Analysis Before Automation

March was something of a legal tech frenzy. ABA Tech Show in Chicago. CLM Simplified Academy and Legal Week in New York. ChatGPT mania. Cecelia. Harvey. Leah. Lots of vendors sharing lots of news. Lots of customers discussing what’s working for them, and what’s not. It gives new meaning to March madness. But amongst all this noise, one thing that really resonated with me was the simple idea that analysis should come before automation, especially for contract projects. So why do so many people overlook this basic rule? 

All too often, CLM projects go like this: a new CLM is selected, and everyone is eager to get something implemented and deliver value. Hoping for a quick win, someone suggests automating a couple of templates and workflows. Someone else is keen to build a clause library and maybe a negotiation playbook. The team gets to work gathering examples and sketching out their current process.  

What about the legacy portfolio of 60,000 agreements? Should we start by loading that? No, we’ve been burnt by that before. Over-hyped AI tools that half work and half don’t. It all ends in massive amounts of human effort. If we start there, it might be months or worse before we see any results (this is, of course, a misconception, but more on that later). 

So, automation becomes phase 1. The team does its best to figure out which templates are the best. Some lawyers opine about which clauses are preferred and which are fallbacks. They argue a bit. They argue some more. You build a few templates and hope that they align with what people are negotiating day-to-day, but who really knows? You hope the new tool is a better experience for everyone, but maybe they hate it and go back to good old Microsoft Word. 

The problem with automation before analysis is there’s too little data, too much opinion and too much guesswork. Your chances of successfully automating anything are much lower when you don’t have reliable data about the status quo. What terms have we negotiated in the past? Which outcomes were we happy with, and was it worth the time and effort? Which deals might go faster with more even-handed language? Which things can we anticipate and pre-approve, rather than routing for approvals that will always be given? 

When it comes to the perils of automation before analysis, examples abound.  Without analysis, we’re left to speculate about how many different variations of limitation of liability clause we’ve agreed to in the past, and how often these differ from our “standard”. Without analysis, we’re not sure how often customers and suppliers accept our unilateral NDA. If we end up signing a mutual NDA 99% of the time, why waste time building a one-way template?  Without analysis, we miss the fact that we’ve accepted 20 different flavors of governing law. So why do we waste time insisting on our NY law preference when we know there are many other low risk options? Without analysis we’re trying to define standards and playbooks without the benefit of real-world data. 

At CLM Simplified, a better approach was proposed during a panel discussion between Lucy Bassli (InnoLaw Group) and Luba Kurbanova (Booking Holdings): don’t start with automation, start with analysis. We now have access to software that can process and analyze tens or hundreds of thousands of contracts in hours or days. Importantly, there’s a new generation of contract analytics software that works well out of the box. We now have a way to understand what our “standard” terms really look like and how often they survived the negotiation process. We can get insight into the data we need to make good automation decisions, and we can do it quickly and cost-effectively.  

Another bonus of starting with analysis is the immediate risk management benefits it delivers. By sweeping your portfolio for clauses and terms you know to be high risk, you’re highly likely to catch things in time to remediate the risk. If you find and fix ten high risk contracts buried in your legacy portfolio, that alone will probably pay for your project. 

Don’t worry. If you’re struggling to automate your contracting process without visibility into your legacy portfolio, there’s hope. Just pause the automation and do some analysis. The data you gain will give you the insights needed to make automation far more effective and successful. These insights are key to achieving “CLM readiness”. Just like the horse and the cart, analysis should always come before automation.