Alan Dix explains “What is AI?”
Author of Introduction to Artificial Intelligence
AI Business is part of the Informa Tech Division of Informa PLC
This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 3099067.
There appears to be no limit to where artificial intelligence can be used. As AI makes businesses work smarter and improves profitability, the International Data Corporation forecasts that cognitive and AI spending will grow to $52.2 billion in 2021.
In addition to autonomous vehicles, predictive maintenance, and customer-facing chatbots, AI can have an immediate positive impact on the bottom line by helping companies select suppliers that provide goods and services at the lowest price and with the least amount of risk.
Here are some opportunities and challenges of using AI to increase procurement effectiveness.
Spend analytics can be armed with AI software to collect, cleanse, classify and analyse expenditure data to help procurement teams identify excessive costs. For example, AI systems can identify when duplicate suppliers were used to purchase the same goods, when urgent purchases were made without using better terms than existing contracts, and when there were suboptimal payment terms.
However, in order to find these savings opportunities, AI software has to be good at classifying data. Statistical and pattern-based AI techniques can have weaknesses dealing with one-off purchases and infrequently used suppliers. They can also be stumped by new languages and geographies, which happens more and more often as supply chains become global. The best way to achieve ROI is to pilot a system where there is a large volume of transactions involving standard repeated purchases, so that there are more opportunities for increased efficiencies.
By using AI, procurement officers can be armed with knowledge about market conditions, upcoming mergers and acquisitions, as well as real-time product and support comparisons. This ensures that there is a data-driven strategy for awarding suppliers, and that procurement is getting the best possible terms.
Using AI also reduces the time required to analyse all of the supporting data. Evaluating responses to a bid process can be reduced by as much as 80%. It can also be used on a continuous basis to provide recommendations of suppliers on demand. Responding to market opportunities in seconds vs weeks can speed up time-to-market by receiving the needed parts and materials quicker.
Guided buying is another AI innovation that enables employees to quickly and easily buy goods and services from preferred suppliers with minimal support from procurement teams. Employees can use voice-activated commands to find the best price or a supplier that can deliver on time where there is an urgent request. Many of these systems enable direct communication with suppliers with embedded rules to ensure that the buying process is compliant with procurement policies.
Many automatic personal assistants also have the advantage of being able to learn from experience. But if the AI system is self-taught, there is the risk that it can be corrupted by outside influences, so communications and procedures need to be protected from hackers or rogue employees. This was evident in the case of Microsoft’s Tay, which was taught by trolls on social media to use inappropriate language and hate speech prior to its removal from the market for further testing.
The majority of organizations do not have a database containing all of the data in their contracts, and they certainly lack an easy way to extract all that information. As a result, there’s no quick and efficient way to, for example, view and compare agreements.
Using AI, companies can review and organize contracts more rapidly, as well as find large amounts of contract data in order to significantly lower the possibility of contract disputes and increase the number of contracts that they can negotiate and execute.
For example, company contracts can be accessed based on renewal dates to inspect conditions and negotiate accordingly. Finance and procurement teams can inspect if pricing discounts are not being consistently applied across the organization in line with contract terms, or even keep track of the wording of specific clauses in different divisions.
The beauty of AI contracting software is that it helps organizations maintain consistency in the terms and usage in all of their contracts, which makes it easier to identify instances of non-compliance, and make sure that less-than-ideal provisions are dealt with quickly.
However, none of the benefits of AI can be realized without a strong data foundation. Firms need to invest in data management as well as data and analytics, in order to have a 360-degree view of their business operations. Only once their CRM, ERP, and financial systems are fully integrated can they access all the data that is required.
Point-to-point integrations can initially appear to be more cost effective when there are only a few systems connected together. But, in time, with more and more data shared with different departments, suppliers, and partners, a third party integration platform can result in lower development and maintenance costs while providing the scalability and consistent data handling that’s needed.
Once companies have a strong data foundation with all of the necessary integrations and data sharing, new machine learning-based platforms can be used to enforce the best procurement practices. Although today AI procurement systems may not always be accurate, machine learning enables algorithms to learn from data, thereby allowing platforms to continuously improve.
As we begin to see spend analysis platforms classifying data at levels of 98% accuracy—the same level as human analysts—it is more and more likely that AI will become a trusted tool for the procurement process.
Tsipora Cohen is the global head of marketing for Magic Software. Tsipora served as VP of worldwide marketing at Connance, BluePhoenix Solutions, and Keyware Technologies and held responsibilities for strategic marketing, product marketing, channel marketing and marketing communication.
Author of Introduction to Artificial Intelligence