Deloitte's Generative AI Report PDF

The future of legal work? The use of Generative AI by legal departments Our perspectives and the results from our Generative AI survey for corporate legal departments June 2024 1 Table of contents + Foreword...

The future of legal work? The use of Generative AI by legal departments Our perspectives and the results from our Generative AI survey for corporate legal departments June 2024 1 Table of contents + Foreword 3 + Executive summary 4 + Section 1: The rise of Generative AI and its impact on the legal department 8 + Section 2: The benefits of Generative AI for legal departments 12 + Section 3: The adoption journey 18 + Section 4: The legal department’s workforce 20 + Section 5: External law firms and legal service providers 23 + Five steps every legal department should be taking now 24 + Postscript: What comes next? 26 2 Foreword Generative AI has dominated the headlines since We are pleased to present in this report the findings of a recent Deloitte Legal survey—“The future of legal work? The use of Generative AI by legal departments”. This work focuses exclusively on corporate ChatGPT became the fastest growing consumer legal departments, and we received responses from senior legal leaders across all major industries. application in history. This transformative technology, Their responses deliver a consistent and urgent message: Generative AI has the potential rapidly to transform the legal sector, unlocking huge stores of value for the legal department. Our clients are with its unique ability to generate human-like content, ambitious about this technology. They predict high levels of adoption over a short timeframe, far has inspired individuals and businesses to experiment, exceeding historic technology trends in the legal sector. Our clients are also pragmatic, anticipating to learn and to create new sources of value. It has also potential threats to the workforce and recognizing the need to invest in reskilling their people. engendered a noisy and somewhat febrile atmosphere, These insights support our view at Deloitte Legal: Generative AI’s impact on corporate legal with widespread speculation about Generative AI’s departments will be far reaching and transformational. The technology’s unique capabilities are both well suited to legal work, with its focus on text-heavy, unstructured data—and highly versatile, with impact on our personal and professional lives. broad application across a range of legal activities. This means that the potential upside is significant. However, if legal teams are to unlock benefits in their anticipated timeframes, we suggest that At Deloitte Legal, we want to bring clarity. We are they need to act decisively. Our report offers a framework that every legal department can use to capitalize on the opportunities that Generative AI presents. confident in Generative AI’s potential to accelerate innovation by combining the capabilities of humans We sincerely thank our participants for their time and thoughtful contributions, and we hope you find and machines. This confidence is informed by our the report helpful and inspiring. work with clients: we are helping organizations across Richard Punt industries use Generative AI to re-imagine business Deloitte Legal Global Leader models, build corporate and social value, and inform a vision of the future—safely, securely and responsibly. 3 Executive summary Generative AI is increasingly a top far. We believe it has the potential to drive organizational priority for business leaders genuine, sustainable change in the way legal across industry sectors. A recent Deloitte services are received and delivered. survey of CEOs found that a majority (79%) We wanted to test this hypothesis with our “The impact on legal expect Generative AI to transform their organizations within three years.1 Business clients—to understand their perspective on will be seismic. This is Generative AI, the benefits it can deliver, the expectations are that corporate legal barriers to adoption, and their predictions a once in a generation departments are part of this transformation— both through adopting Generative AI as to how the technology will change the opportunity.” technology in the delivery of legal services legal ecosystem. This was the context for —Richard Punt, Deloitte Legal Global Leader and enabling the business to adopt this new our Deloitte Legal survey: “The future of technology at pace, safely and ethically. legal work? The use of Generative AI by legal departments”. We took the views of However, despite the transformative promise senior legal leadership at 43 of our closest of ‘legal tech’ over recent years, the work and largest clients. We summarize the key of the legal function has remained largely findings in this executive summary. The unchanged. Legal technology solutions have full survey report that follows presents our delivered only incremental improvement. clients’ views alongside our own market At Deloitte Legal, we are convinced that insights on the future of legal work combined Generative AI is a technology of much with our practical experience of using this greater consequence than we have seen so technology internally and with our clients. 1 State of Generative AI in the Enterprise 2024 | Deloitte US 4 Our report explores five areas: professionals. We see this as a positive impact of productivity gains where in-house lawyers will have more time to focus on More strategic deployment of in-house lawyers driving reduced reliance on external counsel (and associated more strategic, interesting and impactful work. reductions in cost). 