Stéphane LACHENAL (SAMSE Group Treasurer) is in charge of growth operations, productivity issues and the quality of information provided to internal customers. At COLAS, François Levasseur is Director of Transformation. At FORUMDIMO 2024, they talk about their use of CashOnTime to automate accounts receivable, DSO and WCR. A workshop led by Fanny RONDET, CashOnTime Sales Manager at DIMO Software.
How does accounting deal with cash?
At SAMSE, the customer service department acts mainly as a first-level collection service, making out-of-court reminders. With regard to collections, the majority of tasks have been dematerialized and automated, in particular lettering thanks to the CashOnTime Allocation solution. The absorption of customer flows has been structured at treasury level. At SAMSE, some fifteen people are dedicated to this task, as there are two business sectors: B2B and B2C. They manage payments, whether made online, by SMS, by e-mail, etc. In the B2C sector, although the baskets are smaller, the volume of transactions is much greater.
COLAS has a highly decentralized organization. In France, the Group manages a large number of large companies via 200 profit centers – each with a bank collection account – scattered throughout the country, with 1 to 3 members of the administrative team close to local operations to handle invoicing, reminders, debt collection and customer collection. François LEVASSEUR explains: “Our customers often consider that a single IBAN corresponds to a single legal entity, and that they don’t need to manage 200 COLAS France IBANs in their ERP to send each payment to the right branch. We were looking for a tool that would enable us to link incoming payments with the corresponding invoice as standard“.
What prompted you to think about cash optimization?
For SAMSE, the motivation was productivity, as Stéphane DELACHENAL explains: “In the construction sector, the preferred method of payment is automatic LCR (Lettre de Change Relevé). Some companies want to take back control of payment and change their terms and conditions, switching to bank transfers. This has led to a decrease in LCR-type payments, which was a simple method for us, and an increase in transfer-type payments. The situation became more complex, creating a lettering problem.
COLAS, for its part, has changed its calibration: “Until 2020, we were based in France on 6 regional companies, which were able to link their collections with their 40 or 50 profit centers. At the beginning of 2021, for reasons of governance and simplification, we merged these 6 entities into a single one. We ended up with customers with a territorial and national reach who paid us on a single account for potentially 200 business units”.
DIMO Software: the choice of consulting and support
Initially, SAMSE encountered a problem with a local authority made up of five town halls but a single treasury. This meant that one settlement could be used to letter the accounts of several town councils.
“These were often small, overdue invoices, which meant a lot of time-consuming lettering. The solution worked very well right away,” recalls Stéphane DELACHENAL. François LEVASSEUR admits to having consulted several software vendors, most of which were focused on DSO optimization, which did not correspond to the company’s needs: “We are strongly oriented towards local authorities, but we have rather limited room for maneuver when it comes to improving the speed of debt collection. DIMO Software really focused on our problem: automatic lettering of incoming payments (automatic identification of the principal with the corresponding invoice(s)). They proposed a POC over a few days, with a set of tests on our invoices and customer collections. We achieved a 65% recognition rate without any configuration effort, which validated our choice of DIMO Software”.Change management: a real issue when it comes to optimizing and automating workflows
According to Fanny RONDET, it’s important to give business users control over the tool, so that they buy into the new project and can manage discrepancies or make trade-offs. “At COLAS, within three days everyone was ready to use the solution. It took no time at all to get to grips with it, and adoption of the tool was almost instantaneous”.
The approach was different at SAMSE, where the solution initially ran on empty: “We then indicated to the operational staff the tasks to be carried out and displayed the scores. Offering a tool that works well reassures and relieves colleagues”. Stéphane DELACHENAL adds: “As our ERP was rather complex in terms of cash-in-hand, it proved much easier to use the CashOnTime solution, which was much more fluid and fun to use. The product’s ergonomics and ease of use were much simpler. It’s nice to have a modern, user-friendly tool that doesn’t require 15 clicks to complete a transaction. What’s more, the same screen is displayed for a population of treasurers who are used to doing reconciliations. It’s a tool that integrates perfectly into a central or decentralized business process.
