One thing that legal, contract management and human resources departments have in common is the need to deal with a lot of documents. While contracts, purchase agreements, performance reviews, benefits claims, invoices and resumes may look alike on the surface, the information inside is generally specific to the situation. That means human labor is required to understand what’s in them. This profusion of unstructured and semi-structured information can make finding, comparing and analyzing data a nearly impossible task.
Humans also make mistakes, particularly when details are involved. The International Association for Contract and Commercial Management has estimated that poor contract management costs companies about 9% of revenues on average. Common problems include missed deadlines, disagreements over contract scope, performance failures due to over-commitment and pricing disputes. It’s challenging enough just to keep track of the latest version of a document; analyzing thousands to find patterns or spot problems is beyond the scope of even the most seasoned professionals.
This is where the new breed of artificial intelligence technologies like natural language processing and machine learning have exciting potential. Computers that understand conversational speech can interpret even complex queries to provide rapid access to information that would be difficult for people to find using standard search terms. Machine learning is a complementary technology that starts with a simple set of rules and repeatedly cycles through a database of information to identify common patterns and learn new ones. For example, starting with a knowledge base of language that is common to problematic contracts, machine learning algorithms can scour millions of similar contracts to find other warning signs that humans might miss.
The combination of these tools has been demonstrated to surpass the ability of even domain experts. For example, one study that compared the performance of 20 experienced lawyers with contract management software that was trained to spot problems in nondisclosure agreements found that the computer achieved an accuracy rate of 94% compared to 85% for the lawyers. And worked three times faster.
AI is ideal for environments in which large amounts of data are involved. For example, machines can sort through invoices from multiple vendors in different geographic locations to identify opportunities to negotiate better prices. They can monitor dates and delivery promises to alert contract managers of upcoming agreement expirations or performance guarantees. They can even be trained to sniff out traits that are common to fraudulent warranty claims to reduce the risk that such claims will be processed in the future.
In the HR domain, AI is already making a dramatic impact on the way organizations hire employees and manage their performance. Software can scan batches of resumes to flag the candidates who most closely match job requirements. They can also look for inconsistencies that indicate candidates to avoid or check applicants’ claims against public records.
Natural language processing technology can even accelerate the process of vetting qualified candidates by engaging in Q&A interviews using. HR organizations can use the same language recognition capabilities to enable employees to fulfill their own information requests, such as understanding benefits or filling out claims. Pay equity can be improved by machine learning algorithms that compare salary data to job descriptions and performance histories and take into account local cost-of-living statistics to ensure that employees are being compensated fairly.
Thanks to cloud computing, these and other advanced AI capabilities will be delivered as a service, making their revolutionary capabilities available to everyone.
Download a new white paper from market research firm Emerj to learn more about how AI is helping automate key business functions.