Sly Connect V2

What is Sly Connect in the first place?

Sly Connect is a powerful tool that facilitates B2B cart creation from structured documents, mostly coming from ERP systems. Sly Connect is powered by a smart parsing algorithm that detects the format of your customers' ERP-generated orders, saves the format for future use and automatically fills the shopping cart in your webshop. But what are the benefits of using Sly Connect? Medium and small businesses don’t have to worry about costly EDI or e-procurement integrations, which are only affordable by big companies due to their high implementation costs. 

In contrast, the integration of Sly Connect in a webshop requires just a few lines of code. Furthermore, it takes around 2 minutes for new customers to configure and save their import format for future use. The onboarding can be done without any support from the shop owner. Every upcoming order takes less than 10 seconds. As a result, Sly Connect saves time, increases customer loyalty and decreases retention.

‍How does Sly Connect work?

The customer's order can be selected via "drag 'n' drop" or by clicking on the highlighted field. The file will be parsed and the preview opens automatically - the setup for a new customer can begin.

In order to select the first field, Order Number, the customer needs to click on the corresponding button. Afterwards, the customer can select the corresponding value for the Order Number by clicking on the parsed PDF. The following fields Article Number, Quantity and Article Reference can be selected in the similar way. For these fields, the customer needs to select only the first item in the PDF. The remaining items are automatically recognized and marked. If the selection is not applied correctly, for example there are too few or too many selected items, the sensitivity of the parsing algorithm can be adjusted via "less" and "more" buttons. If all mandatory fields are selected, the initial setup is completed.

The parsing algorithm calculates the positioning of values based on the position of the field values of the first product. Based on the selected values, the algorithm reapplies the format of previous parsed orders. Since the formats of the customer's documents are usually similar, the algorithm can recognize and extract the products independently for future orders. Consequently, all subsequent orders can be placed in a very short amount of time, less than 10 seconds. Currently, Sly Connect supports generated orders in PDF, XLS, CSV, IDoc and XML format.

‍Improvements

Algorithm

Although Sly Connect has a rich set of capabilities and our customers show great satisfaction with the tool, we always try to set the bar higher and further improve our tools to fit customer’s needs even better. Sly Connect is integrated by several customers. Because of the user feedback we noticed that some PDF files have structures which were challenging to parse automatically. Article numbers were, for example, positioned on one page and the other relevant data of the same product already wrapped to the next page.

In rare cases the parser couldn’t properly match these distributed values to each other and identify them as a single product.

In another case the position of the same kind of data (Article Number) varied from page to page within the same PDF. Therefore the parser missed some occurrences and could not process all the order items properly. 

For us at Sly, this was a motivation to make our algorithm even better and solve many more edge cases. During a brainstorming session we came up with the idea to introduce the concept of “row” in the algorithm to mitigate these edge cases.

The concept of “row” has the following meaning: Other fields (e.g. Quantity or Article Reference) are related to some article number and cannot be positioned before or after the row in which their article number is placed. With this concept, we can have fields that span on multiple pages, and on different positions in the PDF, but if they belong to the same “row” they would still be matched in the final selection. The image below shows the Article Number (page 2), Quantity and Article Reference (page 1) of item 03 on different pages. However, with the new algorithm, the selection was correct. Quantity and Order Reference are related to the Article Number of item 03. The red square represents the one “row”.

As a result of our improvements, Sly Connect can handle these “imperfect” PDF files much more reliably.

‍Flexibility

One of Sly's core values is flexibility. Therefore, we also wanted to provide more flexibility for Sly Connect customers, and not to restrict them to only four predefined selection fields which were originally provided by Sly Connect (Order Number, Article Number, Quantity, Article Reference). Since different customers have different requirements, we implemented the fields to be super flexible. Now we allow up to 8 completely customizable fields! The customers can decide the field labels (in different languages), field minimum length, whether the field is mandatory, regular expression for allowed field content and many more options.

‍Evaluation

In order to evaluate the new algorithm, we performed intensive testing to compare the old and the improved algorithm. After some iterations we came to the point where the new algorithm performed better in all test cases. Some tests showed that the new algorithm can correctly select all fields, which the previous could not achieve or both algorithms could properly select all fields, but the improved algorithm provided better user experience, since less adjustment with “less” and “more” buttons were required. Our customers also had a chance to test the improved algorithm and their favorable feedback confirmed our testing efforts. Customer’s satisfaction is very important to us, and gives us motivation to further improve our solutions.

‍Summary

Sly is thrilled to improve its services and to provide solutions that best fit the customer’s needs. Once we saw the potential for improvement in Sly Connect, we were determined to find solutions to further improve Sly Connect. To resolve the issue regarding page wrapping within one product, we introduced the concept of “row”, meaning that fields (e.g. Quantity or Article Reference) which are related to an article number, must be positioned in the same row as the related Article Number.

Furthermore, we made custom fields fully customizable to allow customers to define which values they want to extract on the PDF orders. This provides greater flexibility and expands the potential application of Sly Connect into all kinds of business domains.

We performed intensive testing of the new algorithm and our customers were very satisfied with the improvements and so were we! We will not stop here. We already have many ideas and will keep the same pace and focus on creating the best software for our customers. If you are interested in testing or getting more information about Sly Connect, don’t hesitate to contact us. Sly Connect Website / Request a Sly Connect Demo / Contact us by Mail
‍

‍