Head of Innovation Abby. I write about artificial intelligence, intelligent automation, and how global businesses can gain a competitive advantage.
Getty
The need to keep teams connected and business processes moving has driven the urgent need to implement new intelligent automation platforms to accommodate the shift to work from home (WFH).According to the most recent polls A study conducted by Sapio Research on behalf of ABBYY showed that 98% of IT decision makers said they made digital transformation investments to improve WFH, with the most common technologies being Intelligent Document Processing (IDP) and process automation technologies such as Process and task mining.
The speed at which people can turn today is increasing. The proof-of-concept, which typically takes nine months to a year, will be installed within three weeks. McKinsey & Company confirmed the trend in October The 2020 survey claims that the share of digital or digital products in company portfolios has accelerated by as much as seven years.
There are several reasons for compressing the implementation timeline. On-premises solutions need to be moved to the cloud so remote workers can access files, systems and applications from home. Workers who used to print documents and papers in the office and bring them to their colleagues’ desks can no longer do so. Paper documents must be digitized and automatically entered into software applications.
Additionally, a fragmented workforce reveals disparate and broken processes that lead to delays and errors, frustrating customers and employees alike. Likewise, customers need access to service and customer service representatives in a socially distanced environment.
A new approach to improving process automation is needed to align with new ways of working and accelerate business outcomes. If you encounter any of the following challenges with your current intelligent automation platform, here are some solutions that might work for your company.
Low performance, accuracy, and/or straight-through processing (STP) rates
Automation solutions that fall short of your performance, accuracy, or STP goals can arise in a number of ways depending on your company’s needs.
Employees may need to look after automated software and manually correct data to achieve the required accuracy of data identification and extraction. Especially in document processing, your IDP technology may not be able to read handwritten text or various forms of ID cards, leading to more human intervention. Also, you may have promised 95% or higher accuracy, but in reality, you’re only getting 30% or lower.
Many companies have KPIs with STP rates of 90% or higher, but these companies also encounter exceptions that require employees to correct data, manually send data to another process step, or override alerts.
You can improve accuracy performance by viewing process workflow in two ways: how employees think the process works and a data-driven approach where you can see a timeline of events of when, where and how each step of the process flows. Looking at these two sets of data will show you where the disconnect is so you can improve performance and results.
Additionally, augmenting your current platform with machine learning techniques that can “read” and “understand” content and quickly understand where information is being sent in downstream processes may be worth considering.
Your intelligent automation solution is not scalable.
Lack of scalability can take many forms. The first clue is that your solution is not ready out of the box, requiring you to invest a lot of time in training machine learning models to read and understand documents and processes. Another sign is that the solution is difficult to run in environments where document types vary widely.
For example, you started out with a document processing solution for invoices, but now want to extend it to purchase orders or customer onboarding forms. As a result, you are forced to manually train the solution on these documents, evaporating the efficiency gains.
Lack of flexibility is another real sign of not being scalable. The platform must be smart enough to understand structured, unstructured, and semi-structured documents, and to automatically classify and split documents into different categories without manual intervention.
Discuss with your automation partner long-term plans to expand into other areas of the organization ahead of time, even if it is outside the scope of your current job. Learn about their ability to replicate digital transformation outcomes in other sectors. Even if you’ve been communicating with your technology partners for a while, I guarantee they’ll be happy to share how they’ve expanded your smart automation footprint. Additionally, they often have mutually beneficial economic incentives that can be negotiated.
Most importantly for scaling enterprise-wide, your solution should provide visibility into document process workflows. Without the most important bird’s eye view of the process, it is difficult to prevent bottlenecks and handle exceptions quickly.
It costs more than expected.
During digital transformation, many hidden costs are likely to increase. Some to note include:
• Ongoing charges for professional services to overcome deployment technical issues and/or upgrades.
• Whenever you add a new document or process workflow, it takes longer than expected to train the model.
• Platforms locked into specific automation vendors can become expensive. You want the flexibility to use best-in-class tools.
If any of the challenges above sound familiar, reach out to your technology partner to identify these costs and develop a plan to minimize surprises. In many cases, your tool just needs to be fine-tuned, or you may be using an outdated version of the platform. You can also renegotiate your contract for the additional support you need. Otherwise, keep in mind that it’s okay to remove underperforming tools. Today’s intelligent automation solutions feature composable architectures and low-code/no-code that easily integrate with existing enterprise platforms.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs, and technology executives. Am I eligible?
