Microsoft Interview – Part One: Contract Data + Standardisation

Microsoft is one of the largest and most technologically significant companies in the world, so what they think about how legal needs are met really matters. Artificial Lawyer asked Tom Orrison, Director of Legal Operations, and Jason Barnwell (pictured above), General Manager for Digital Transformation of Corporate, External, and Legal Affairs, a number of questions ranging from legal ops, to legal tech, to standardisation. Here is Part One of the interview. As you will see, there is plenty of useful wisdom in their answers.

How do you view the importance of extracting contract data to improve insights into agreements? Do you do this already and what benefits does it (or could it) bring to the legal team and the business?

Let’s start with ‘Why contracts matter to us?’ Contracts are valuable assets for our businesses because they contain information that can help us plan and operate our businesses in ways that create more value for our enterprise. Enterprises that turn contracts into data can run a tighter business because they have more insight and control into their fundamental value creation engines that build upon bargaining and exchange to create utility for the enterprise and its stakeholders. Contracts are more than ‘T’s and C’s’.

Contracts are a common unit of business intelligence that encodes and encapsulates the things our businesses do to drive outcomes that create value. The heterogeneity of our formats within and across our organizations makes creating that value hard. These useful data are often trapped in unstructured forms.

For example, scanned images stored with limited, manually-entered metadata. Unlocking this business information often requires humans to find the relevant file, open it, review it, and extract a datum or insight. This is expensive, slow, and does not scale. This is a business problem! Legal departments can partner with the businesses to contribute value that makes us a part of the business.

We can start by building processes that transform business information embedded in contracts into structured data. There are several approaches, but let’s simplify them into manual and automated. We can change our processes to put gates on human effort that put data extraction and recording close to other similar manual work. This has fixed incremental unit cost and works well for lower volumes where adapting for agility is important.

We can change our processes to introduce technology that automatically extracts information in the stream of work or where the assets come to rest. This has diminishing incremental unit cost and works well for higher volumes where scale is important. We can also combine these approaches so machines augment humans. Humans process smaller volumes of unknown elements to create data that let us build models that can be used by machines. Creating structured data is easier when we have defined processes.

Turning contracts into structured data unlocks our ability to reason over contract data with the assistance of technology. We can improve cycle times and discoverability, apply smarter approaches to enterprise risk identification and mitigation, and connect contracting data to other intelligent systems and automated workflows. For example, Microsoft leverages technology to extract contract data from our high-volume supplier contracts. We are still learning how to do this well, but the extracted contract data is currently used to:

  • Monitor Microsoft’s early payment discount program with our suppliers, enabling millions in additional savings on an annual basis,
  • Clean-up and maintain contracting metadata in our repository, improving discoverability,
  • Develop (early stage) enterprise-wide privacy and security risk scoring and monitoring, and
  • Develop (again, early stage) augmented and automated contract compliance reviews, which will drastically reduce contract cycle times.

Technology advances are starting to let us extract contract data from less structured and defined generation and negotiation workflow processes. And we hope to be able to share some of those applications in the future.

Tom Orrison, Director of Legal Operations.

Do you think that standardisation of contracts will grow (e.g. oneNDA)? I.e. that using a standard model that other lawyers and businesses use will be a benefit and drive efficiency, and would Microsoft embrace this / does it embrace it already?

We believe standardisation of legal work is good for everyone. Standardisation drives work velocity because it allows us to build good outcomes on common elements that someone else has figured out is good enough. We should expect that standardisation across the legal industry will increase because it is required for the demand growth we are experiencing and the scale we must achieve. What might standardisation look like for contracts?

Let’s imagine there are two levels of contract standardisation. Standardisation of a specific form (e.g. oneNDA, ISDA, and NVCA) and standardisation of the underlying concepts (required elements are present, expressed in several potential forms). Getting an entire industry to adopt one form is possible, but hard. ISDA is a rare success. With NVCA and oneNDA we still see people changing the forms and adapting them for their needs with many changes having little substantive impact and increasing translation cost. We do not have to fully standardise forms to improve our approach.

We believe standardisation of concepts in contracts is a practical approach that embraces our heterogeneous legacy. If we can reason over an agreement to understand what it means with standard descriptions, then we can evaluate if the ‘intent’ is acceptable. Identifying intent requires decoding the strings of words classifying them to a common set of provision types and identifying the categorical type within that class. For example, identifying an indemnity and recognizing it as uncapped and by how much? 

Standardisation of concepts lets us apply machine learning to operate faster in a world of heterogenous form. Machines can power basic work with less human intervention when we deconstruct contracts at a clause level and bind these elements to a known universe of meanings. We reduce time spent on known repetitive work. This allows our human subject matter experts to deliver their greatest impact and focus on the novel and the complex. We hope to see more standard-form adoption, but we continue to expect form-heterogeneity because of the nature of legal professionals and the challenges of their work.

Legal professionals like their preferred language even if an alternative is substantively similar because of risk and effort aversion. The terms have been ‘tested’ with known outcomes and take less cognitive energy to manipulate. The current heterogeneity of contractual terms reflects our path here. People demonstrated value by negotiating and securing favourable change, regardless of the cost to the ‘negotiating system’ that change introduced.

Legal professionals are trained as critical thinkers and to point out how things are different. To get them to start thinking about the system effect of treating everything different is part of the challenge as we adapt and evolve into a new state with more standardisation using forms like oneNDA. 

We believe the bridge to form standardisation is concept standardisation as an intermediate step, but we also embrace that it won’t work everywhere. Sometimes having that sharp lawyer knife to thinly slice why something should be different is critical for landing a necessary business outcome. Therefore, it is important to identify opportunities like non-disclosure agreements and statements of work that can be standardised and offloaded to less scarce resources. We must conserve human time for the truly complex. Standardisation of form and substance gives legal professionals more time to apply their craft to work that should be bespoke.

Thanks very much to Tom and Jason. Part 2 of the interview with Microsoft will appear next week.

Conference Promotion:

Legal Innovators California 

If you found this interesting then come along to Legal Innovators California conference on June 9, in San Francisco, where we will be exploring a range of subjects, from legal ops to legal tech and much more, with a range of great speakers from across the industry.

Ticket information can be found here.

1 Comment

  1. Quick thoughts on what I just read:
    Standardization makes identifying commonalities easier for the lawyer.
    En masse contracting is easier to control.
    Legal should team with business to harvest data.
    An analog might be prescriptions that work for most of the populace – uniformity that is scalable.

    Where about:
    checking understanding behind the nuances of uniform contracting;
    identification of key diversions when uniformity needs to drill down to match the uniqueness of an industry;
    harvesting data that adapts to the individual to aggregate suggestions during future contracting.
    Uniformity is, in part, effort aversion.
    The world will always be heterogenous and culturally diverse, how does working within heterogeneity scale across cultures?

Leave a Reply to Beth S. Cancel reply

Your email address will not be published.


*