Business processes involve routine tasks that are getting increasingly complex with rapid technological advances. Workflow Automation can provide the relief to both the employee and the employer—to the employee opening them up for more creative and value-added work; the employer in cost savings, efficiency, and avoidance of errors. Workflow automation is about automating business processes based on workflow rules where human tasks, data, or files are routed among systems or people. Typical workflow automation examples include onboarding a new customer, adjudicate a quote request, handling an Invoice, process a new claim, etc.
Conversational AI is about leveraging the channels facing the customer, including messaging apps, speech-based assistants, and chatbots to automate communication and create personalized customer experiences at scale. The AI in conversation represents elevating the engagement to a new level of intelligence by appropriately introducing certain components, especially the Natural Language Processing units. At times, the processing involves Machine Learning and sometimes even Deep Learning. The key is to leverage the AI components at appropriate points in the conversational flow.
Combining these two powerful technologies (process automation and conversational AI) is the key to building a smart enterprise and happier customers. With these tools, major insurance provider has already seen ~12M in costs savings after the first year of deployment. Similar impacts are being felt in other industries as well.
Workflow automation is typically achieved by connecting software nodes involved in a repetitive process, capturing human operator’s visual movements on a screen and deploying them as being done by software robots.
An SMB can automate their workflow by connecting a cloud-based application like Google Sheet with their backend structured database. As the google sheets are getting filled-in with data, manually or automatically through other means, their backend database is getting populated for further internal processing. The database tables can then be viewed via a CRM application.
A large banking organization that is processing loan applications can employ software robots to take the information from the submitted loan forms that are on the screen and copy it into the next system to continue processing. This was originally being done by humans who were involved in the routine task of copying and pasting data across systems.
Conversational interface is about being proximal to the customer (the end-user). This interface is the conduit through which a customer is reached, content of their interaction handled, conveyed into a software-based system, stored into a database, used in computing, fetching other related information, and finally efficiently giving back to the customer what is relevant in that engagement. With the recent advances in machine-learning algorithms and with a set of relevant AI technologies, there is a cognitive dimension to the conversation and hence termed conversational AI. This now becomes the realm of NLP (processing) along with NLG (generation) and NLU (understanding), depending on the depth of the use-cases.
For the digital transformation to be on the right path, clarity, and priority in sequence of actions must be established in the strategy. As found in PWC article [1], in the modern era of advanced engagements, the customer is the first stop. They are the driving force for businesses. It is only then that the business processes kick in. So, here the conversations with intelligence drives the workflow automation giving rise to an era of intelligent automation.
Let us consider an example here. A user initiates an engagement by texting in a keyword to a well-known virtual number. This is done over SMS channel. This user is interested in securing a health insurance policy. The content in this engagement is fetched from the backend system and is presented to the user. The two-way engagement continues as a conversation while the backend system is continually engaged for further information. As part of this ongoing engagement, there can be information stored on the backend system. At any point the backend system may choose to reject the policy creation and the conversational interface will deliver that experience appropriately.
While the end-user has engaged, the backend system is kept in sync and may transition accordingly. So, we see here workflow automation on the backend system as well as in the front-end, along with the advanced conversations going on in a seamless flow. Regardless of the user input being voice or text, the user’s sentiments can be derived and fed back to the backend for appropriate actions. Instead of a plain old telephone call where the operator is picking up the call, support systems at the backend operation in this old paradigm is forced to enter data manually into a system (this is often much slower and prone to errors). Here, the cognitive elements of the conversation interface is now extracting the information (structured and unstructured data) and feeding it appropriately to the backend. Thus, the conversation together with the backend workflow acceleration leads to a fully unified workflow automation via conversational AI. As part of the ongoing engagements, a cognitive graph can be built and enhanced for future intelligent engagement handling with the end-user. The system is thus self-learning for ongoing improvement.
A successful deployment is one where the focus is not just on the efficiency of business processes, but also on the means of achieving it. This includes customer engagement; the mode of interaction that is the current trend; future-proof considerations of how the consumer is moving; changing paradigm of software, computing, advanced cognitive technologies, cloud deployments and so on. Instead of narrowly looking at the backend efficiency of automating processes alone, it is vital to holistically look at the use-cases and bringing in the right set of cloud-based solutions, which can withstand rapid changes in the software spectrum and societal trends.
Companies that are deploying Ushur’s Workflow Automation Solutions are primarily first focusing on the Conversational aspects of it. They fine-tune the message delivered to the end users based on the context of the engagement, clearly establish the purpose of the engagement with the user, are very specific in what they expect from the customer, they set reminders to the end users accommodating the users behavioral patterns and finally they convey to the end user if the purpose of the engagement is achieved. They also employ Ushur’s AI-ML modules at the right places in the conversational flow.
The companies choose their integration strategy based on their backend systems. Some have leveraged reporting capabilities of Ushur and using those reports have integrated customer data into their backend systems. Still, others have utilized Ushur’s API that taps into every AI-powered engagement Ushur deploys with the end user and thereby integrates those API hooks into their backend systems. The possibilities are limitless depending on the goals and strategy
