Implications of AI for professional services and law firm marketing

Professionals and techies love acronyms, and the one that has been everywhere this year is AI, standing for artificial intelligence.

PM_Front_Cover_V4.jpgWhen firms with long names moot the idea of rebranding to their initials, I always advise a quick internet search to identify whether those initials have any alternative meaning – such is the case with the initials AI which transport me to my childhood and sniggering in the playground.

There was great excitement amongst the children of farmers when the AI man came to perform artificial insemination, and much mirth as those more familiar with the birds and bees explained to us townies the fertilisation process without the bull.

There has been a great deal of hype about the future of AI in professional services, with numerous headlines proclaiming the decline of certain professions and the loss of jobs to the likes of robolawyers. This article aims to provide a no-bull…. introduction to AI and to explore its many forms and the potential benefits to marketing and business development in professional services.

Some explanations

Artificial intelligence is the theory and development of computer systems able to perform tasks that normally require human intelligence and skills such as speech and language recognition, visual identification and the ability to make decisions. Computer systems, called cognitive technologies, learn by data mining, pattern recognition and natural language processing to mimic the way the human brain works.

It is worth remembering that, at the moment, underneath all this techno-jargon is computer code based on logic – a logic which must first have been defined by a human.

AI has its own extensive language – a useful glossary is available from Tech Republic - and there are many forms of AI, each with its own character and potential uses, descriptions of which are available online here.

Personally, I do not see AI signalling the death or decline of the professions before I retire for a number of reasons.

First, our scarcest and most valuable commodity is expert time and AI will enable the smart professionals to leverage that time more effectively to provide clients with better value. Of course, laggards who dislike change and cling to old ways of working may see their workload decline, but forward-looking professionals will adapt their practices. The labour market is fluid and high-quality people will gravitate to the employers that offer the best tools to work with, including AI.

Second, taking the legal profession with which I am most familiar, AI is introducing its own challenges and requirements for professional support, particularly in law where the regulatory environment struggles to keep up. The European Parliament only called for EU-wide rules on the liability of robots and artificial intelligence in February 2017.

This is positive in that it creates a need for advisers to consult and draft regulations to cope with the innovations, and the innovations themselves create new legal challenges. For example, drones and driverless cars provide new work for lawyers working in areas of transport, product liability, insurance, property, data protection and intellectual property. There will be a change required in the legal skillset with a move from jobs in low value repetitive areas towards complex tech-related areas.

Third, there is a limited supply of AI engineers for the technology sector, and an even smaller pool of professionals with high technical skills. The education sector has some catching up to do to provide a large enough pool of talent to feed potential need. A search of UCAS for a degree combining law with other subjects threw up the following results:

[This image cannot be found]

Law schools are missing an opportunity.

Opportunities for AI in professional services marketing

Taking each of the 7 P’s of service marketing in turn, how is AI being used? And how might AI help those of us that work in professional services in the future?

1. Product (services)

Any activity which is voluminous and repetitive has the potential for automation, but when combined with deep learning from neural networks, the hardware and software can perform multi-layered calculations, more like the human brain, but quicker and more accurately. This provides an opportunity to improve services and offer better value for money to clients.

Examples are emerging in areas such as:

  • planning and scheduling of any service involving regular dates for renewals, reviews, compliance activities;
  • streamlining the workflow for claims in the insurance industry;
  • automated fraud detection and know your client checks in the financial services sector;
  • improvements in forecasting sales demand, capacity management and financial modelling;
  • product development based upon evaluation of client feedback combined with more extensive market research such as sentiment analysis;
  • improvement of document review during audits, litigation and corporate transactions.

The real estate team at Berwin Leighton Paisner have benefited from extensive PR about their use of RAVN’s ACE robot to automate data extraction in standard documents and the benefits they have achieved in terms of increased efficiencies, mitigating risk, cutting costs as well as increasing staff morale.

There is also the opportunity to improve knowledge management using natural language processing. For example, ROSS Intelligence identifies legal authorities relevant to a particular question. In research by Blue Hill Research comparing AI-assisted research with traditional electronic legal research tools, ROSS performed better in terms of quality of information retrieved (percentage of relevant authorities retrieved and percentage of results that were relevant); user satisfaction and confidence; and impact on research efficiency (up to 30 per cent quicker).

In the world of pensions and investment management, Inalytics is being used to determine the link between investment processes and performance, and RAVN is being used by The Abraaj Group to search and access both proprietary and commissioned knowledge sources to visualise and navigate relationships between experts, content, investments, and investors in a way that was not possible before. Lexis Nexis recently announced that they ar testing a chatbot to offer users a more conversational approach to accessing their research assets which comprise around 60 billion documents.

2. People (professionals and clients)

Speech recognition first appeared on the iPhone in 2008. With only around 80% accuracy we all had plenty of funny stories to tell from our trials. In 2015, it had improved to 92% and by May 2017 it had reached 95%, which is considered to be the threshold for human accuracy. It is probably one of the most established forms of AI in use within professional firms for dictation and transcription.

Computer vision helps computers to recognise faces, which has huge scope for security identification (my bank has just offered this option to me).

AI offers an ability to better understand our clients though the automated analysis of a greater range of data, particularly unstructured data such as customer emails, social media, audio and video.

I love the idea of some algorithm crawling through a firm’s client files, billing data and information on the internet to provide me with an accurate and informative data file, rather than having to crunch through lots of poor quality data to get it to a state of a basic mailing list, never mind achieving the nirvana of trigger data for cross-selling.

