Document understanding revolutionizes processes for E/D/E

E/D/E is a family-owned business with a history stretching back over nine decades. Now in its third generation of leadership, the company pays for €9.3 billion worth of tools, technology, and equipment its members buy every year. Retail companies in the group can then sell the goods through stores, while commercial businesses use the supplies in construction and manufacturing.

Running such a large operation, with thousands of members, suppliers, and partners is complex. It requires a huge web of processes, information, and payments that need to be efficient and effective.

Achieving this can be a challenge, especially when the business has such a long legacy, with older systems and IT. Which is why Holger Gerlach, team leader in the finance department at E/D/E, is always keen to explore new ways to streamline work, boost return on investment, and transform costs.

“Back in 2019, I was talking to an old colleague,” Gerlach begins. “He mentioned how his company was using automation to transform operations. I thought we could do the same and the idea was born.”

Working with a colleague in IT, Gerlach chose UiPath as a technology partner. He then presented his plan to the CIO and finance director. Both were instantly taken by the concept of automation, even suggesting a partner that could help with the implementation: Office Samurai, based in Kraków, Poland. A proof of concept started at once.

A promising start

The inaugural automation dealt with the validation of VAT details. “When we send items to other member states in the EU, we don’t charge VAT. It’s tax-free,” Gerlach explains. “But to remain compliant with regulations, we must ensure customers have valid VAT identification.”

To achieve this, E/D/E had to pull reports from its SAP system, which created an Excel document. Then a colleague had to add customer details showing their names and addresses. This spreadsheet then went to a tax specialist to validate the information. It was a manual process that took six hours each month and cost €500 in accountant fees. “It was inefficient. A drain on resources, yet had to be done,” Gerlach recalls with frustration.

We automated the process. The robot took the details from SAP, added information, and then used a script to check the list against German and EU tax databases available via public records websites. Not only did the automation reduce the time taken to 50 minutes a month but saved on fees too.

Holger Gerlach • team leader in the finance department at E/D/E

“We presented the results to the CIO and finance director and got a smile,” Gerlach says. “They asked wryly why we’d not done this sooner. It was a great success for a business which can be a little old fashioned. We’d proven we could use a cutting-edge technology to improve processes and save money.”

Expanding the role of automation

To ensure the smooth deployment of the software, Gerlach knew he’d need staff on-board. “We had some lunch and learn sessions with colleagues. At first, I worried people would think we were going to replace them with automation, but they didn’t see it like that. One person said to me, ‘I have too much work to do. Getting help from a robot would be great.’ It was a fantastic response.”

A second automation followed, creating an SAP report each day in one department to send to another, where it was printed and actioned. It was an additional small success, which led to the first big test for the technology.

Making the most of technology

“We offer customers a service called factoring. This is where we buy unpaid invoices to help boost their cashflow,” Gerlach says. “Before we take on the invoices, we need to check all the debtors and calculate the risk in doing so. Normally there are between two hundred and a thousand to check, so staff did it manually.”

Then a larger customer asked for help. It wanted to sell E/D/E invoices from 20,000 debtors. “It would have taken six months to check them,” Gerlach confides. “We simply couldn’t do it, which would have lost us the contract. So, we asked Office Samurai for a solution.” An automation assessed creditworthiness by downloading all the necessary data from a credit report to process it automatically.

The task took just three and a half weeks, saving months of work and securing the contract. It was a major turning point, further proving a new and novel way to automate, which also had a tangible commercial impact.

Adding document understanding

“We have a central warehouse in Germany which holds over a hundred thousand items. Our members can order what they need from here, with 22,000 items being requested via 6,300 deliveries daily,” Gerlach explains. “We have 11 colleagues in a team to manage inventory, updating SAP and reordering replacement stock as needed. It was a huge manual process.”

One of the 11 team members left the company, which put extra strain on the remaining staff. Having heard about automation, they asked for help from Gerlach. “At first, we said it would be too hard, because we worked with 700 different suppliers which all had different order confirmation documents. We thought we’d have to build a different template for each one.”

Gerlach put the challenge to his partners at Office Samurai, who suggested document understanding as a solution. With this technology, an automation could extract all the information it needed from the different documents and then process them. It worked like a dream.

Recently, another team approached Gerlach with a problem like that faced by the warehouse management team. “We said to them, don’t worry, we’ve got something that will help. We’re now exploring their process before we deploy the same automation.”

E/D/E now has two software robots handing 25 processes. Each automation pays for itself within a year, after which every efficiency savings represents a return on investment.

The secret to success

Reflecting on the use of automation at E/D/E, Gerlach says, “It’s not huge, but we’re growing quickly.” He’s also keen to point out that E/D/E’s achieved its success so far without an internal team. “It’s just me and two other people from the IT department, doing it in addition to our main jobs.” The secret has been working in unison with the IT team and partnering with Office Samurai.

Together, we’re a powerful force, getting great results and enjoying the work. We receive ideas for automations from colleagues, do an initial exploration and then ask Office Samurai to suggest how they’d make it work efficiently.

Holger Gerlach • team leader in the finance department at E/D/E

Planning for the future

For E/D/E, the plan is to continue the careful roll-out of automation with a focus on document understanding. Contemplating his experience, Gerlach says he has three pieces of advice for other businesses. “First, get the right support. Working with Office Samurai has been fantastic. An external partner can implement on your behalf, bringing a wealth of experience. They can also be proactive, offering new ideas, suggesting new technologies and finding ways to address problems we can’t.”

Secondly, he warns not to go too fast. “Don’t try to do everything straight away. Go step-by-step, especially if you don’t have much internal resource. Ensure you get people on board from all levels of the business then build a pipeline of automations and deliver them carefully.”

Finally, Gerlach says that to make document understanding a success, it’s vital to carefully train machine learning models. “While we were training the robot for the warehouse team, we had different people involved, all of whom had a slightly different approach. This confused the robot. Office Samurai was able to provide best practice guidance and train the robot to get better accuracy and outcomes.”

Considering E/D/E has operated for over 90 years, with all the legacy and history that brings, it’s fast embracing a cutting-edge technology that puts it ahead of many similar organizations. And while its history might be in ironmongery, hand tools, and workshop equipment, a new generation of tools will aid its future. Namely, automation, document understanding, machine learning, and AI.