Triage - the most underestimated word in Legal Operations (part 2)

In the previous post in the series, we looked at why the triage process is so important to an efficient and effective legal function, and why it can sometimes be neglected. We then delved into the triage process, examining the first three stages in more detail, and noting how we can make these as efficient as possible. In this next post, we’ll complete the detailed examination of the triage process and summarise how a Digital Operations Platform can be used to create a best-in-class, efficient and effective triage function.

So far we have covered the intake of the work instructions, their review and their prioritisation. Next we need to collate any additional materials and information to allow the lawyer to work on the matter.

Digital operations platform

4. Collate Additional Materials / Information

Before the lawyer gets hold of the matter, it is important to ensure that he/ she has the right information available to support his/ her analysis, and that all additional materials that may be required are attached to the matter. Therefore the triage agent must be able to assess what type of matter it is and what else the lawyer is going to require. This, in most cases, can be straight-forward, especially if the incoming work is of a similar nature, e.g. property leases to renew. If the work is to be done manually, a simple playbook will suffice, but there is also scope for automation if the work is repetitive enough; through workflow, the additional documents can be sourced from document repositories or external sites. If information is simply missing from the original work request, then automation can email the originator to request the necessary documents.

If there is variability in the incoming matters then either more manual effort will be required, or machine learning can be introduced to work alongside the rules-based automation. ML can be used to classify the type of matter so that the standard additional documents required can be identified e.g. it’s a lease renewal therefore we need ownership details from the Land Registry. Artificial intelligence can also be used to identify any unstructured information on the work request or email that may be suggesting additional documentation; Natural Language Processing (NLP) can be trained to identify extra comments in emails, for example, that can be codified into a document request.

5. Allocate Work Request to Legal Resource

Now we are ready to send the matter, along with all the associated information and documentation to the legal resource to process it. This can be through a case management system which is ideally integrated into the lawyers preferred way of working (usually Word).

But it is usually more complicated than simply putting the matter into someone’s work queue. The legal resource that is taking on the work could be internal or external to the business, for example if some tasks are outsourced. We also need to ensure that it goes to the person with the right skills, and that they are available to do the work and do not have too much on their plate already. Which means we also need to know roughly how long the work will take. And we need to make sure that we are using the most cost-effective resource we have, especially if there is a choice between internal and external, and where there are SLAs that need to be met.

The simplest solution to allocate the work out is to use a round-robin system, but this wouldn’t usually satisfy all of the above criteria. A simple automated rules-based solution could do a good job of balancing all of the different priorities so that each lawyer had an equal amount of work, and that they were matched as well as possible against their skills. However, things get really interesting when we start to feedback the performance data from the lawyers; in this way we can start to tune the estimation and prioritisation models as we find out who does which sorts of work better, or who does them quicker, or who copes better when workloads are high, for example. Which brings us on to the final stage of the triage process.

6. Monitor and Report

In mature legal functions there are tools to measure the performance of the lawyers beyond simple time recording. This is important when SLAs are in place, and to ensure that the work is being done as well as possible. The best types of tools are those which are unobtrusive and work invisibly alongside the lawyer, (so no having to press the ‘chess clock’ button on the screen, for example, every time a bit of work is completed). If the work has been set up well from the start (see stages 1 to 5!) then the monitoring and reporting becomes much, much easier, as well as being inherently more automatable and effective. The nirvana here is being able to automatically allocate work to lawyers based on an estimation and prioritisation model, which is then continually updated based on how well the lawyers responded to those pieces of work.

Estimating how long the work is going to take means the model learns from lots of previous, similar examples. It also relies on the triage system (whether people or computer) being able to classify the work accurately. By continually feeding back how long the work took compared to the estimate, the model can be further refined. It can also help identify ‘outliers’ where a matter may have taken longer than expected but due to the lawyer not having the requisite skills or experience, or having too much much work to deal with. All of this information is crucial to making the department as effective as possible.

Digital Operations Platform

In all of the stages we’ve described above, including the first three from the previous blog post, we have been calling on different automation techniques to help make the process more efficient. We’ve talked about portals, robotic process automation, machine learning, natural language processing and good ol’ workflow. But we need to be able to organise and orchestrate these different technology elements so that they can work in unison with each other. Which is where the Digital Operations Platform comes in.

As well as natively enabling much of the key automation functionality referenced, think of the platform as the glue that sticks all these different parts together so that they work collaboratively and efficiently together. And running through the whole process, from start to finish, is the data, which the Platform manages and controls. In triage, it’s vital that we get the data in the right format at the beginning and that we persist it right the way through the process, so that we can extract the inherent value that it provides. That data needs to connect to, and be processed by, the automation tools, as well as providing the inputs that are used to train the ML systems. It’s also the data that the lawyers use to process the matter, and that is used to monitor and report on performance. A Digital Operations Platform, therefore, enables a triage function to operate at maximum efficiency and creates an environment for the lawyers to be as effective as possible.

For more information on how a Digital Operation Platform can transform your triage function, get in touch or register to attend out ALXLegal automation workshop. Read more here.