The Challenge

Since the beginning, Peter was continuously looking for answers to the client’s questions. As a consultant in greenhouse cropping of tomatoes, the questions depend on many context parameters (previous and future climate, the balance of the crop, time of the year, the place on earth, high tech or no-heated greenhouse…) making the advice complex.

Experience, visual recognition, and knowledge management are important. But his biggest problem was finding quickly resources and relevant information to help his customers. “I know I have an article about this, but where?” To look up the information did cost me easily 30 minutes and sometimes much more – Peter says.

Moreover, grey literature is important since it describes the small innovations and tricks to solve a problem. “I did tear out the revues during the last 30 years all those interesting articles. I even did sort them in categories.  But I have no time to scroll through all of them. When making presentations I take the time; they are good learning moments. How to find my information back in a structured and fast way”?

The Solution

The solution is based on scanning and OCR (Optical character recognition) all the grey documents. Meanwhile giving that document the facets (taxonomy tree classification) to help retrieve efficiently, filtering the search results by topic. The author group can help to find better what you need: research, own advice, grey literature, colleague’s advice, commercial product description. We do this also with the files on the personal computer and the most important manuals. “A lot of work? Yes, but I found a company with a professional scanner, it makes life easier.”

Since InnoGreen works internationally, the documents are in four languages. We made a translation list with more than a thousand words including all the horticultural ‘jargon’ words. The search function now finds the corresponding words in texts of the three other languages.

Another part of the solution is to visualize the findings in its context in the document. It is obvious that the word ‘cold’ has another meaning when we look for the cold in winter or in summer, or of the irrigation water or of the roots. Hence, every search hit is visualized with its corresponding text paragraph which makes a faster accurate finding of what they were looking for.

Peter worked for about one year to find out what they needed, how they searched for information and mainly how to categorize the information.  A full search text would give too many answers and must be fine-tuned by using the facets. They looked to the major players on the market but also to the open-source providers of semantic full-text search programs. The open-source program is well adapted to small companies’ request but they still need adaptations, and this proved to be a bottleneck. “Finding people to adjust the software to fit our needs became a nightmare. From this learning, we went back to the major providers but more focused on a local service to adjust the program to our company’s needs.

Also, the investment cost to fine-tune a program was a hick up. Luckily Xenit proposed an acceptable offer. Top reasons: multilingual possibilities, meta-data management, and personal service.”

Faster results with Alfred Finder

Innogreen is a knowledge-based company. Since we search for information constantly, if we find and read a given document in 3 minutes instead of 30 minutes, we gain multiple hours per week. Furthermore, we can improve the quality of our answers to the needs of our customers by finding more better information.

One employee can waste up to 10 hours a month on documents inefficiencies. Working on incorrect or outdated files, wasting time searching for documents or looking for the right version, generating duplication, those behaviors decrease dramatically the productivity and increase the operational costs. Challenges that most probably your company is facing, just right now.

Alfred Finder is a web application, that allows you to quickly find documents on an Alfresco back-end, preview them and process metadata.

How does this project support Innogreen’s Digital Transformation?

We still use our intuitive selection to map and stock a document but we link immediately the facets to the file. This avoids a second reading. We understand now that a click less or more represents a huge difference in time.  It gives a good feeling that the information is now in the cloud, a safer place than a portable computer and a more accessible place than the rack with documents on the attic.

Based on our new data structuring we can envision making full use of our available archived historical data, where in the past this was rather erratic.

However, Integration with other applications will present a challenge since the bottleneck of integrating existing highly different solutions in one superstructure is not cost-efficient for us. To have one base structure that we can use as the foundation for future developments may avoid this issue.