The article Investing in AI? Strategy Principles Say Go Lightly and Later by Hugh Simons in the American Lawyer (18 May 2018) argues that law firms cannot gain a sustainable competitive advantage by investing in artificial intelligence (AI). I question his thesis in this post.

Mr. Simons, a “Ph.D. is formerly a senior partner at The Boston Consulting Group and former chief operating officer at Ropes & Gray,” explains that strategy generally has two elements: hygiene and differentiators. The former is the base level attributes that clients expect all firms to have. The latter “are the elements of a firm’s offer that clients perceive as being different across law firms and where such differences are of value to a client so that they incline a client to choose a particular firm or pay a higher price.”

He then asks if AI can create a sustainable advantage and answers in the negative:

” I don’t see how it can. The AI systems will be provided by third-party vendors selling to all firms. Anything special one law firm can offer clients today will be offered by al firms tomorrow. The window for being differentiated, if any, is short.”

I Tweeted this article on Friday, asking if we should draw the conclusion that firms should build their own AI. Ed Walters, CEO and co-founder of Fastcase, offered up the first answer:

Alex Smith, an Innovation Manager at Reed Smith, offered his perspective on whether firms should build their own AI:

A couple of others chimed in, agreeing with Simons that one cannot gain competitive advantage from AI and firms should focus on offering clients the best service by whatever means necessary. Personally, I am on the side of Ed and Alex: I do see how building tools, especially with proprietary data or expertise, can create sustainable advantage.

Noah Waisberg, Co-Founder and CEO of Kira Systems, did not mince words in disagreeing with Simons:

I like this point and agree with Noah. I have argued in the past that large law firms have few apparent advantages from size. Recently, however, in the age of Big Data, I tuned that argument to acknowledge that access to vast amounts of data in bigger firms can be an advantage.

I Tweeted back to Noah and he replied. I regret not making my Big Data point but, in any event, Noah’s reply about sustainable expertise advantages seems compelling to me. If he’s not right, how do we explain the long-standing, near unassailable positions of top NYC and London firms:

Before coming to a close, I feel obliged to share one other point of view. Kevin Gidney, the Founder and CTO of Seal Software, a legal AI company pointed out that one firm’s data will not be as powerful as an aggregation of multiple firms’ data:

Conclusion

First, on form, This is a great example of Twitter at it’s best, with multiple points of view and a robust conversation. Of course, following it all on Twitter can be hard as Tweet threads can split, as happened here. And not all Tweets may be fully formed or completely argued points (that certainly applies to mine). So forgive me if I have misconstrued (or omitted) any Tweets on this article spawned by my Tweet about it.

Second, on substance, I find the views of Ed, Alex, and Noah persuasive: either by building unique AI systems (or products) or by harnessing huge amounts of data and applying expertise in training, I do think that larger firms can gain a sustainable competitive advantage.

Kevin raises a good point though about aggregating data. I would need, however, to understand the ownership model of the training in more detail to comment further on that.

I see a future where it’s even possible that firm-built AI tools or big-data-powered + custom-trained commercial AI tools become a virtuous circle. What does that mean? Think about high-end practices today. Many firms can handle a deal or litigation but only a handful have the experience to handle the biggest and most complex deals and litigation. They tend to get that work and therefore maintain the experience advantage over time. Why won’t AI be the same and create the same advantages?


Update (21 May 2018) – I shared my blog post with Hugh Simons by email and, with his permission, I share here his reply to this post:

I think it’s terrific that the Twitter discussion focuses on differentiation. It’s great that all agree that this is the key issue. And I’ve no monopoly on truth, so I respect that others’ views differ. I read the counter argument to my point being that the customization of the 3rd party applications will create differentiation. This is entirely logical and reasonable. I confess I don’t foresee this happening though. Why? Because it didn’t happen when people made the same claim with other technologies I’ve seen over the last 25 years–it didn’t with ERP systems (SAP, Oracle, etc.) for industrial companies, on-line banking solutions for retail banks, or e-discovery solutions for law firms. Rather, AI will become table stakes–a must have for sure, but not a differentiator. Hence, the time to invest is when the ‘must have’ becomes a ‘must have’, which is later rather than sooner.