Automation & AI in Contract Lifecycle Management: Why Pretty Good is Great has been saved
Automation & AI in Contract Lifecycle Management: Why Pretty Good is Great
Targeted automation can go a long way
Use of automation and artificial intelligence in legal, contract and commercial activities are not new things but nor are they universally embraced. Despite a big push to purchase more, user satisfaction results can still be mixed, often due to differing expectations and perceptions.
Despite the proven use cases around automation, or even artificial intelligence, reluctance and rejection still remain amongst some practitioners. Questions raised may be, “well, this isn’t 100% accurate, so a person still needs to be involved”. Or, “this only automates a part of the work so a person is still involved.” In both cases, this may be true but, if we follow that logic, should we also get rid of cars, too? Let me explain.
Cars offer a tech-enabled way to make transportation and movement quicker than a purely manual effort, aka walking. Putting aside climate concerns, cars and other forms of transport have become ubiquitous but, even when going from point A to B, you still have to walk for portions of it. Despite this, most of us see the advantage of having a vehicle in the middle of the process to hasten what would otherwise be a fairly arduous task without the help of this technological solution. So why is there sometimes a reluctance or disappointment in a CLM or other tool that “only” automates 70-80% of a process? We accept this in other parts of our lives, why not here? The “Wave of Automation” to replace lawyers, contract managers and the like may have been oversold over the years. That’s unfortunate, but not insurmountable. There are a number of practical examples where ‘pretty good’ is actually great.
Intake & Triage
In the past, business users would have simply emailed or called legal to ask for support and the legal or contract team would jump into action to diagnose, assess and action the contracting needs for the business requirement. Today, there are technologies that allow for business users to put their requirements into an intake form for personalised assistance. If pushed further, through the use of intelligent design, a first contract can be drafted in a self-service manner. Today’s technology can even allow for a first level review and parse out what needs to be escalated and what does not. In all of these cases, a professional needs to get involved to handle escalations and exceptions or draft more complicated work. People remain involved, but that is a good thing because we still need judgment and business insight. However, I would argue that this is far better than having someone send an email “Can someone help me with this?”
Most of us have probably used an extraction tool or at least received an email or invitation from one of those companies to help you with your regulatory compliance, be it GDPR, IFRS 16, IBOR transition and so on. Originally, and some still do, these tools functioned like “Control F” on steroids. They also had accuracy rates of 70%-80% and we were happy with that to get more day-to-day activities done. A lot of the better tools have evolved and can extract full obligations through a mix of pattern and semantic recognition. This is great for two reasons: (1) obligation tracking is the first big step in proactively managing contracts for better business and monetary outcomes; and (2) in all my years I still have yet to find the person who really enjoys searching for clauses, and then doing cut & paste into a spreadsheet. The hit rates on the better tools are in the 80-90% range and that is certainly better than not doing it at all.
Most IT and services contracts contain complicated KPIs, discount systems, billing rates and other metrics which show how the service is being performed. These are data driven activities that most customers consume through hastily put together slide decks that reduce the data into the ever classic RAG report (red, amber, green). Because of the decoupling of data from the actual requirements in the contract, occasionally you get what many call the “watermelon effect” or all green on the outside, but really red on the inside if you look at it critically. But now – there are tools that take contractual requirements and use those to connect to raw data coming from suppliers. Not only does this show if the KPI or SLA is hit, but also the economic impact and can check if that matches up to the invoice. In order to get this right you need to invest time with helping the tool translate the obligations – see above – and refining the outcome. It may not work exactly right the first time, but with a couple checks it works incredibly well and gives you real information that can give clarity on actual contractual adherence and perhaps re-assess the commercial structure with partners.
There are many more areas where automation and AI can and are useful and they reduce the amount of time a legal professional has to review or work on an activity. That is fantastic because having a professional add their value in the right spot, at the right time is what you need. There are many ‘business as usual’ activities that don’t require an effort (or cost) basis of a high level of sophistication. So I argue that we should use those skills where we need to and embrace where technology really makes it better. And let’s use other skills to make sure the technology is doing what it needs to do to get the job done. Pretty good can be great. And who knows – maybe one day there will be a car that picks you up from your bedroom.