Unleashing the Power of Automated Process Discovery for Enterprises

Automation and AI Transforming Organisations

Automation and AI have entered a new era of maturity, boasting scalable, secure, and stable platforms that can transform enterprises. Despite this impressive progress, many organisations still struggle with adoption and scaling challenges, primarily due to two significant hurdles:

• Lack of access to business process owners and subject matter experts (SMEs), which hampers the identification of potential automation targets and their return on investment (ROI).
• A desire to transform processes before automating them, under the belief that this approach reduces effort during the build phase and leads to more efficient delivery.

Stay on Track

Like Gromit in our picture, automation delivery teams need to stay on the right track with sufficient automation targets to make effective forward progress. The answer to overcoming these challenges lies in the discipline of Automated Process Discovery.

Traditional Approaches are Counter-Intuitive

Traditional discovery techniques are counter-intuitive in the realm of automation. They rely on extensive human effort to engage with business leaders and process SMEs, manually review and map out the process landscape, and then identify potential automation targets. Subsequently, data on the number of process transactions and average handling times must be collected to calculate the ROI from automation and prioritize the best targets for automation.

Process Complexity

Additionally, an assessment of process complexity and ease of implementation is conducted at a high level, as a deep dive into each process would be too time-consuming to complete. Convincing business leaders to embrace an automation ambition level of approximately 40% of their process landscape can also be an arduous task, consuming substantial time and effort.

Failing Manual Discovery

All these challenges result in manual discovery activities often failing to provide automation teams with enough viable targets that offer sufficient ROI in terms of benefits. Consequently, automation teams may resort to asking users for their preferences on automation targets, which often yields suboptimal results. Users frequently request the automation of high complexity, low-frequency tasks, typically the ones they dislike performing themselves. Such choices can burden automation teams with an additional workload of triaging, reviewing, and rejecting unsuitable automation targets, leading to counterproductive outcomes., not to mention trying to support a wide group of stakeholders.

Automated Process Discovery

To overcome these hurdles, Automated Process Discovery emerges as a powerful solution, positively disrupting traditional discovery engagements. Automated discovery tools provide unparalleled insights into tens or even hundreds of processes simultaneously, offering estimates on ROI, process complexity, and ease of automation implementation. The main advantage of these tools is that much of the analysis is passive and doesn't require extensive input from business SMEs. Business analysts and process SMEs can then review the outputs of the analysis and compile lists of automation targets with the most significant potential ROI in a few weeks instead of months of manual work. Such automation tools enable scarce automation resources to have a far-reaching impact while reducing the burden on business SMEs in the identification of suitable automation opportunities. This capability helps organisations break through the scale out glass ceiling and keep their automation pipeline fuelled with high-ROI targets and maintain effective progress. Let's explore the various approaches to automated process discovery.

Process Mining

Process Mining is specifically tailored for well-engineered ERP and CRM systems such as Oracle, SAP, and ServiceNow. This technique has been in use for some time and involves processing a year's worth of transaction logs from these applications. By interpreting this data, Process Mining visualizes business processes, complete with variances, based on real-world data. The output is a prioritized long list of potential automation targets ranked by ROI. Process Mining eliminates the need to gather MI data on volumes and average handling times since this information is derived directly from the logs. Moreover, decisions regarding process complexity, ease of implementation, and process variances suitable for automation are all based on real-world data. Automation leaders find this type of output highly beneficial, as it provides assurance to governance boards regarding decision-making processes and increases confidence in the expected returns from automation investments.

Task Mining

Task Mining shares similarities with Process Mining but focuses on the front-end capture of user activity. By analysing mouse clicks and keyboard inputs, Task Mining maps out complex, long-running customer processes across various applications for each business function analysed. Task Mining is great for gaining rapid insight in a single business function and can identify approximately 15 to 20 automation targets from analysing just 2 to 3 users' activities for a week. The insights generated align with those of Process Mining, offering the same level of information with a fraction of SME involvement.

Communications Mining

The new entrant in automated process discovery is Communications Mining. With the rise of Large Language Models, made famous by Chat GPT, intelligence on customer service requests and outcomes, which were previously buried in emails, CRM notes, and chatbot exchanges, can now be categorised and automation opportunities identified. This technology is already proving transformative, as evidenced by Octopus Energy, which reported a 42% reduction in email processing through the implementation of these techniques.

The New Value Chain

Historical business process re-engineering and Lean principles were based on people-based value chains, where releasing benefits could only be achieved by reducing time spent by people on delivering processes and labour-intensive development and regression testing, Fast forward to today’s automation-based value chains, a single robot can accomplish the work of six people at only a quarter of the average process worker's salary. This astounding efficiency translates to a massive 24 times the effectiveness, pound for pound of virtual workers versus physical workers for transactional work.

Early Automation Releases Benefits

Automating a process without undergoing any transformation can release 60% to 80% of people's time as the work is transferred to a software robot. Consequently, organisations should focus on automating their processes as quickly as possible and reap immediate benefits through "bot shoring." Further process efficiency optimisations can be achieved later using data derived from your newly automated process that now reports on every transaction undertaken.

The Upside of Automated Discovery

Additionally, the upside of automated discovery technologies extends beyond identifying automation targets. High-volume analysis of processes also reveals the most efficient ways to execute these processes, aiding business users to accept the level of standardisation that automation demands. Automated discovery naturally identifies the most efficient custom and practice way to execute a process and does away with the need for separate 'As Is' and 'To Be' process investigations as the common happy path and all process variances are recorded. This significantly facilitates consensus among diverse groups of process SMEs on consistent processes and accelerates the pace of automation deployments, positioning organisations for further optimisation in the future.

Adopt Continous Discovery

So what approach is a bet for me? The answer is to leverage all available methods, depending on the specific business area targeted for transformation. Process Mining, Task Mining, and Communications Mining all offer unique insights and advantages that can drive successful automation initiatives. However, it is crucial not to treat Discovery as a one-time activity at the beginning of the automation journey. Instead, organizations should adopt a 'continuous discovery' approach, empowering their automation teams with powerful automated process discovery tools to continually maximize automation opportunities and accelerate the automation journey. By embracing Automated Process Discovery, your automation train will become unstoppable, much like Gromit in our picture!

David Cameron is a Co-Founder of Aivantor Ltd