LegalTech Blog | Legartis

AI in contract review: best practice with Arvato | Legartis

Written by Kerrin Klüwer | Dec 23, 2022 11:41:41 AM

Artificial intelligence speeds up contract review: Arvato Bertelsmann is introducing AI from Legartis in and outside the legal department. Find out why this collaboration has quickly proved successful and how AI can easily be implemented in an interview with Dr. Marc Hansmann.

In what context did you decide to introduce a Legal Tech solution at Arvato?

Before the introduction of a Legal Tech solution, we had a classic bottleneck situation with our contract review. This was mainly due to the fact that only a few experts could review DPAs and this usually under great time pressure. This quickly led to delays in the whole process. We wanted to extend our contract review to less experienced colleagues and find a way to have comparatively standardised contracts like DPAs reviewed outside of the traditional legal department.

What was the result of your deliberations? How did you arrive at a concrete need for action?

Our aim was to shorten the initial contract review process. This task is demanding and takes a relatively long time, because I need to first familiarise myself with a 20- to 40-page document. We wanted to put the initial reading, understanding and commenting in the hands of a software. This way, in the following step, the lawyers have the annotated contract at their fingertips and can jump to the critical points; they no longer have to make the effort of reading through 40 pages from A to Z, discussing them and returning them to buyers or customer managers. So we thought, since this process is repetitive and there should be a certain standard in the company anyway, can’t we automate it?

At Arvato, did you have a clear idea and target image of the right solution for you?

For us, three aspects were decisive for the success of the solution. The review of a data protection contract takes about 45 to 60 minutes, and our goal was for this time to be reduced. The software manages the initial review in about seven seconds, although of course I have to look it over again. 

The second aspect was the question of whether, with this legal software, we could also give the review to non-lawyers. The software makes the process faster for lawyers, but I also wanted to be able to relieve them completely. Our goal was that everyone could use the software without needing a supplementary Excel checklist and still work according to all company standards. 

Thirdly, we wanted to keep a uniform standard in auditing: can the software ensure standardisation while maintaining quality?

So our goals were time reduction, standardisation, user expansion and workload reduction.

Which factors were ultimately decisive for Arvato?

Above all else, we wanted to empower more people to review contracts. A technological solution should therefore be very intuitive and not require any additional effort. Legartis offers exactly this intuitive usage and as an add-in in Word, our users can also work in a familiar environment. The hurdles are extremely low. It takes virtually one click and then I'm in the review. So I don't have to teach anyone a tool, but can work directly in the solution.

The decisive factor, however, was professional quality. At Legartis, you’re mostly lawyers yourselves. So we trust Legartis to develop software that we can pass on. We can empower users with it without having to look over their shoulder. Those were the two most important points: intuitive use in Word and professional quality. And, to some extent, we also relied on the recommendation of Allen & Overy.

Would you like to learn more exciting insights about Legal Tech and best practices? Then click here for our interview with Nicole Steuer, General Counsel of Rexel Group.

How did you implement the contract review software after the decision was made?

In the implementation of this project, it was first and foremost important that all stakeholders were picked up. That's management, that's the teams, the legal teams, but also those outside the legal teams that we want to empower to review DPAs. That’s why we’ve already run manually-reviewed contracts through the software again and found that it produces the same results. Through this exercise we built more confidence in the software and noticed that the AI is still wide awake even on page 60 of the contract, finding things where humans start to get sloppy.

What are the effects of using Legartis internally?

In a word, we’ve become more efficient. I no longer have to read a contract from A to Z, but can be guided by the Legal AI and look at critical clauses right away. We’ve become faster in the initial review regardless of who’s doing it.

The application is also very, very intuitive. The add-in takes you by the hand and tells users point by point what they need to do. You need little or no prior technical knowledge to review a contract. In figures, this means that instead of only two or three employees who used to be able to review DPAs centrally, we currently have seven users who review contracts with the tool. We’ve been able to significantly expand the number of people who can review DPAs.

Something more difficult to qualify, because you have to qualify something that hasn’t happened, is of course the issue of risks. We’ve put our company policies in contract playbooks in order to reduce risks with the software, because with the contract playbooks we ensure a certain degree of standardisation. The seven users total, that is, the four additional users, now definitely review contracts in accordance with company guidelines. They don't overlook anything, they don't forget anything, because they are guided by the playbook and by the AI, and in case of doubt, would also have to ask themselves: "why deviate from the norm on one occasion?"

Does the introduction of Legal Tech also change the way you are perceived externally?

Yes, one aspect we haven't mentioned yet: the legal department is very rarely suspected of being particularly progressive, innovative and digital. By using a Legal Tech solution, we can actually score points as an employer brand in a completely different way. That is a positive side effect. We notice this especially in conversations with candidates, but also with current employees. Working with AI makes you stand out.

Are there any other quantitative results you have achieved in the review process or legal department?

We can't immediately save three lawyers, the use case can't be calculated that way. But quantitatively, the results can be measured very well in terms of processing time: from 45 to 60 minutes in processing a DPA, I have reduced the time with Legartis for the initial review to 10 minutes. The AI is ready in a few seconds, then of course I have to look at the clauses again and consider the AI's suggestions; but after 10 minutes, I’ve gone through the whole contract. I notice the effect mainly because there is time left over to deal with issues that are really critical.

Another thing: I gain 100 per cent of resources because I can expand the user base for contract review. I can enable employees who are already in the company to carry out the initial review with little effort. This saves a lot of time in the company simply by being fast, agile, and not slowing down processes which might have been the case before where the lawyers first needed to have time for 60 pages worth of reading. We are simply much more agile. 

Thank you very much for mentioning these concrete figures. In summary: what are your key takeaways from the collaboration so far?

We have to include all stakeholders when we introduce AI. Working with AI is not yet commonplace everywhere today. We have to get people on board emotionally; the technical side is not the problem. But we have to let the users build up trust in the AI until to some extent they rely on the software. That takes time and must not be underestimated.

Do you want to know more about automated contract review and the implementation of Legartis in Legal Departments? Feel free to download our detailed solution brief. 

 

The interview was conducted verbally in the webinar with Arvato Bertelsmann and Allen & Overy on 17.11.2022.