By Henning Lund - June 15, 2017
Everybody is talking about AI, and it is here to stay. Salesforce and Microsoft have been playing on the AI drums for quite some time already and I must say it’s a part of the future I’m looking forward to. But most companies have more down-to-earth problems to solve before the AI wave really hits them.
I remember making some of my very first presentations in the ERP industry some 30 years ago. The “pitch” was around avoiding islands of data by implementing an ERP-system. And fair enough, an ERP system has helped a long way down the road, but other data has entered the scene and specialized systems e.g. for CRM are challenging the traditional ERP philosophy.
Today, we have new islands of data. Big data. Disconnected systems like CRM and ERP, and there is no other good reason for this than lack of time or knowledge about the possibilities to fix it. Or both.
I have to admit that ii is only a year ago I learned how simple a data integration between Microsoft Dynamics (365, NAV, AX, GP, CRM) and Salesforce can be made. My paradigm has always been based on an ETL framework where data is extracted for the source system, transported via mail or FTP-servers to a place where the target system can read and load the data. I remember that my very first data integration project was between a Mainframe SAP-based system and a mobile Microsoft Dynamics solution, and this was made by simple export and import of ASCII-files (aka text files for the younger generation). And it was pretty complex work to define it and set it up, and there was quite some development time involved in that. Good for me as the vendor though, and it turned out to be a long-term happy client of mine.
Read also: A Beginner's guide to data integration
I’ve learned a lot since. You don’t have to extract, transfer and load data as batch jobs. Modern technologies can detect new and modified data for transfer, and read and write directly between the systems – bi-directionally if you want. XML programmers can do amazing things with data transfer, but when the integrations are a bit more demanding, data volumes are a bit bigger, the organization is a bit more sensitive when it comes to detecting and solving potential integration problems, then the data integration platforms come in play to shine.
It took me a little while to really understand that I’m representing an amazing data integration cloud solution. The rich functionality to deal with any business or technical design, the ability to completely avoid programming and a mind-blowing performance that is making the solution robust even when connection is poor.
Read also: Should data integration be so hard?
How do you explain that something that, by nature is super complex, can be made super simple? We’re all happy about our smartphones that can do phenomenal things for us, and we believe that it always works, seamlessly and with all the information we could possibly need at our fingertip - as Bill Gates wrote in his book long ago. And yet, still, every day, people accept to work with disconnected systems. Making double data entries, and manually converting quotes in one system to orders in the other system. Should our phone really work better that our customer-facing systems and financial/business operating systems?
It’s time to act. There is no good excuse to accept inefficiency. And for many companies in today’s market, lack of infectivity is not just annoying. It’s what is preventing your company to truly shine.
In any case, if AI does not have accurate and sufficient data to process, it will just be misleading AI providing wrong recommendations, and eventually AI will figure out that your data is not consistent. Therefore, we humans need to step up. Get your data organized. Get your data integrated so that you can make informed decisions based of real data, accurate data.