Webinar report: Process – getting the most out of AI

AI is a fact of life. No organisation can ignore the relentless progress of technology and remain competitive.

But do we need AI? And if we do, how can we ensure we get the most from it?

Our webinar, Process – getting the most out of AI, provided actionable answers. Justin Turman, an in-house lawyer of 10 years and computer programmer for a further 10, has long experience in adapting technology for in-house legal teams. 

Through his Amsterdam-based business, Automate Office Work, he provides consulting services, custom solutions and education to organisations looking to streamline business operations.

Justin joined Steve Bynghall, founder of content consultancy provider Defiantly Digital – and a founding CLL contributor - to demystify AI for in-house lawyers.

Start with your processes

The best way to plan for AI is not to buy AI. It's to understand your processes. Look to convert the chaos of work into pattern and structure. AI will multiply and accelerate chaos, not solve it. So, the start point is to make every step of your processes explicit. 

Delegate question: What’s a typical example of chaos? A business emergency such as a cybersecurity incident that needs immediate attention. This calls for structure so you can give the business the information it needs when it needs it.

The Bakery Example

To illustrate his approach to pattern and structure in workflows, Justin devised the Bakery Example, a breakdown of the five key elements of a process.

The Triggers

In the Bakery Example, triggers for baking cakes include replenishing stock to required levels and responding to events such as special orders. 
In other words, triggers are both proactive and reactive. In-house legal teams will recognise the parallels between the baker’s workload and the need to maintain contracts and prepare for audits while being ready to respond to a crisis or exploit an unexpected business opportunity.

The Actors

When a process is triggered, it’s down to the actors to make things happen. In a bakery, there may be a store manager who greets customers and presents the product range while bakers in the kitchen prepare the cakes.

Who does what in the legal team should be clear across the organisation. Do internal clients know who to approach? Does one person operate a triage process and assign work to individual lawyers? Be sure to define who the actors are in your team, because if no one – or everyone – is in charge, your chances of success diminish.

The Inputs

Just as a bakery will have a shop, a phone number a website and an order form, the in-house legal team needs appropriate input mechanisms. Give your clients ways to approach you that they are comfortable with.

Chaos at this stage sets the tone for the whole process. Create robust interfaces rather than relying on random emails. A good interface will ask relevant questions that enable you to direct the query to the right lawyer. These questions will also provide analytics that’ll highlight common requests and legal needs across the business and enable to you to structure your legal services.

Also, make this experience identical across multiple interfaces. Whether a client is using an app, a website or talking face to face with a lawyer, the data points being requested should all be the same. This is where you can deploy AI. Online forms and chatbots can be trained to follow the script, capture the data in consistent formats and do away with weeks of email conversations.

Delegate question: How do you get people, especially senior colleagues, who are reluctant to use chatbots to follow input processes? Use a printed order form that follows the unified process. Ask the questions over the phone or face to face, then input the answers as though you are the client. This way, you get the data you need in the format you need while staying sensitive to the culture in your organisation.

Whatever technology you use for inputs, think of AI as the frosting on the cake, not the entire process.

The Processes

The processing stage is the part that most lends itself to AI. In a bakery, AI would remove any friction that stops bakers making cakes. In the in-house legal department, it’s the tool that reduces time spent on repetitive tasks, freeing lawyers up to concentrate on strategic work.

Delegate question: How do you automate tasks that are niche or unique to each user? Focus on higher-volume tasks where clients can self-serve. Enable clients to grade the severity level of their matter, eg high, medium or low. You may need to train the business on making this assessment.

And just as a baker may outsource specialist skills, such as cake decorating, to an external provider, so too will in-house legal teams place work with external counsel. In this event, align their offering to yours so your clients enjoy a seamless experience. 

Delegate question: How should we process information that is too sensitive for AI systems? Create a dedicated system for people to upload private information and give you permission to use it for a specific purpose and period.

The Outputs

The outputs stage is where we switch our perspective from provider to customer. Think back to the trigger that brought the client to you in the first place. Remember, whether it’s a contract or advice you’re providing, it’s part of a wider strategy. You’re empowering the business to move faster with less risk and greater confidence. To maximise your contribution, think about the most effective way to deliver your outputs and integrate them into your client’s processes.

The AI paradox

AI will only work for you when you’ve nailed every element in the Bakery Example. Yet paradoxically, once you have, you may no longer need AI as the tools you already have may well be sufficient to automate your processes. Similarly, jumping straight in with AI before you’ve analysed your processes increases the chances of poor outcomes.

So, when planning any form of digital transformation, make a thorough assessment of your processes and business needs first.

And only then buy the technology.