Do you ever compromise by buying a business solution that seems cheaper (on its face, or at first) — even though it doesn’t meet your requirements? Then, as you regret that decision, you go on buying bolt-on fixes or a patchwork of extras in hope that your needs will eventually be met? Perhaps you add a few hands to hold all these pieces in place, with fingers crossed that it will all work together. Eventually when you add up the costs, you’ll realize that you ended up spending way more than the price of the best solution would have been.
This scenario is entirely possible when it comes to business process automation solutions, too. There are many options for intelligent process automation, also known as Intelligent Automation, so how should you choose?
Let’s talk automation
The real question is: Are you hoping to remove some bits and pieces of inefficiency by automating some tasks? Or are you looking to transform?
Let’s consider the example of email intake at a customer service desk or common mailbox. This is a universal problem in every industry: at a logistics company where the customers/buyers frequently ask about shipment status or confirm sales orders; finance and accounting teams where vendors track payment status; insurance claims teams that receive hundreds of follow-up emails regarding claim statuses; banks where customers often reach out with various requests regarding address updates, account statements and loan information; and so on. This is probably the most fungible of all use cases.
Let’s divide the problem into its different steps.
- Classify the email: What is the purpose of the email? Which category of request applies: Is it a query, perhaps requesting a status update? Or actionable, such as updating an address?
- Route the request to the correct team.
- Read the email and extract key data points. For example, if it is a request for an account statement, the important fields would be: Account Number, Account Name, Statement Period, etc.
- Action it. This could be data entry into the target application, querying a database to fetch the status, look-ups and validations, etc.
- Close the thread by responding to the client/requestor with the final status of the request.
A quick-fix approach to the problem could be to keep steps 1, 2 and 3 manual (put all the data into an Excel spreadsheet) and bring in an RPA bot to perform the data entry and send the closing email.
Another quick fix may be to put in some keyword-based rules to classify emails, where anything not classified (because of the limitations of a rule-based approach) goes into an “Others” category — which, let’s face it, doesn’t really solve for much — and then maybe use another RPA bot for data entry, system look-ups and some validations, etc.
Transform the work
If you were to zoom out and to take a look at this approach, the tasks and the overall workflow remains exactly the same, except replacing a few human tasks with bot tasks, and instead of a person doing it, there’s a bot. Just automating manual processes as-is is a band-aid approach with limited gains. However, for truly innovative organizations, who recognize less is more and want customers and their employees happy, the way to go is to transform the task. The real benefit of automation is realized through considering the problem end-to-end. The “counting bots” approach is counter-productive because it encourages automating as-is processes. Bear in mind that your business process may already be inefficient, not due to any failure in your organization, but just relative to new methods available.
Let’s look at a truly transformative approach to solving the problem of email intake with automation:
- An API call fetches an email from the mail server, and extracts all the key data points from the email immediately.
- These data points are used to auto-classify and route the message to the correct team using machine learning, then action the email and close it.
Let’s point out the inefficiencies automation removes:
- The API call is much more efficient than a bot navigating your mailbox and sifting through your emails. [Tech used: APIs]
- Information extraction has been reduced from two steps to one. This includes opening email attachments and extracting the data points from them simultaneously. This reduces multiple touches by bots and people (upon email arrival, routing to team, final check). [Tech used: OCR reading email attachments, machine learning extracting data]
- Extracted information is used to classify and route the email, then to look-up, validate and update systems and applications. [Tech used: machine learning for auto classification, RPA for working on applications, rules for performing checks and validations]
- Throughout, bots play role of “maker,” and do all the heavy lifting of going through content and analyzing it. If required, people are brought in to the workflow [“human in the loop”] to verify and sign off information.
- Security and Operational analytics offer insight and governance – the platform tracks automated process efficiencies and inefficiencies.
Summarizing, to achieve a transformational level of automation, the tech components required include API, OCR, RPA, Rules, Auto Machine Learning, Workflow, Analytics and Security.
In both the short- and long-term, it’s faster, simpler and more cost-efficient to choose a provider who offers a process automation solution — not just a tool. WorkFusion’s Intelligent Automation Cloud includes all the technology components needed for automation in a single unified platform. If you’d like more information or a demonstration, please contact us.
of Intelligent Automation
of Intelligent Automation