CashOnTime takes advantage of machine learning and Artificial Intelligence
Customize settings
Fanny RONDET explains that CashOnTime differentiates two things: “There’s machine learning, but there’s also a parameterization layer. We customize the parameters according to our customers’ activities and businesses, to build strategies that deliver results right from the first file injection. The AI learns from users’ actions, and from encounters between banking information and invoices, to enable the tool to go beyond the simple context of the information contained in banking files at a given moment. This is what will enable us to achieve significant lettering percentages.
CashOnTime also relies on the identification of the paying customer. There are four criteria:
- IBAN in bank file.
- the invoice number or part of the invoice number, if announced.
- history: the solution takes the bank’s main wording and checks its archives to see if any transactions have already taken place.
- keyword” search: starting from the main bank code, the solution uses an algorithm developed by DIMO Software to focus on “main keywords”.If it finds that there is an incoming payment marked “transfer received [Company name]”, it takes the company name, analyzes the invoice database and the customer database to retrieve all the accounts linked to the company in question. This process makes it possible to clearly identify the principal before launching the lettering strategy.
Total invoices
Basically, the tool starts from automatically retrieved invoice numbers, which it then aggregates, and looks to see if the amount of the incoming payment is similar. It then applies a customized lettering strategy, for example, by grouping invoices with the same due date. It can then go back to the creation date, an order number, etc. It is possible to add the overdue balance of the customer account, the customer account balance, etc. If several options are available, CashOnTime Allocation will display proposals for the user to consult for manual arbitration.
Lettering for key accounts
Fanny adds: “The solution also makes use of transfer advices when the customer is a major customer, with lettering on tens, hundreds or even thousands of invoice lines.In many cases, the major client communicates the information upstream, generally in the form of native PDF files, Excel files, etc. sent by e-mail. This data is fed into CashOnTime, which reserves invoices and highlights potential deductions. CashOnTime also dematerializes debit notes and credit notes, so that they can be used when collection arrives.A turnkey approach
DIMO Software deploys CashOnTime in “turnkey” mode, supporting its customers from start to finish with clear steps for project deployment and CashOnTime expert consultants dedicated to implementation. The solution exploits the data made available according to the customer’s needs. Some go very far in machine automation. Others prefer to have user controls on certain projects, in the form of partial processing”.
Order to cash: absorb daily flows and harmonize processes
In addition to saving time on low value-added tasks, the implementation of CashOnTime also enables rules to be restated and ensured that they are respected from one subsidiary to another, or from one manager to another. An initial audit enables the rules to be redefined to ensure immediate data processing. A connector with the customer’s ERP reinjects all the necessary lines into CashOnTime Allocation. The bank file is the source of the lettering trigger, which takes place between 6 a.m. and 7.30 a.m. when bank calls are made. The automatic percentage is immediately sent back to the company’s ERP and accounting systems. This enables the data to be sent to CashOnTime Collection, which then collects it on the basis of updated morning collection data. “This global link in the order-to-cash chain is crucial for effective data management,” assures Fanny RONDET.
DISCOVER CASHONTIME ALLOCATION
Visible benefits after just a few months
François LEVASSEUR points out that after one year, the improvement is very quickly visible: “We are on average on 70-72% automatchings, with peaks of 80-85%. But there are inevitably a few remaining invoices or collections to be processed manually, as we also handle invoices from local authorities with various problems – notably price revisions – and therefore for which not everything can be matched… DIMO Software was involved right from the scoping workshop at one of our branches, with its administrative manager and the customer’s accountant. In fact, we were looking to implement the tool in line with our teams in the field, and with transformation in our sights.We used to have 60 FTEs (full-time equivalents) dedicated to cashiering, but now they’re a bit multi-skilled and do part-time cashiering. We’ve moved from a centralized to a decentralized model. We set out on a very broad deployment, with many users. We gave all local users the possibility of handling the two residual daily cash receipts not processed by the solution themselves, in a tool with more satisfactory ergonomics than an ERP“.
