IBM Watson: It’s (Almost) Elementary

Tuesday, May 12, 2015 - 05:02

Because he rubs elbows with lots of self-styled disrupters of the law business, it’s tempting to slap the “disruptive” label on Paul Lippe. He did, after all, help popularize the whole “new normal” thing as one of the ABA Journal’s Legal Rebels. He did develop Silicon Valley-based Legal OnRamp, an innovative online collaboration engine for in-house and outside counsel launched in cooperation with Cisco Systems. And, now, Lippe and OnRamp have become the first implementers in the legal space of IBM Watson, whose high-powered (and Jeopardy-winning) cognitive computing and natural language processing capabilities have spawned predictions that a machine will one day render lawyers superfluous.

So if displacing the bar isn’t disruptive, what is?

Lippe, who doesn’t suffer fools – or fads – gladly, will have none of it. A graduate of Yale College and Harvard Law School, he veered west from the white-shoe path and migrated to Silicon Valley, where he was senior VP and GC of the widely admired electronic design automation company Synopsys, and later served as CEO of Stanford SKOLAR, a medical digital library and e-learning company sponsored by Stanford Medical School. So Lippe has been steeped in innovation for much of his career. That seems to explain why he’s never liked the “legal rebel” label, and why he’s spent a fair amount of time dissing the legal disruption crowd – or at least picking apart facile explanations of the changes afoot in the legal market.

Lippe, taking a more nuanced view, offers two competing labels to go along with disruption: eruption and interruption.

Yes, he says, disruption of the sort popularized in “The Innovator’s Dilemma” by Harvard Business School Professor Clayton Christensen is one possibility. Successful businesses – including law firms – can see their advantages slip away when unexpected upstarts arise and capture their markets. Think Kodak.

But disruption has a Johnny-come-lately quality that undermines its utility as an explanation for change, Lippe says. If you follow the Christensen model, the legal disruption horse strolled out of the barn quite a while ago when corporate law departments evolved from a cheaper alternative to a better alternative for much legal work.

Lippe has even less use for the interruption crowd. These include partners of big law firms who pine away for the pre-2008 era of seemingly endless growth – and dismiss disruption as a corruption of the natural state of legal affairs. Loathe to admit that their too-good-to-last thing is, indeed, too good to last, they hang their hats on interruption and wrap themselves in the warm and fuzzy blanket of a raging deal market.

That leaves eruption, which Lippe calls “an explosion of new service models.” Lippe’s OnRamp is an erupter, and he is quick to point to other erupters he admires, some of them well-known in-house change agents such as Cisco’s Mark Chandler and FMC’s Jeff Carr, others lesser known but no less impressive, such as Carla Goldstein at BMO and David Allen at MassMutual.  

But the most eruptive force isn’t a GC – not yet, at least. According to Lippe, that honor goes to Watson, which Lippe believes will spur the laggards to clamor up the diffusion curve and accept the reality of what’s happening under their noses.

I’ve known Paul Lippe for some years. So I was delighted when Paul, along with Randy Nickel, OnRamp’s VP for Engineering and Operations, offered to give me a peek under the hood of what they’re up to with Watson. What I saw was both more profound, and, in many ways, much simpler, in an Applesque kind of way, than I expected.

Paul, Randy and their team, including partners such as Riverview Law and various law schools, are confronting complexity head on. But it’s not the issue-spotting variety rooted in today’s legal education model. That’s yesterday’s news, Lippe says.

“The basic conceit of law school is that you are uniquely solving this problem,” Lippe says. “That’s not reality. You’re solving exactly the same problem as everyone else in exactly the same way. That’s reality. And variation is not useful.”

Paul’s offer of proof springs from the complexity spawned by Dodd-Frank and other post-debacle financial regulation. Working on a major project for a UK bank (one of the world’s biggest financial institutions), his team is helping it develop a so-called “living will” (a/k/a “recovery and resolution plan”) – a blueprint for dismantling a too-big-to-fail financial institution’s affairs without government support. (Sound easy? Last summer, the Federal Reserve and the Federal Deposit Insurance Corp. flunked the living wills of 11 of the biggest banks.)

Most lawyers would put this type of work on the highest end of the value spectrum – or, if not the highest, far removed from the commoditized end of the scale. Lippe disagrees. He points to the Lehman bankruptcy and its $2 billion in legal fees as work most would put in the brain surgery bucket. He readily concedes that winding down a failing institution such as Lehman is a complicated and important project. That’s not, however, because of the legal reasoning power required. Managing that type of complexity requires a different suite of skills beyond the reach of most lawyers. Not that they couldn’t do what needs to be done, but they couldn’t do it with the speed and efficiency and accuracy required in the face of such complexity.

That’s why it’s better, Lippe says, to let lawyers (or smart people trained for legal-type work) do what they do best and let Watson do what it does best. That’s a potent combo. It is not, however, as some fear, an example of a machine displacing the reasoning function of lawyers.

Lippe and Nickel show me how their team is identifying and sifting through thousands and thousands of documents – mainly contracts to which the bank is a party – that would need tending in the event of a failure. They are pulling them apart – disaggregating them – and organizing, classifying, coding and storing millions of pieces of information. They are then reconstructing them into something that will be useful if and when needed (and will pass muster with the regulators). When the OnRamp team completes the heavy lifting of deconstructing 10,000 contracts, Watson jumps in and makes digital mincemeat of the other 90,000.  

So Watson the Disrupter isn’t really supplanting the legal reasoning function. Rather, Watson the Erupter is supplementing it, but only after lawyers teach it how to “think” about a bank’s contracts.

In doing this, Lippe says, Watson is casting a harsh light on what many lawyers really do – “humdrum processing work” at “obscene” rates. In other words, Watson will show us just how rare “bespoke” legal work really is.

Lippe calls what OnRamp is doing “Legal by Design.” And he says most people just don’t get it.

“In the popular press, this usually gets discussed as a problem of cost or law firm business models,” he says. “But it’s mostly a problem of complexity.”

So it seems Watson is not going to displace legal reasoning anytime soon. (Maybe that will be left to DeepMind Technologies, the London-based artificial intelligence startup founded by Demis Hassabis and snatched up by Google for a cool $650 million.) It is, however, going to expose just how low end much of today’s high-end legal work really is.

And maybe that’s one reason Lippe is trying to interest law firms in partnering with him on helping their clients tackle complexity such as living wills or M&A due diligence. With Watson on the case, it’s elementary: If you can't beat ‘em, join ‘em.

Joe Calve, a former editor and publisher for ALM, leads business development, marketing and communications for McGuireWoods.