Share It
As a consultant on AI technology for some of the legal industry’s leading companies, I spent a considerable amount of time converting first-time users of Artificial Intelligence. However, it always amazes me that despite the investment companies make in advanced analytics, some users are still hesitant to implement AI within their standard workflow. According to a recent LexisNexis study, 92% of companies plan to increase their use of legal analytics in the next 12 months. So why are some not fully adopting AI into their business?
I remember consulting with a mid-level associate at a top firm in charge of her first large-scale document review. The partner on the matter gave her autonomy to run the project. Among many complexities, there were tight deadlines and the parties needed to agree on a discovery protocol. My pitch to use AI (specifically, Continuous Active Learning) was met with some trepidation, as she had no prior experience with the technology. It required me to do some education and mitigate her perceived risks in using this process. In the end, the anticipated value she would gain from AI far outweighed any of her concerns. After the project, we had a great story to tell and she was eager to brag to her colleagues about the outcome.
Investing in technology requires a lot of time and effort from both a financial and an implementation perspective. Think of all the hours you’ve spent researching, configuring, training, marketing, and managing your company’s AI technology. Yet, for all that effort, you want to ensure your clients leverage the full potential of AI because you know the value it will bring to your clients. Maybe you are just starting with your first AI project. Perhaps you’re switching to a new product, or maybe you’re relaunching an underutilized product.
The most important thing you can do to increase your adoption of AI technology is to have a systematic approach to engaging more clients. Having a plan ensures that you’ve accounted for potential risks or areas of resistance. Attempting to increase adoption without a proper adoption plan can have the following adverse effects:
- Improper alignment between opportunities and AI capabilities;
- Client’s expectations do not match project outcomes;
- Barriers to entry in the use of AI remain high;
- Projects veer off course and lead to unnecessary errors or delays; and (the most critical)
- Little to no realized value.
I’m not going to address things I would consider as indirect steps to increase adoption. Things like webinars, lunch-and-learns, best practice guides, and blog posts are all great ways to showcase your AI capabilities. Still, they seldom have an immediate and direct impact on expanding adoption. If you are looking for something you can implement today to boost your AI adoption, these five steps will help you achieve that goal.
Step 1 – Find the Right Opportunity
In the early stages of your adoption plan, you need to be selective about finding the right opportunity. I’d much rather fish in a lake I know is stocked with bass than throw my lure in the middle of the ocean. Proper alignment between every opportunity and your AI solution is the first step to ensuring you’ve secured an AI devotee. So how do you know that you’ve found the right opportunity?
Consider external forces. Depending on your maturity level using AI, be careful to consider the project’s external forces. For example, is the project low or high stakes? What are the turnaround or time constraints? Are there any procedural constrictions (e.g., discovery order, judicial direction, etc.)? What is the nature of the adversarial relationship between opposing parties?
Take a look at the data. You’ve heard it before when talking about data – garbage in, garbage out. That still applies here. Be mindful of the amount of data in the project to ensure there is enough for the efficiencies gained by AI to make an impact. When I’m evaluating a new opportunity, I’m looking at the content of the data. For example, are there a lot of images, semi-structured data, foreign language, etc.? And I always like to take a statistical sample of the data to estimate the prevalence of my relevant documents. Too much or too little of a good thing in AI is bad.
Know your audience. There are several personas to look out for when assessing a new opportunity. According to American sociologist Everett Rodgers’ book, Diffusion of Innovations, there are five types of adopters for new products. So first, segment your clients into the adopter types listed below.
- Innovators – these are your forward-thinkers who have past successes using advanced technology and have an appetite for risk. They can sometimes have some connection to the scientific discipline, such as Intellectual Property litigators.
- Early Adopters – look for younger users with a high interest in learning about technology and in need of some successful outcomes. These are your thought leaders or those who are socially active in the legal industry.
- Early Majority – these users are typically open-minded, have some influence they can wield, and tend to be conservative when adopting technology. These people stay close to Early Adopters, who leverage their opinions when deciding to adopt technology.
- Late Majority – you’ll find this persona type to be very risk-averse, usually tenured at the company, and unlikely to be very social (i.e., never go to any of your AI training).
- Laggards – have the “this is the way we’ve always done it” mentality and are usually forced into decisions. They should be the last on your list for adoption. Look no further than the user sporting a Blackberry.
Second, determine how many of these users are currently leveraging AI, which will give you a baseline of who to approach first. Again, start with innovators and end with laggards.
Step 2 – Provide Upfront Value Proof Points
Now that you identified the right opportunity, you can focus on pitching AI to your team. Some people hate pitching ideas because it feels “sales-y” – I’m one of those people. But when you know your suggested solution will provide real value, you must educate. Leverage past success metrics or ask your service provider to supply you with some. Here at Casepoint, we have several metrics and references we give to our clients. These are a few you want to make sure to address.
Increased Accuracy. Your clients will spend less time sifting through meaningless data and more time on important information. I’ve seen first-hand how AI improves the consistency of decision-making. The time and space between interconnected data points are shorter which leads to more consistent calls.
Increased Efficiency. Your team will learn faster by leveraging AI. Your client will benefit from reducing the amount of time associated with understanding content within the data and issue-spotting. The team needs to know what is in the data set as quickly as possible and the best way to do that is to use AI.
Reduction in Costs for Low-Value Tasks. Of the many goals and benefits to AI, the reduction of the overall cost is a clear driver. . The true goal related to cost is to reduce the cost of low-value tasks. For example, there is a significant reduction in the time spent reviewing low-value information (either wholly or marginally non-relevant) when using AI. To illustrate this point, let’s assume we have a corpus of 100,000 documents – of which only 20% are relevant. By reviewing the entire corpus, your client will spend most of their time sifting through irrelevant documents.
