As my colleague Pedram Abrari recently pointed out in The New Cloud: Where Big Data Meets Machine Learning, AI tools can help enterprises fully exploit their customer data assets to build competitive differentiation and optimize commercial relationships in ways that weren't possible before. We all know how hot artificial intelligence and machine learning are these days. That’s why companies are looking to these technologies to help them squeeze more value from data across their organizations. One potentially revenue-rich target is the data that’s locked up in customer contracts. But how do you get to it? Accurately and effectively?
AI is not enough
By using AI but not forgetting that it needs a second ingredient to work well: humans. Ones that can accurately vet and understand what’s important and what’s not. AI is a powerful tool that can deliver great results when it’s guided by equally powerful human expertise that distills and vets key data within contracts. Without that crucial overarching human intervention and input, the results of an AI-heavy initiative to uncover the right customer intelligence will likely be disappointing.
Some overly enthusiastic advocates of AI seemed to have missed the idea that artificial intelligence is only one ingredient, albeit an important one, in a successful recipe for effectively capturing key data from customer contracts and existing systems such as CPQ, CLM, CRM and billing systems.
I was recently in India where I met our Digitization team and I realized, after being with these experts, just how much machine learning technologies need orchestration by human expertise:
1. To ensure accuracy - AI systems need to be fine-tuned to discover the nuggets of information hidden in the organization’s contracts – all of them, and in whatever form they are presented in the documents. Companies tend to underestimate the complexity of their contract data as I discuss in Turns Out Standard is the Exception. Yet we’ve found that even for a category as simple as Document Type, error rates can be surprisingly high when data extraction isn’t supported by human review and validation as a part of the AI process.
For effective contract intelligence, you need a high degree of confidence that the extraction output is clean, accurate data. And for that, you need an application of human expertise and methodologies.
2. To deliver business value - To provide truly valuable and actionable information that delivers tangible revenue to the business, a data extraction initiative needs to do much more than simply extract language from contract documents and make it searchable and reportable. Indexed summaries of contracts still put the burden on the end-user to interpret the terms and determine if there’s any value. Real value results from human expertise baked into the extraction process to build the business context around the extracted data. Users of contract information need to know the whys and the hows, as well as the whats and whens – in other words, they need to know not just what was agreed upon and when, but also why action is needed and how to proceed.
The result of a 360-view
With a broader view of their customers, users (SalesOps, Finance, Business and Contract Management) can identify and address real business challenges like revenue growth opportunities, missed renewals, billing inaccuracies, and compliance risk issues. They can build a comprehensive, 360-view of customers and maximize the commercial value of those relationships.
Businesses are finding out that when they start out with unrealistic expectations for what AI can accomplish and over-rely on it, they end up with inaccuracies and inconsistencies. They often must bring their people back into the mix to get everything back on track. But here’s the thing, you can’t add a key ingredient after you’ve baked a cake. Far better to make sure right from the start that you have the right ingredients, in the right proportions, at the right time, and that include experts like our team in India.