The legal sector has been in a challenging position, since at least the financial crisis of 2008, when the economic downturn hit many of its clients. Notably, this included the insurance sector, a sector that was hard hit due to its dependency on stock market investments, and on which many law firms depend for a large percentage of their fees.
The result was persistent downward pressure on the legal fees insurance firms were prepared to pay. This drove in turn a growing insistence on law firms to demonstrate innovation in their operations in order to reduce their fees to insurance customers at panel selection time.
This event and its subsequent fallout is well documented, and resulted in numerous M&A activities in the legal sector worldwide and in the UK, partly driven by diversification to reduce exposure to a single sector.
The effects of COVID-19 on the legal sector are examined briefly here, in the context of its pre-COVID need for the sector to invest in modernisation, as well as the current and future effects of COVID on legal operations. We will briefly look at how the legal sector can reduce its costs by moving its costly on-premise IT operating model to cloud, and touch upon the benefits it can gain from AI services through automation across the its lower-value legal support services.
Pressures before COVID
Six months before COVID hit the UK, The Law Society’s publication, Gazette, published an article in July 2019 entitled Corporate cut legal function spending. (e)
Based on an EY survey of over 1,000 in-house lawyers across 25 jurisdictions, it found that “82% of companies plan to reduce legal function costs, and 42% plan to do so by more than 10%.” The report also highlighted that 60% of lawyers complained their firms had also “struggled to attract and retain the appropriate talent needed in today’s legal function.”
EY’s law leader was also quoted as saying that the survey supported a “generally acknowledged position that legal functions are increasingly being challenged to do more with less. New operating models will need to evolve to address these challenges.’
Pressures during COVID
Some practices in law firms were hit hard, even in the first few months of the pandemic. There were very few aviation claims, as passenger flights dropped by 95% according to CNTraveller (g). This figure, incidentally, is echoed by this author’s own independent observations of a >90% drop in non-freight flights based on 24x7 monitoring of ADS-B flight data gathered in the greater London, UK region during the first 2–3 months after lockdown in the UK.
Wolters Kluwer, a leading provider of “information, business intelligence, and regulatory and legal workflow solutions for the legal sector” released a report (f) on the impact of COVID-19 on the legal sector, just one month after the lockdown in the UK began. It noted that new workarounds were needed for common physical tasks (signing, reviewing, negotiating, etc.), which “inevitably involves more digital solutions in our daily routines.”
It also commented that “This situation will no doubt be a wake-up call for the legal industry to challenge their processes and align them with their people, needs and scaled digital solutions.”
One survey respondent commented, “COVID-19 is clearly an accelerator insofar as we are required to continue to counsel clients on every day and urgent matters as well as strategic matters, as if we were operating on a business as usual basis.” Another stated, “More ‘digital’ is certainly the right answer to avoid another collapse for legal professionals, but only if we take the time to analyse where we fell short with our processes.”
Pressures post-COVID
A recent survey (e) by The Law Society’s Law Management Section suggests that law firms are “forecasting a 10–20% drop in revenue for the 2020/2021 financial year.”
Whilst some law practices have worked through COVID relatively unscathed, the same report also notes that the Commercial and Corporate practices of law firms were the second worst hit by COVID, with current transactions not completing and new instructions not materialising. Even as the UK government has eased restrictions, a reduced pipeline of new instructions is set to impact fee income, as ongoing uncertainty impacts business confidence.
In addition, with 75% fewer cars on UK motorways (b) during the pandemic and fewer vehicles on the roads generally (traffic levels dropped in the UK to levels last seen in 1955,(c)), a longer term decline in road traffic accident (RTA) claims against insurers has only been reduced further. Road traffic in cities has dropped too, with 71% fewer cars on the road in London, and 75% less in Manchester (d)
Until both society and business return to their pre-COVID levels of activity, which is not guaranteed in the near future, law firms are going to continue to suffer in these and many more areas connected to human activity, such as aviation claims.
Another report by Armstrong Watson (a) concluded that “Many firms have demonstrated their agility in adapting to a new unprecedented normal and the law firms that still remain when this is over will be leaner, more efficient and well equipped to deal with clients from any location in a profitable way.”
What’s a Law Firm to Do?
Three high level IT strategy options are available, with differing consequences and risks:
OPTION 1:Do nothing. This in reality is not an option for long term survival in the legal sector. Doing nothing in what has been a (previously) successful law firm is inviting an acquisition, as seen post-2008.
OPTION 2:Go Cloud Native. Taking everything to the cloud is an attractive option on-surface. However, most law firms utilise numerous off the shelf applications, a myriad of custom-built system integrations and employ IT staff with mostly on-premise technology skills.
Combine this with risk & compliance teams who are quite rightly naturally cautious as part of their job, and this is a long road to take before cost-saving benefits are yielded. Not because of the cloud environment, but because of the internal friction in dotting every ‘i’ that slows down moving to the cloud.
Even a relatively risk free “lift and shift” approach to putting existing applications on cloud provider managed hardware can be difficult, as many application vendors in the legal space are as slow as many law firms in going to the cloud, and still don’t provide a supported cloud option for their software.
In other cases, software vendors may no longer exist, but their software continues to provide value for critical legal business functions and cannot be reliably transplanted into a cloud environment. These and, for instance, those server applications that require access to the underlying hardware in order to install hardware dongles to use their software.
OPTION 3: Pick the “Low Hanging Fruit”. This third way is the least risky of the three options, and one that will appeal to many law firms. Identifying then swiftly automating necessary but costly, repetitive tasks, such as invoice processing can be done at far lower cost and less risk than, for instance, full-blown case management workflow digitisation.