Feedback on the cash flow chain
It seems that good lettering requires good debt collection, and vice versa. However, it can happen that data doesn’t arrive quickly enough from one department to another, because there may potentially be errors that could trigger wrong information in the communication with the customer.
Stéphane DELACHENAL recounts how customer accounting was criticized for not moving fast enough: “We were on an old model where, to be able to contact a customer account, you had to make a reminder and go through the ERP to get a billing statement. This wasn’t very efficient, and it took a very long time when we were dealing with scattered entities that we couldn’t consolidate. Lettering enables us to make accurate reminders on a daily basis, whereas in the past, certain reminders could be sent out without any real supervision. We can also manually process outstanding balances on complex files. We have formalized a new organization at Group headquarters and in our subsidiaries, which generate sales in excess of 100 million euros. From now on, everyone has the same tool and there are no more discrepancies between entities. Internally, this has given credibility to communication between the various departments, by highlighting our mastery of flows. It’s a powerful tool for improving overall performance.François LEVASSEUR explains that with CashOnTime Allocation, 70% of the transactional work has been eliminated: “It’s the machine that does it, and that brings up all the data. What’s more, the tool offers an attractive screen for supervising everyone’s actions. Residual receipts are correctly processed on a daily basis. In the past, collections were sometimes carried over to the end of the month, at the time of the cut-off and calculation of the bonus/malus on customer debt in our branches. Now, we can monitor the residual amount as it comes in. Another benefit is that when staff return from vacation, only 30% of collections have to be processed, rather than 100%! More monitoring, more supervision, and a smoother pace of accounting. We also have automated bank reconciliation.
Major challenges in the order-to-cash context for 2024-2025
Stéphane DELACHENAL believes that we need to free up time for electronic invoicing and instant transfers, and prepare for dematerialized invoicing: “We don’t know whether there will be a codified format from the moment the invoice is issued. Collection is the lifeblood of private companies. But the more steps are added to the payment process, the more preparation is required. We’re going to be more qualitative, more managerial, more analytical when it comes to data, because Europe is starting to put on the pressure. We have a large number of interfaces with a general-purpose ERP, whereas SAMSE works more in specialized fields. Communication with the publisher is complicated, as France is a small country for the world’s major ERP publishers. As a result, we have to develop specific bricks”.
After completing the deployment of the CashOnTime Allocation solution, François LEVASSEUR realized that it was possible to set up high-performance processes outside an ERP system that was limited in terms of ergonomics and functionality. We can find application bricks to handle new processes, for example to capture expenditure commitments and supplier invoices. The supplier chain is full of complexities, involving many different players and principals who issue purchase orders received in the ERP. Suppliers need to be diligent and disciplined in the information they enter on their invoices, provided that the authorities can also help standardize the fields (IBAN, etc.) for the purchase order, which would facilitate matching operations”.
A few points to watch out for
To get a project off to a good start, François LEVASSEUR believes that the tool’s deployment and implementation must be well defined. It’s important to get in touch with users to analyze their needs, and find a good compromise to ensure that the tool makes a visible contribution to user acceptance. “On the advice of DIMO Software, we opted for a virtually open-ended configuration, so that any user can see the cash receipts of his or her branch neighbor. This means, for example, that a user can see whether an invoice can be cashed at a nearby branch. We’re gaining in fluidity.Stéphane DELACHENAL, for his part, believes that machine learning is very powerful, but that we need to be vigilant: “If an operator makes too many mistakes at start-up, we’ll have to modify the learning process. We achieve a 70% success rate after running the solution for a week. On the other hand, you have to take into account the specificities of each company, because the machine capitalizes right from the start. In addition to productivity, we also need to work on quality. We’ve achieved 100% lettering on some well-done statements. The solution has made our lives so much easier, we couldn’t do without it.
Do you have any questions about automating your accounts receivable?