Personal performance is already coming under the spotlight. Premonition is an example of machine learning which mines court data to identify trends in the performance of firms and lawyers, including the identification of good and bad outliers. The Miami-based company has crunched court data in the US, the UK, the United States, Canada, Virgin Islands, Netherlands, Ireland, India, Australia and New Zealand and provides subscribers with a shortlist of lawyers with a proven track record in a field of law for a type of case. They stress that the lists are indicative, rather than absolute, and other factors such as personal chemistry, cost and availability also need to be considered when instructing a lawyer. This provides real-time data on a lawyer’s performance in a certain court, before a certain and seems to me so much more relevant and useful than the anachronistic legal directories – it is both independent and up to date. Could the days of the legal directories be numbered?

3. Promotion

2017 is the year when global advertising expenditure on the internet is expected to overtake television advertising.

AI is primarily used online to sort through large amounts of user-generated information to establish patterns, and provides the ability to tailor promotions on social media to a degree that was never available with traditional above-the-line advertising. Advertising inefficiencies are immediately obvious, so it is possible to adjust the direction of the marketing budget with speed.

We are all familiar with online adverts which have been ‘tailored’ on the basis that ‘you liked that, so you might like this’. In theory, machine learning will make better and better recommendations, subject to it having sufficient data. At its simplest level, if a customer has recently searched about ‘making a will’ then the machine might suggest that they also need a power of attorney.

However, sometimes common sense fails to prevail. When I added some nappies to my grocery order to donate to a refugee centre, the algorithm wrongly concluded that I needed a whole range of baby paraphernalia.

4. Price

Kira Systems report that contract review times can be reduced by between 20% and 60% for due diligence exercises.

Clearly that represents a huge potential cost saving for the client, which may seem like a huge potential loss to their adviser. However, clients do not always want an absolute reduction in cost, and may prefer an improved value proposition. For example, on a global corporate transaction it may become possible to review thousands rather than hundreds of contracts for the same budget.

One area that the legal profession is struggling with is the growing pressure for fixed fees and for volume work, data analysis will be able to help firms to better understand their cost base and plan or manage fixed fee work.

5. Place

Most readers will be familiar with the power of geo-positioning, if only for their Google results, but this has huge potential.

In the film Focus, Will Smith plays a con man who influences his target with a range of subliminal messages on his journey from the airport to a gambling event. The online equivalent of this is known as hyper-local targeting where messages can be streamed along a prospect’s regular commuting route (I find this quite spooky).

Artificial intelligence is already having on how offices of the future are being designed and savings can be expected in the cleaning of internal floors and waste disposal by autonomous guided vehicles, external windows will be cleaned by drones and even toilets and showers could be cleaned by actroids (a humanoid robot with strong visual human-likeness).

Security and maintenance are other areas where significant cost savings are predicted alongside a reduction in risks for personnel, as an increased use of sensors, cameras and high-speed communications is combined to detect intruders or faults, diagnose problems remotely, schedule replacement of parts, and monitoring and patrolling can be done autonomously.

I am just waiting to hear from my accountant whether a robotic hoover will be tax allowable as part of my research for this article.

6. Physical

Bots are robots without a physical presence. Chatbots (originally called chatterbots) are software programmes which use natural language processing to simulate conversational interactions with humans and use automated processes triggered from these interactions. These replace humans, often in call centres, and are being used extensively in banks in an attempt to provide customer service.

My personal experiences have not been very impressive – as a recent request for a cheque book required me to send the paper form by post, and a need to increase my online daily payment limit required me to go into a bank (the nearest now being several miles away and involving a trip of more than an hour). I was surprised that neither of these requests could have been at least collected by the bot, if not processed. Some joined-up thinking is required.

I know one firm looking to scan a bank of 20,000 wills which seems like an ideal job for a robot.

7. Process

Artificial intelligence is all about process. Just as robots revolutionised manufacturing processes, AI will revolutionise those information processes that can be standardised, where the volume is high enough to facilitate software development; and where there is a clear potential for a return on investment.

In order to train the computer, a key requirement is a large enough quantity of data for analysis, partly because of the set-up costs (although these can be expected to reduce over time) but more importantly to achieve the statistical robustness to underpin reliable output.

When IBM’s Deep Blue beat world chess champion Garry Kasparov in 1997, many saw this as simply a show of brute force as the computer could evaluate 200 million positions in a second, within a very narrow field of expertise (only chess). Today, in the healthcare sector it is considered unlikely that it will be possible to train machines regarding very rare diseases where there is simply not enough data for the machine to learn from.


As with anything new there are concerns, some more real than others. While potential job losses grab headlines (and sell books) two thirds of the respondents to a survey by PwC into US financial services in 2016 said they were not relying more on machines because they were ‘limited by operations, regulations, budgets or resource limitations’.

Research by Deloitte and Efma (April 2017) into perceptions of AI in the financial services sector revealed the following as the biggest concerns:

  • 12% hacking / cybercrime
  • 12% scarcity of technical talent
  • 12% limited understanding of data technology
  • 9% sustainability
  • 9%  existential threat to humanity
  • 8%  no ownership / accountability.

Readers of Professional Marketing will be glad to know that alongside cyber-security experts, data scientists and chefs, the recruitment website Indeed does not believe that robots will replace the need for people to work in marketing, communications and design, as ‘these jobs require social intelligence and new media literacy skills that … are not within the realm of robot programming’ - Phew!

Back to Blog

Get in touch ...

We are always happy to have an informal chat. As a member of the Law Consultancy Network, if we cannot help you we may be able to introduce you to a specialist who can.

Tell us your communication preferences