An example of a relatively simple document processing solution I’m currently working on, using cloud storage and three ready-to-use cloud services from AWS demonstrates what can be implemented relatively quickly, including “human-in-the-loop” checking for documents that are borderline cases:
Here, a PDF invoice is saved to cloud storage, turned into plain text by Textract, and its fields and their associated values understood by Comprehend.
Each converted PDF is automatically scored with a confidence rating, and any questionable, low scoring conversions, or previously unseen fields or values are passed to the A2i service for in-the-loop human checking (and correction, if necessary).
Finally, the contents of the invoice are written out to a CSV file for ingestion into an existing process, such as a finance system.
This fastest-cost-benefit approach with minimal risk to existing processes can yield swift returns. It also gives the IT function in an on-premise law firm insight and an appreciation of the benefits and operational requirements of a cloud solution: Cost savings, security, scalability, and the flexibility to adapt to fluctuating workloads without paying for underutilised hardware that is sized for the quarter end or year end billing cycles that plague most law firms.
This approach also gives IT departments a route into cloud, without the perceived risk and long lead times a full blown, immediate commitment to fully cloud native brings, even if a lift and shift is possible. It also mitigates the risk of re-skilling IT staff, allowing them to migrate their skills and gain familiarity with new technologies and operating procedures over a longer period, whilst still gaining business value during that period.
Finally, by choosing to use pre-trained AI services instead of doing their own machine learning, costs are contained and results yielded in a shorter timeframe.
Is Legal Well Placed to do ML?
Mostly, no.
The legal sector was relatively late to machine learning, and typically buys off the shelf solutions, then integrates them into line of business sytems, rather than building their own. Their business is law, not software development. It’s reasonable to state that legal sensibly prefers to spend its money attracting top legal talent, not top ML engineers.
As such, building entire ML based systems is not practical for many law firms, as many of the skills required to build and train ML models are beyond the capabilities of their IT departments.
So what is the alternative to building your own machine learning solutions? Putting aside bespoke solutions that address a single problem, one answer is to use commoditised AI/ML services, such as those outlined above.
Managed AI, not ML
This is a key point to emphasise. We’re not talking here about training your own ML solutions. Instead, we’re proposing the integration of pre-trained ML solutions in the form of Managed AI services into your existing IT systems, bypassing the need for machine learning skills. Managed, because hardware does not have to be selected, procured installed and configured. Building blocks developers can use immediately to swiftly enhance value of existing systems.
This building block approach allows a relatively modular solution to be built using pre-tested, scalable services, on a pay-as-you-use pricing policy. Costly hardware is not over-provisioned to provide for once per year or once per quarter peaks, as is typical in most on-premise legal systems, most notably database platforms. (They can be migrated to cloud too, but that’s a different, bigger discussion.)
Advantages of Managed AI services in Legal
The advantages of managed AI services are simple: They are pre-trained machine learning services, wrapped up, commoditised and ready to integrate into existing legal system workflows with no machine learning knowledge required. So you get the benefits of machine learning, without having to hire expensive in-house expertise to do machine learning. It’s already been done for you.
By way of the example shown above, by integrating services such as these into an existing document processing workflow that is heavily dependent on error prone, manual input, existing staff can be re-deployed or trained up on more meaningful and rewarding tasks, or reduced through natural turnover. Then, as confidence in AI services grows, “assistant” tasks such as contract review and discovery can be performed in a cloud environment, often at a cost much reduced over that of proprietary software solutions.
Conclusion
AI is not a magic bullet. We’ve had far too many of them in IT. However, post-COVID, reducing operational costs and being able to spin up and innovate quickly with modular, pre-trained AI based solutions which present minimal risk, low ongoing cost and low initial investment, should be considered as a level headed proposition for law firms that face post-COVID cost-cutting in their manual processes.
AI services can be viewed as an option for a cost-saving first step by a naturally technology conservative and cautious sector towards the cost savings cloud can yield, without going fully cloud native.
For legal innovators, the case for experimenting to gain business advantage with cloud based AI instead of expensive on-premise hardware can be justified on a cost basis alone, whilst using it as a parallel first stepping stone into broader cloud services, without the risk and cost often associated with such ventures.
References
Title image copyright (c) Wayne Howes from his book, London in Lockdown, which can be purchased at: https://www.howesimages.com/londoninlockdown/
(a) Armstrong Watson, The Impact of Covid-19 on the Legal Sector https://www.armstrongwatson.co.uk/services/covid-19-and-your-business/impact-covid-19-legal-sector
(b) European Transport Safety Council, Huge drop in traffic in Europe, but impact on road deaths unclear https://etsc.eu/covid-19-huge-drop-in-traffic-in-europe-but-impact-on-road-deaths-unclear/
(c) Car Magazine, UK roads deserted in lockdown, as traffic drops to 1955 levels https://www.carmagazine.co.uk/car-news/industry-news/coronavirus/uk-road-travel-plunges-to-1955-levels/
(d) Tom Tom, What can traffic data tell us about the impact of the coronavirus? https://www.tomtom.com/blog/moving-world/covid-19-traffic/
(e) The Law Society, Gazette, Corporates cut legal function spending https://www.lawgazette.co.uk/news/corporates-cut-legal-function-spending/5070803.article
(f) Wolters Kluwer Legal & Regulatory, (Lexology, Apri l28th 2020) Perspective on implications of COVID-19 on the legal sector https://www.lexology.com/library/detail.aspx?g=eda62995-f08e-42a9-862c-b1b19fddc761
(g) CN Traveller, Mark Ellwood, April 2020: Coronavirus Air Travel: These Numbers Show the Massive Impact of the Pandemic https://www.cntraveler.com/story/coronavirus-air-travel-these-numbers-show-the-massive-impact-of-the-pandemic
