Evolving Scope of Artificial Intelligence in Procurement Technology<\/p>\n
From augmented procurement to autonomous procurement, in this article, we\u2019ll take a brief journey of the evolution of procurement.<\/p>\n
In the history of Artificial Intelligence (AI), the year 2014 has been a turning point. That year in May, Google revealed a new\u00a0prototype\u00a0of the driverless car. In June, a chatbot named Eugene Goostman beat the \u201cTuring Test\u201d heralding a new era for using human-like bots for automation.<\/p>\n
Incidentally, 2014 was also a key year for Zycus. Having launched our Procure-to-Pay platform in 2012, we were at the threshold of realizing the immense potential A.I will have for procurement. Looking back, it was one of the first steps that took us to where we are now, and towards Merlin. But first, let\u2019s go back in time\u2026<\/p>\n
For the past 15 years, procurement has always been juggling between quality and quantity. On the one hand, they had to deliver strategic inputs involving time, analytics, and intelligent negotiation. On the other, the requirement was to review and process thousands of transactions and records like a machine.<\/p>\n
To strike that fine balance, the demand for a digital transformation grew stronger. Companies started moving away from ERPs to dedicated source-to-pay solution suites. Few of these solutions, or rather the best of them, also came with inbuilt A.I capabilities.<\/p>\n
The first generation AI-based procurement solutions used essential machine learning technologies and used probabilistic models to aid spend classification and contract or supplier analytics. Such technologies require the right data sources to learn from and be effective.\u00a0 Hence, it is natural that only established procurement solution providers can provide meaningful AI solutions to their customers.<\/p>\n
For instance, Zycus was the pioneer in AI-based spend analytics using Auto-Class, a patented AI automation platform. Auto-Class can work on a massive volume of unstructured spend data (PO\/invoices) and automatically classify and assign the correct category codes to each line item using the knowledge gathered from over a trillion dollars\u2019 worth of processed documents.<\/p>\n
The second-generation AI-based procurement solutions evolved more recently as more and more companies started automating their core processes providing a more developed \u2018base\u2019 framework for AI to grow. The focus of AI also shifted partially from a back-end computation platform to \u2018smart insights\u2019 helping users with actionable insights. This provided what we called Augmented Procurement experience by leveraging technology to increase procurement control.<\/p>\n
This stage also saw AI use cases expanding across the entire gamut of source-to-pay processes. Some instances of 2nd<\/sup> generation AI solutions at Zycus focused on improving turn-around times.<\/p>\n Automated three-way matching of PO\u2013Invoice\u2013GR, converting supplier pdf invoices to eInvoices, and tagging Non-PO invoices to existing contracts were some examples that stood out in terms of their extensive adoption.<\/p>\n For example, Guided Buying routed users to preferred buying channels (favorite suppliers, contracts, catalogs, etc.), even for free text searches. Spend monitoring and mining looked to provide detailed insights on savings opportunities hidden within detailed spend data based on experience. Smart Assist in contract and supplier management helped in adhering to compliance and timelines with timely insights to the right stakeholders.<\/p>\n This was also the time when Natural Language Processing (NLP) became a part of procurement applications with conversation platforms like Zycus Ask-Anything encouraging human-like conversation with machines. The synergy between solutions and their users were changing and changing fast.<\/p>\n The topics for most procurement conferences held in 2018 focused on AI, Intelligent Augmentation (IA), etc., as the path forward for procurement. But as per Gartner, \u201cWithin the market for procurement, the current penetration of the addressable market for AI is estimated to be within the range of 1% to 5%.\u201d*<\/p>\n Why is there such a massive gap between expectations and reality for procurement ambitions?<\/p>\n There are two reasons which we at Zycus identified, and the insight is the driving force behind the third-generation AI-based procurement solutions <\/strong>designed by Zycus that will result in transformational change.<\/p>\n First of all, AI based processes are being created in bits and pieces. They provide a partial advantage but not the power of an integrated AI process across the solution suite. Also, the development does not cater to the needs of individual organizations despite customer-specific learnings.<\/p>\n Secondly, AI solutions are still used as a computational advantage. As a result, they are dependent on the end user\u2019s readiness to leverage them. They are not \u2018autonomous,\u2019 i.e., not able to take actions on their own.<\/p>\nOther instances of 2nd generation AI solutions at Zycus were more strategic<\/h4>\n
Today, we are in the first quarter of 2019<\/h3>\n
I<\/h4>\n
II<\/h4>\n