Simpler: Building a Robot and How to Think of ADR Tech

Kluwer Arbitration Blog
June 16, 2020

Please refer to this post as: , ‘Simpler: Building a Robot and How to Think of ADR Tech’, Kluwer Arbitration Blog, June 16 2020,

Not all technology is born equal. In 1739, the French inventor and artist Jacques de Vaucanson unveiled a head-scratching automaton that fascinated contemporaries for more than a century. It was a ‘digesting duck’, which had the apparent ability to eat kernels and actually digest them. The invention was a trick, of course, and it was ultimately debunked, but what is interesting is that the first time anyone was curious enough to take a screw-driver to it and have a look inside was in 1844, a century later.

Vaucanson’s duck is epitome of something that technology does to us often, even to this day: it kindles the imagination and human hope of mastering matter. It taps into the myth of making the universe ‘light’, as semiologist Roland Barthes would say in the 1950s by way of the sinuous design of the Citroen DS. This fascination is often independent of any apparent practical benefit. For instance, you can easily get perfectly good biological ducks to do the same trick as Vaucanson’s contraption, consistently, robustly and relatively cheaply.

Technology, in other words, can take a life of its own in human imagination and can take over our more immediate human goals. This can happen with legal and dispute resolution technology. But it need not to, and the alternative is not glorifying the ball-point pen, but, instead, building tech from a human perspective: tech that is ‘legal-by-design’.

Technology is becoming a second ‘discipline’, alongside law, animating arbitration proceedings. This is not new or COVID-19-related, but, certainly, not being able to meet in person for months had removed much of the resistance to technology in ADR proceedings.

What is perhaps less attended to is that arbitration, unlike litigation, requires prior agreement to even get off the ground. Adopting an arbitration clause can be difficult enough during complex commercial negotiations and the dispute resolution clause is often among the last to be tabled. Adopting an arbitration jurisdiction can be especially challenging for already concluded contracts that do not contain an arbitration clause but would otherwise benefit from one. Take, for instance, Brexit and its uncertain effects over the enforcement regime across the UK-EU border looming no later than the end of 2020. National court jurisdiction may raise concerns with legal risk management that the New York Convention 1958 does not. But there are other situations where switching to arbitration might make sense, for instance where the current dispute clause is entirely silent on jurisdiction or the choice is no longer adequate out of concerns with political instability or corruption.

Switching to arbitration in case of existing contracts is challenging because it requires persuading counterparties to change status-quo, perhaps mid-performance under the underlying contract. This can be disruptive and, possibly, costly. It may be resisted even out of inertia or risk-avoidance. Crucially, receiving a request to amend a disputes clause in an active contract may ring alarm bells of distrust, pre-empting a cost-benefit analysis which would otherwise counsel in favour of agreement.

Can technology be employed to help with these challenges? How would one proceed to develop it? I will share here some recent experience with building such technology and what we learned about legal, and especially ADR, tech, in the process. The project is called sArb (for ‘Simplified Arbitration Reference Facility’). Its firs incarnation lives here: and it was a pro-bono collaboration with BIAC, an Arbitration Institution, and UiPath, a robotic process automation company.

The first insight was therefore this: Concluding an arbitration agreement for an existing contract requires removing distrust so that the parties may be free to consider their best interests. This is something we learned from mediation. In mediation, often the job of the mediator is half done by changing the ‘architecture’ of the exchange from a two-player escalation game to a three-party discussion.

The second insight was that there is one way to reach agreement but a thousand ways to derail it. This is an insight that we took from the ‘nudging’ literature. In the words of Nobel Prize behavioural economist Richard Thaler: “If you want people to do something, make it easy.” Removing all unnecessary barriers to agreement is not just desirable, but key and each practical hurdle can act as a bottleneck to agreement or as an excuse for each party to fall back on the status-quo.

The third insight is again from psychology, namely that: in dispute resolution, what makes the process legitimate is the fairness of the process and the respect and voice given to the participants. There is long-standing empirical research on this, especially in crime prevention and criminal adjudication, but it is also common sense: if people only cared about winning, half of all litigants would invariably abhor the legal system. This is not what we see. If we conceive our challenge to facilitate arbitration agreements covering existing contracts as one of mediation, then this insight should also apply to it.

With those insights in mind, to overcome the challenges of switching to arbitration in case of existing contracts the idea behind sArb was to build an online facility that helps parties safely propose and conclude arbitration agreements. Following on the first insight from mediation, this is done with the insertion of an Arbitral Institution into the exchange in order to disable unnecessary friction and distrust between parties. The proposal goes to the Arbitral Institution and is then channelled in the name of the institution to the counterparty. The Arbitral Institution acts as a focal point of trust.

The second insight required a focus on facilitation. To make it easy and scalable, the process deploys state-of-the-art robotic process automation (RPA), document assembly and electronic signature collection, from end-to-end. The robot assembles the agreement, provides information in the name of the facilitating institution and collects signatures electronically. This simplifies the hassle and bustle of traditional contracting by orders of magnitude.

The third insight requires that we ensure that the parties’ freedom and control are preserved throughout the process. Freedom and control must trump the logic of automation. The proposer must be free to safely and painlessly withdraw their submission up until signing the Agreement but can rest assured that their signature will not prejudice them if the agreement is not completed. The counterparty’s choices are also clear from the outset: it can ignore the proposal completely or it can act on it knowing that while the document is available for signature, it has not been withdrawn.

In sum, the process assembles the document and correspondence, sends out information to parties and ‘talks’ to the electronic signature provider in a seamless flux that makes it easy to agree and painless to disengage unless all parties are on board. The cost savings of using automation allows making the process available for free. The hope is that it will ‘nudge’ businesses to their own benefit, while maintaining them in full control throughout.

Finally, building this process has forced us to think hard about legal tech, especially in ADR. It taught us in particular that it needs to be less about tech, and more about humans. Start from humans, their sensitivity for autonomy, fairness and transparency, which permeates much of the practice of law and adjudication.

If I had to boil it down to a checklist, I would say that a valid ADR tech proposal must tick at least the following boxes:

  • Take human psychology seriously;
  • Facilitate rather than complicate; and
  • Preserve the parties’ freedom.

How much of the future of technology in ADR will be about leveraging human institutions and intuitions and how much about ‘digesting ducks’ will depend as much on technology providers’ familiarity of the legal world, as on the ADR community itself.