Step 3 – Remove the Barriers to Adoption
Okay – your team is now sold on the value gained using AI but is still apprehensive about the technology and workflow. Your next step is to remove these barriers to adoption. There are several recurring themes I see when trying to expand AI adoption.
It’s Overly Complicated. Despite the widespread use of AI, some clients still feel that the technology or process is complex. Sometimes the process can appear disjointed, especially if your end-user needs to leverage disparate systems. In addition, you may hear the workflow itself has too many steps or conditions. A clear sign of a complex workflow is the need for a subject matter expert (“SME”). Here at Casepoint, we’ve made a concerted effort to address all these aspects through our single AI-integrated platform, multi-step simplified workflow, and minimalist user interface that provides you with the information you need. As a result, our system can guide you or your clients through the process without the need for an SME.
It Costs Too Much. Here is a barrier that is entirely antithetical to why we are pushing the adoption of AI. Leveraging AI should save, not increase, costs. Avoid any add-on or bolt-on technologies that can come with a hefty price tag. At Casepoint, we understand this is a potential barrier to adoption, which is why our end-to-end unified cloud solution not only provides clients with a seamless experience with AI but is included at no extra charge. Also, be on the lookout for additional service fees that come with using more of your current solution.
Perceived Risk of New Process. Let’s begin by stating the obvious – change is hard. If litigators already associate using AI with some level of risk, it is incumbent on you to remove any of these concerns upfront. You can do this in a few ways. First, make sure you have a simple, documented, and repeatable workflow to share with them. Second, make sure the workflow is adaptable to various use cases. For example, in a CAL workflow, your steps may be different for a broad discovery response than a time-sensitive internal investigation. Third, use a product with a fast, unified, and simple interface. Lastly, leverage your past success stories using AI. If you are just starting, ask your service or technology provider for examples.
Step 4 – Proactively Monitor and Report on The Process
You’ve successfully identified a new opportunity, communicated the value of using AI, and addressed any concerns. Now, you need to back up all of your statements with action. This step is one of the most important because it builds credibility between you and your client (internal or external). It lets them know that you are equally invested in a positive outcome and will work hand-in-hand with them to achieve success. Additionally, it creates an opportunity to guide clients through the process, if they need it. Let’s look at some things you can do.
Quick Status Checks. The operative term here is quick. You want to provide periodic and brief updates or check-ins to ensure they have a positive experience. The quick status check is your opportunity to solicit feedback and address any in-flight issues that may arise.
Standardized Reporting. Set a preferred formal reporting cadence with your client and send a standardized report showing AI performance. If the report’s consumers can’t figure out what the status is in the first 10 seconds, they may find little use for it. Since you will likely send this via email, picture them opening an attachment and trying to read multiple tabs of a spreadsheet. They’ll be frustrated or may feel the need to call you to interpret the information. Make sure the information is easy to consume, is one to two pages, and contains visualizations. Establish a cadence beforehand to avoid inundating them with too much information.
Provide Encouragement or Course Correction. Try to become an active participant in their project. Look for positive outcomes or potential issues when you prepare the report. If you find something that exceeds expectations, make sure to let the client know immediately. Alternatively, if you see something that could have adverse downstream effects, take swift action to alert and address those issues.
Step 5 – Demonstrate Value Realization Through After Action Report
The most critical step in our plan to expand AI adoption is to deliver client value and ensure that value is known to the client and realized. Let’s assume everything up to this point has gone to plan. But, what good is all of that work if your client can’t articulate what value they gained, or worse, doesn’t even realize the process was valuable? How do you ensure these two things don’t happen to you?
Show Operational Efficiency. You should have an idea of the efficiencies gained through the application of AI over a manual process. Things like reduction in the level of effort to understand data volumes, increase in throughput due to better organization of data, or higher accuracy in pinpointing critical information are great examples. Be sure to have a baseline to compare these metrics. If you don’t have some benchmarking metrics, you can ask your service provider. Here at Casepoint, we’re happy to provide our clients with some battle-tested benchmarks and targeted data points.
Document Cost Savings. Using some standardized metrics, build an estimate that projects the total cost had you not leveraged AI. You can compare this figure against the actual incurred costs to highlight the financial savings. Don’t limit yourself to just this opportunity. Do your research or ask your client how many other projects would benefit from the same workflow and summarize those savings. Give your clients a view into what the future would look like if they implemented a repeatable AI process.
Demonstrate Improved Accuracy. Artificial Intelligence is exceptionally good at increasing the quality of work product through precision improvements. Look for instances where agreement ratios between algorithms and operators were in sync and communicate how often this happened. When we use AI, it enhances our ability to separate the wheat from the chaff while simultaneously focusing on documents with similar content. For many of my clients, this proves to increase the consistency of their work product.
If you follow these five steps to increase AI adoption, I’m confident you will have a great story to tell others. However, sometimes having a great story to tell is not enough. The ability for the client to retell that story is a critical piece to expanding AI adoption. I’ve found it far more effective for someone other than myself to speak to the successes we’ve attained together. To do this, you’ll want to make the process easy for your client to share the great news with others quickly. If possible, create a visualization or infographic that is easily digestible and copied into an email with little effort. At a minimum, you’ll want to send some form of communication to your client, thanking them and summarizing your success in a clear, concise, and compelling manner.
I’d be remiss if I didn’t point out that Casepoint’s technology and services will help you achieve all five steps. As a Client Success Manager, it is my job to ensure Casepoint’s valued customers achieve success in every aspect of their journey with our company. To learn more about how we can work with you to help drive AI adoption at your company, please contact us!
