacit in that the subject matter experts can make complex decisions but are unable to articulate the logic behind such decisions as required to automate. We overcome this constraint by supporting a unique decision mining technology and methodology that can induce tacit knowledge from examples of decision making supplied by experts.

In the example below, an expert financial advisor used his tacit knowledge to define the table of examples below describing the type of Trust required for the various customer circumstances and requirements:

From the above table of tacit examples, our Decision mining engine automatically induced the decision tree below that can be used to automate the Trust advice process.

  1. Decision mining using machine learning

Many organisations have data that can be used to model the propensities of risk events such as churn, credit etc. Such predictive models can be used to tailor/customise the conversations with customers by considering such risks. Whilst deep learning can be used for such purposes, its lack of transparency and auditability can present a real adoption challenge for most business users. Viabl.ai overcomes this by supporting machine learning that generates transparent and auditable decision trees from data.  This has the added benefit of using the same decision representation for both Tacit expertise from experts and decisions learnt from data.

The decision tree below describes the propensity of a motor insurance policy to give rise to a fraudulent claim and is derived automatically by machine learning from historic data about policy details and fraud.

Delivering on Intelligent Automation

Digitally capturing expertise and knowledge is only half of the equation. Being able to act upon and optimize complex workflow processes is the other half.  Once expertise is digitally captured decisions can be executed using XpertRule’s Viabl.ai Decision Engine. The software with its integrated Conversational AI capability can integrate with any customer contact channel including web, chat, voice and mobile. Decision automation reduces operational costs whilst increasing customer satisfactions by:

  • Empowering every contact centre agent with the expertise of the best experts and with a 360-degree view of the customer whilst engaging in complex and tailored conversations. This leads to reduced calls handling times/repeat calls, and increased customer satisfaction as a result of receiving accurate resolutions quickly, without escalation and without having to make repeated calls.
  • Diverting calls to customer self-service channels that can handle complex conversations at a far lower cost. In our experience in a complex technical support contact centre in a telco, it was possible to successfully answer 12 percent of call using self-service.

This approach to powering complex customer conversations through conversational decisioning is radically different to all the other conversational AI & digital assistants on the market.  It allows organizations to achieve a degree of automation in customer engagement not previously possible (Fig 1.).

By integrating advanced decision engineering with Conversational AI capabilities, our Viabl.ai platform is uniquely positioned to deploy back-end data and decisions as a fully integrated function of the customer engagement. Capturing documented, tacit and machine learnt expertise and using this to drive complex conversations with agents and customers, reaches into realms of automation that were not previously possible. It also allows businesses to extract more value from existing attended automation solutions by upskilling the front office employees to handle more complex cases rapidly, cost effectively and consistently.

We’re bringing our Viabl.ai capabilities to market by partnering with Emergence. We’ve integrated the capabilities outlined above with Emergence’s new Decision Maker offering.  Decision Maker uniquely combines the Viabl.ai platform with consultants to bring customers not just the technology they need to deliver better customer and agent experience but crucially to identify where and how to successfully deploy decision mining in their business for maximum impact.

 

Next reads:

Sign up
to TechKnow.

Sign up to TechKnow to get all our latest news and jargon-free insights, straight to your inbox. 

By clicking subscribe, you accept our terms and conditions.