Commercial real-estate company JLL\u2019s automation of its source-to-pay process, like many enterprise IT projects today, had its origin in the pandemic.\n\nAs COVID began to close offices around the world, many enterprises faced challenges managing or monitoring cashflow, JLL among them.\n\n\u201cWe were watching cash coming in and cash out the door, given the situation that most large organizations were in,\u201d says global CIO George Thomas, who oversees enterprise IT at JLL, including data centers, ERP systems, and other software infrastructure in support of the company\u2019s 300-plus offices in over 80 countries around the world.\n\nAs the pace of the pandemic accelerated, it became harder for finance systems to keep up: \u201cThere was the high volume of the accounts payable turnover, with long processing timeframes due to paper checks, siloed systems operating and interacting with our accounting systems and our human capital management systems, with no real-time visibility across the firm for accounts payable,\u201d he says. On top of that, payment approval processes were not streamlined and lacked a consistent global tracking mechanism.\n\nWorking with the CEO, CFO, and chief procurement officer, Thomas identified an opportunity to automate the source-to-pay process using artificial intelligence and machine learning on top of JLL\u2019s ERP systems to deliver greater efficiency and a great user experience, a project that has earned JLL a CIO 100 Award for IT innovation and leadership.\n\nVisibility\n\nAs Thomas and team formulated the project, they asked themselves what benefits end-to-end automation of the source-to-pay process might bring besides better cycle time and better visibility into the cash flow.\n\n\u201cThe soft benefits quickly turned into hard benefits, because we then looked at the number of people who might be required to do this, and it was going to be significantly lower than the number of people that did it previously,\u201d he says. \u201cWe could move them onto other things that we were required to do or aspire to do.\u201d\n\nWith senior executives sold on the benefits of the project, says Thomas, the team moved on to identifying high-level requirements, then \u201cdetailed planning in an agile fashion across various design sessions around the world,\u201d and finally, release planning with more and more of the operational groups within the finance, procurement, and IT functions.\n\nInitially, JLL\u2019s goal was to standardize both the technology stack and the business processes around the world, although in the end some local customizations were required to accommodate regulatory needs or compliance with government policies.\n\n\u201cIn hindsight, had we done a little bit more work up-front to understand those things a little better, we would have prevented some undo-redo work,\u201d Thomas says. Nevertheless, in the 22 countries where JLL has automated source-to-pay, \u201cWe have standardized the tech and the process almost 80% of the way.\u201d\n\nWith a strong business case, securing executive buy-in and funding for the project was relatively easy, and the biggest challenge turned out to be in managing change across the organization, he says.\n\nThe amount of change that JLL\u2019s operational units needed to go through to perform the new processes presented a challenge. \u201cThe underestimation of that cultural aspect of change management that we needed to bring into the planning cycle, \u2026 if there was better appreciation for that, we would have probably been done much faster,\u201d he says.\n\nIn addition to the human factors, automation also meant technology change. \u201cWe introduced several new technologies that needed to connect to one central hub, which did require an intense level of testing to properly sync all of the systems up,\u201d he says. While that made the process a little longer, the team lost no time in realizing value from the project: \u201cWe didn\u2019t wait till all 22 countries were implemented to start extracting value. We put it into a market, we understood that the business case had legs to stand on, then we went on to roll out into the other countries, so value was being added along the way.\u201d\n\nSupply and demand\n\nAnother factor complicating the automation process: It wasn\u2019t just JLL employees and IT systems that had to adapt; there are demands on suppliers, too.\n\n\u201cIn a program like this, we have a tendency to ignore the supplier transformation that\u2019s required,\u201d says Thomas. JLL must exchange data with its suppliers through the new framework, meaning allowance must be made for them to plan and implement the program, too. \u201cWe may not always be the highest priority for them because they have a range of programs that they need to complete for their corporation as well,\u201d he warns.\n\nHe recommends being transparent with suppliers early in the process, setting out the requirements and the benefits, and allowing time for them to adapt.\n\nSuppliers can make or break automation projects like this in other ways, too. \u201cAfter the first implementation, we understood that supplier data is not always clean,\u201d he says. \u201cData capture will need a good data quality program around it.\u201d\n\nBased on his experience gained from automating source-to-pay at JLL, Thomas says he would take a slightly different approach to future automation projects, taking the time to understand localization needs earlier in the process, and kickstarting the change management process much sooner, so that it advances hand in hand with the technology changes.\n\nAlthough Thomas dwells on delays caused by some early decisions, he says it only took a few months for the transformation to start producing results. \u201cOnce we started seeing them, it surpassed anyone\u2019s expectations. That then allowed us to demonstrate that we needed to go faster in these other markets, and it was quite easy for us to fast-track the implementation of the other countries.\u201d\n\nAs a result of the project, JLL is seeing faster invoice processing and a greater volume of payments being made electronically, further reducing cost.\n\nBased on the success of the source-to-pay automation project, Thomas has made extending the use of intelligent automation at JLL one of four strategic imperatives for his team over the next two or three years. The other three are enabling a hybrid working environment; making IT infrastructure secure yet adaptable; and delivering an intuitive user experience for employees and clients.\n\nThe automation of source-to-pay plays into some of these as well: Hybrid working is much simpler when you no longer need physical access to documents such as purchase orders or invoices, while a user experience that doesn\u2019t involve scanning papers and cross-referencing files in multiple systems on-screen is also more likely to be intuitive.