When algorithms review contracts
The shortage of skilled workers is putting legal departments under pressure. AI applications promise relief. Artificial intelligence shortens time-consuming contract reviews.
According to the KfW-Ifo Skilled Workers Barometer, no other industry in Germany would lack as many skilled workers in 2021 as the legal and tax industry. One consequence: while companies are gaining efficiency in many business areas through digital tools and automation, they are losing the pace in their legal departments: Legal audits take time, concentration and expertise. Until now, they were among the tasks that could not be accelerated digitally. However, this is now changing.
Artificial intelligence (AI) is making its way into legal departments. Where lawyers used to have to spend hours reading through documents and shares, smart algorithms are supposed to deliver insights at the push of a button – and compensate for human weaknesses on top of that: Because unlike lawyers, an AI does not tire. Once configured, it delivers consistent quality and virtually eliminates errors.
Up to 40 percent faster
So will artificial intelligence soon replace lawyers? No, but it can already significantly shorten time-consuming examinations. Especially in contract review, AI applications deliver amazing improvements.
With its semantha® solution, the Karlsruhe-based AI company thingsTHINKING is one of the pioneers in the “para” legal tech scene. “With semantha®, lawyers save up to 40% of the time they would otherwise have to schedule for contract reviews,” emphasises CEO and founder Sven Körner. The secret behind the efficiency gain: the AI algorithm is able to understand documents at the level of meaning and can thus assist with various processes involving documents or text.
AI application understands texts
If business partners change agreements or contracts, lawyers have had to manually compare existing and new ones. Although there are tools that display identical or different passages in digital documents, the results are usually not reliable. This is because the applications match texts word for word. As soon as content is formulated differently from the term set of the software or at a different position in the document, it is not recognised.
This is where AI applications like semantha® show their strength: Since the software understands the meaning of words, it recognises content in every formulation. Users can analyse old and new text documents and have defined passages checked, or have innovations and changes pointed out to them.
Man has the last word
Artificial intelligence also helps to identify risks in completely new contracts: If companies link the tool with internal and external databases, semantha® can check contracts for compliance with DIN standards or distribution agreements, for example. The software flags problematic passages. semantha® follows the human-in-the-loop approach, the final check always remains with the user. “But the AI does the preliminary work, which radically shortens the path to the decision,” says Körner.
Mainstream in the future
The quality of this preliminary work is the key differentiator between currently available AI contract review tools. Many algorithms need to be trained over months to successfully match company specifications with new documents. Some tools offer pre-trained AI, which reduces the duration and cost of training. With semantha®, however, this learning phase is completely unnecessary. This is because the AI reads and understands texts in a similar way to a human. This makes the application cost-efficient and also more versatile than other solutions.
In the meantime, business leaders have shed their initial scepticism about AI. Instead, the race to see who can most skilfully use the still young technology to their advantage is on, also in the legal sector. Analysts predict that the legal tech market will grow from 27.6 billion euros (2021) to 35.6 billion euros in 2028. Sven Körner is certain: AI will soon become standard in legal work: “Digitalisation is driving the pace so fast that legal departments without AI will become growth brakes. No company can afford that.”
Author: Paul Henkel. Picture: AdobeStock / Pablo Lagarto
This article appeared in a special supplement of the “Neue Zürcher Zeitung” as part of the “Swiss Innovation Forum” on 19 November 2022.