Layers of Process Driven AI for Customer Success

Updated: Aug 1

In business, process is everything. Processes are inextricably tied to operational efficiency, profitability and the quality of the customer experience. The more efficient a business’s processes are, the faster and more agile it becomes. That’s why today’s largest and most successful companies obsess over their processes. It’s not surprising why – process improvement is estimated to boost success rates by 70%.

Yet visibility is a necessary condition for efficiency. To improve processes, you first need to be able to see and quantify them. Only then does it become possible to make the best intervention. Having this level of process observability delivers great benefits: business leaders say it helps them to rapidly identify and resolve issues (38%), speed up development (35%), and lower operating costs (32%).

However, process visibility and improvement depends on good data that covers the entire lifespan of the process. Change and transformation leaders need the right tools to understand their processes from end to end, and extract quality data from them – whether they are structured or unstructured.

Process Mapping: Laying the foundations

A process can’t be understood until it’s been modeled. Until then, it is only a collection of actions or events. Creating a process model is the act of tying these events together in a sequence, understanding how each stage relates to the other. Only then do you know where to look for process efficiencies and improvements.

Traditionally, this modeling was performed manually through process mapping. This is the process of interviewing users and stakeholders in a business to find out what processes exist, who is responsible for them, and where the blockers are located. Process mapping is a very useful exercise, and it can sometimes be the only option available for understanding a business’s processes and how people interact with them.

The drawback of this approach is that it is both time consuming and highly subjective – based entirely on stakeholder opinions. In large organisations with hundreds of people and processes, it has the tendency to produce models that end up as convincing fictions.

Process mapping often represents processes as people see them rather than what they really are. People have the tendency to oversimplify when explaining things, and that means missing out key parts of processes and forgetting about exceptions and error handling. Models can also be subject to implicit bias, as employees tend to view the parts of the process where they have the most control as more important, and downplay the steps that don’t involve them.

Process Mining: Explaining the ‘how’

When it isn’t paired with hard data and misses key steps, process mapping can be misleading for decision makers. This is a huge wasted opportunity when many of the most important processes are structured extensively in data. In any modern enterprise, actions taken by employees and bots are automatically recorded by the business’s various IT systems. This data represents a treasure trove of insight to help understand processes and augment their efficiency. Yet extracting all of this information and putting it into a useful format would be extremely costly for the business.

Fortunately, Process Mining greatly expedites this process. Process Mining tools take the structured process data that exists in corporate IT systems and converts it into useful event logs – which summarise key information such as timestamps and user IDs for deeper analysis.

This serves two purposes. Firstly, Process Mining uses event logs to create a comprehensive visual process map. This provides unrivalled process understanding and accuracy, revealing every stage of a structured process based on objective data rather than subjective opinion. Process Mining also provides a detailed picture of process efficiency. By analysing these event logs, change and transformation leaders can see what parts of the process cause the most delays and bottlenecks. It’s also much easier to see which parts of the process are unnecessary and can be eliminated.

Process Mining reveals the ‘how’ in every process. It gives you a detailed snapshot of how each structured process works, and highlights areas of inefficiency for targeted analysis and intervention.

Process Driven AI: Discovering the ‘why’

Process Mining is a critical tool for process improvement. However, it won’t be enough to take an enterprise to the highest levels of competitive efficiency. It’s limited in that it can only be used on processes that have been recorded in structured data. What’s more, while it shows you where efficiency exists in a process, it cannot explain why. Root cause analysis and targeted problem resolution will need more input than Process Mining provides.

While it often goes unnoticed in the enterprise, the majority of business processes are in fact unstructured and so produce unstructured data. This describes most ‘human’ interactions in the business, which are mediated through conversations and communications. Even when performed through digital channels like email, these conversations and processes remain based on unstructured data that can’t be easily analysed.

Yet, just because these processes are inaccessible doesn’t mean they aren’t having an impact. In fact, these unstructured communications processes can be some of the most inefficient and wasteful. Manually processing emails, for example, comprises 40% of the average employee’s working day, and can cost a business up to $10,000 per employee each year. Much of this work is critical, but a great deal of time is also wasted on repetitive, low-value tasks that could be automated or removed altogether.

To gain visibility into these processes, you need Communications Mining, an application of Conversational Data Intelligence. Process Driven AI uses Natural Language Processing to convert unstructured communications like email and chats into structured, machine-readable data. A Conversational Data Intelligence platform then allows you to analyse all this data en masse and at speed.

Not only does Communications Mining allow you to monitor unstructured processes in the business – you can understand the root causes of process inefficiency, whether the process is structured or unstructured. Customer motivations and complaints, the causes of breakdown and inefficiency – all become accessible through Communications Mining.

For example, a specialist insurer used to analyse the communications work of its highly valuable but overstretched underwriting team. It discovered that 97% of work was being incorrectly routed to an underwriter first, requiring them to waste precious time forwarding the requests on to the right people. After using Re:infer to understand, route and automate all email requests correctly the first time, the customer has achieved $240,000 in savings and reduced handling times to almost half.

Process Mining shows the ‘how’ of business processes. But it’s Process Driven AI that reveals the ‘why’. Both are needed to tackle inefficiency at the source and create competitive efficiency for the enterprise.

2 views0 comments