IWD: Data Classification – AI and RPA are NOT taking over the world according to this expert

We mentioned here in our overview of eWorld last week that we met a lady who ‘cleans, fixes and classifies dirty spend data’ and maintains it for ongoing efficiency and accuracy of your S2P process. As an expert, she calls herself The Classification Guru. We thought she would make a great person to feature as a 'woman in procurement' around International Women's Day. Her area of expertise is in high demand as businesses realise the increasing value and power of good data.

We asked her a few questions about the role of emerging tech in the world of data classification:

“When it comes to AI in the data classification universe there will be no takeover of the machines - well not for a very long time,” she tells us. “Don’t get me wrong, AI is an essential and valuable tool when it comes to data classification, we can now classify millions of data in a fraction of the time it would take to do manually, but this is only valuable if that data is classified correctly. And how do we check that? We use humans.”

So where does your role fit in? – “If we look back at the beginning of the process, many of the AI systems in use now are based on machine learning, however, the data that it is learning from has to be completely accurate and clean, and the only way to ensure that is by the human hand, or eye in this case. Hence what I do.

I have worked for many years in data classification, working alongside AI to improve data. Many people are under the impression that all these machines are classifying your data, and yes it does do a lot of it, however, if this were the case I wouldn’t have been in work for all these years, or be running a manual data classification business. If we had ‘nailed’ AI classification there would be an industry leader selling their product to the market, but as it currently stands there are a number of companies using many different methods, scripts and rules all trying to perfect their product.”

Why haven’t we perfected AI-based classification yet? – “Well, unlike some industries where AI can completely replace human intervention, data classification is not a black and white, right or wrong process. There is context, variables and conditions that all have an effect on data classification.

An example I use is DHL. To a professional services company they are most likely to be used for courier services, but if a manufacturer is using the same company they are more likely to be involved in their supply chain and providing logistics services. The AI would have to see the supplier and the industry the data comes from several times to learn and perfect this, helped of course by humans.

Depending on how the machine learning works, it can also misread invoice line descriptions. Something like “Taxi from hotel” can be confusing and the data could end up being classified as a hotel rather than a taxi, all needing to be rectified manually.

One supplier can also be many things depending on the spend amount. If a hotel has a line value of up to around £5,000 then most likely it will be for accommodation. Anything higher than this and it could be venue hire, but it is all very much dependent upon whether there is an invoice line description that can be used as a guide. This makes it very tricky to write rules for and for AI to learn from, bringing back in that need for the human touch.”

So where can AI help in this process? – “AI can really shine through for ever-changing items such as lab chemicals or electronic parts. What would take a considerable amount of time for a human to research, find and classify each time, once learned AI can classify it at lightning speed.  This is of course after the initial manual classification …”

And what of RPA and its role? – “There is a similar concern around Robotic Process Automation. A number of organisations are deploying this technology across many departments to improve extremely manual tasks such as document scanning, order processing, CRM updates and payroll processing.

While initially employees may feel threatened by this new technology, the reality is that RPA is taking on many of the tasks that humans detest. These are the manual, time-consuming and mundane tasks that can be processed efficiently, faster and more accurately than humans can, unlike the AI for classification previously discussed which requires a huge amount of human involvement.

If introduced into the organisation correctly, employees will understand and appreciate that RPA is there to help free up their workload from these tasks so that they can focus on the areas that need more skill required from human involvement. The reality is that with a number of these tasks it takes time; they are open to cut and paste, spelling or typing errors when undertaken by humans and they will often go undiscovered until an issue is flagged, which can cause further expense to rectify. As long as the RPA is thoroughly tested before implementation then this improves the experience both internal and external to the organisation.”

Will we ever get away from human intervention in this space? – “As data is ever-changing and updating, I can’t see a time when there won’t be a need for human involvement in data classification, even as a minimum to update any new or unseen data. While full dependency on AI may work in some industries, I don’t think this approach will work when applied to data classification and the same applies to RPA, there will always be a need for some human involvement at some stage.”

“I have by no means seen and experienced all AI tools and I would happily be proved wrong, however at this stage and for the coming years, I am not concerned about the fate of my future.”

As we've said, eWorld is a great opportunity to meet industry experts that you might not otherwise come across. Thanks to Susan for sharing her insight with us, we will be hearing from her again soon on “the dangers of dirty spend data.” And tomorrow we'll be featuring some 'voices of women in procurement.'

Share on Procurious

Discuss this:

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.