SELECT distinct(b.rankyear) FROM magazine_details a left outer join ranking b on b.mag_id=a.sno where b.app_status != 0 and a.cat_id=79 and b.rnkid is not null group by a.year order by a.sno desc limit 1
SELECT distinct(a.year), count(*) as count_yr, a.image_name,a.url,a.label1,a.label2,a.date, b.rnkid,b.rnktitle,b.rankpicname,b.weburl,a.cat_id,b.app_status,b.mag_id,b.master_stat FROM magazine_details a left outer join ranking b on b.mag_id=a.sno where b.app_status != 0 and a.cat_id=79 and b.rnkid is not null and b.rankyear = '2020' group by a.year order by a.sno desc
Artificial Intelligence is causing a worldwide impact on our daily lives. Robotic Process Automation (RPA) is a popular way to introduce AI into Finance departments. I predict that this trend will grow over time. I have heard that many RPA installations have failed to be rolled out properly. My experience with RPA did not end in failure and has been positive.
I began my RPA journey by speaking about getting ready for AI to financial executive peers. I am the CFO of ULVAC Technologies, Inc. and I was approached by an RPA integration partner called DocuPhase. They attended a financial conference called Accountex in Boston where I gave my first talk on AI.
A couple of DocuPhase employees attended my talk and later asked if I have started with a vendor who could deliver an RPA solution that I described. I had done research on the major RPA bot vendors at the time and DocuPhase was not on my radar. I had a met their sales team and participated in some online demos. I had a good feeling about company representatives I met and I requested a quote.
My goal was to automate the vouching of AP invoices into my ERP system called Visibility. My requirement was to have electronic invoices be translated by a bot and entered into my ERP system through an API (Automated Programming Interface) via my SQL server that Visibility uses. I could have chosen to have the data be input like an Excel macro that duplicated the screen steps of my AP person, but I chose the API route to update the correct SQL tables directly for speed and future ERP upgrades. My Visibility system needs to have a bot match up the invoice to a received purchase order part number and line.
As the CFO, I still need my CEO’s approval to move forward with this plan for internal control reasons. I had to justify my request to implement RPA like all our capital expenditures.
Since we are using AI with RPA, this is called Intelligent Automation (IA). This is a trend that will make RPA “smarter” in time as advances in AI are made
The highest manual transaction volume in my department is in AP. Bots work more efficiently than humans to do data entry into a computer. My company was going through a growth phase, and AP became a bigger burden and it was forecast to be a bigger burden as the volume of transactions was projected to grow. I was able to justify that the cost of my RPA solution was less than adding an additional AP person. I had excess capacity forecasted in my financial model on bot availability.
We bought an OCR (Optical Character Recognition) Server so we can scan paper invoices to DocuPhase where a bot will extract the invoice information needed to vouch the invoice on my system. A bot uses AI to figure out the details that were scanned. The bought is taught algorithms to identify invoice numbers and purchase order numbers from a variety of vendor invoice formats. AI can also help the bot learn about exception handling. When the bot has trouble processing a vendor invoice, the invoice is digitally set aside for a human to find the reason and help the bot handle future invoices. The bot can also learn over time using machine learning. If it recognizes patterns of how we clear up exceptions, we can teach it how to avoid the exception reason in the future.
Since we are using AI with RPA, this is called Intelligent Automation (IA). This is a trend that will make RPA “smarter” in time as advances in AI are made.
Now if you read and thought, “hey, this sounds easy!”, it is not. There has been a lot of behind the scenes work between my Visibility ERP vendor and bot integration partner called DocuPhase. There were field mappings and programming to make this solution work. In theory, once the mapping works on my system, the next Visibility customer will have an easier time to upgrade to DocuPhase’s RPA/IA solution as the groundwork has been done on my system. My work was to scan hundreds of invoices to train the bot.
What do I attribute to my success? I need to like the vendors that I use. I had developed a good relationship with people from Visibility and the same for DocuPhase. I also had carefully chosen to vendors that had a mutually beneficial interest to get my RPA/IA solution working. My two vendors can now partner in the future based on the knowledge they each gained from my RPA/IA solution. I hope that the knowledge captured from my system can help existing and future Visibility ERP customers integrate RPA/IA from DocuPhase with ease. I have since spoken at both user conferences to share my story. We started with an implementation plan. We needed cooperation from everyone involved. I gave my DocuPhase vendor access to our Visibility test server to prove out the solution before my real data server was introduced to the RPA/IA solution. I had patience to make sure that there was good communication between my vendors and that vendors had the appropriate people working on this project. The lion’s share of the work was definitely between my two vendors. My company focused on testing the system.
My next goal is to rollout this automation to other parts of my business where there is manual data entry.