1 The impact Capability—Generative AI provides opportunities to improve Short to medium term increase in the use of shared A once in a generation step change in productivity legal service delivery, either by elevating the quality of legal service capabilities/alternative legal service providers Our clients are bullish about the technology’s impact for corporate work or by enabling the legal department to offer entirely new (ALSPs) where those providers are early adopters of Generative legal departments. 79% of our respondents believe that services and insights to the business. Better access to, and AI and pass on these benefits to in-house legal teams. Generative AI will have a moderate to significant long-term insight from, legal and business data will enable more data led impact on how legal work is performed. This percentage is decision making and more proactive risk management by Legal Changing composition of the workforce and in particularly striking given the immaturity of this technology, which teams. Anticipated benefits from survey respondents were house/external mix presents a cost efficiency has been available commercially for less than 2 years. slightly lower for this category, reflecting the renewed focus now opportunity of 20–30% needed on good data management and the longer timeframe Almost half of our clients go further, predicting that Generative expected before we see these benefits scale more broadly. Demand on AI will render some legal tasks entirely obsolete. Our legal survey highlights the most impacted legal practice areas are Whilst this technology is still in its infancy, benefits can be realized contracts, legal operations (including knowledge management, today. In our report, we have included Deloitte case studies that spend management, document review and eDiscovery), M&A and illustrate these benefits across a range of Generative AI use cases regulatory compliance. including contract review for regulatory compliance, eDiscovery, M&A due diligence, IP value leakage prevention, service schedule 30% 40% 35% 2 The benefits generation and contracts simplification. More than just efficiency, GenAI will enhance 10% 3 The legal ecosystem 5% 10% the legal experience and unlock altogether new GenAI will drive a fundamental shift in the legal capabilities 40% workforce and the in-house/external mix 50% 45% Generative AI benefits for corporate legal departments sit across three core areas: Whilst there has been a lot of focus in the industry on Generative AI use cases, we believe more focus is needed on the longer term 20% 5% 10% Efficiency—Respondents overwhelmingly scored efficiency and impact on the future shape of corporate legal departments. 46% productivity gains as the biggest anticipated benefit from this of survey respondents expect the Legal department will stay the Typical legal Targeted GenAI Broad GenAI ecosystem today adoption adoption technology. This supports our own experience at Deloitte where same size, but with meaningful changes in composition, seniority we have seen efficiencies of 25–50% across a broad range of and skillsets. We expect a number of key changes in how legal Self-service ALSP/shared services legal use cases. work is delivered that support this view: Automation Lawyers Experience—Two thirds of our survey respondents Increased business self-service and automation, freeing External legal resources Ops and tech staff anticipate an improved experience for Legal team’s business up in-house team time and capacity to meet increases in customers through Generative AI. A similar proportion future business demand without increasing headcount. expect an improved employee experience for in-house legal 5  An increase in operational, technology and change ethically is a critical enabler and accelerator for organizational Although expectations are raised, clients’ confidence in their skillsets and capabilities to implement these changes. adoption at scale. external lawyers’ ability to use the technology to enhance the Lawyers will need to develop new skills to work with this client service experience are low: enhanced service from technology on a daily basis, and new specialist capabilities Combined with the legal department being a high impact, high law firms ranks lowest in the list of benefits that GCs will be required in Legal Operations such as data scientists, opportunity area for use of Generative AI, this provides a real expect from Generative AI. This is an opportunity for a prompt engineering and AI solution architects. An increased opportunity for corporate legal teams to be higher up the priority robust conversation between law firm and client. Corporate legal investment in change management and driving adoption will list for investment than has historically been the case for GCs. Our departments should not just ask their external law firms how they also be critical. survey respondents see Generative AI funding coming from two are using Generative AI, but specifically what benefits they can different sources—both Legal functional budgets and enterprise expect to receive. The winning law firms will likely be those that We see these shifts providing a positive and exciting future for wide AI investments—highlighting the importance of influencing move beyond ‘innovation theatre’ to demonstrating and delivering corporate legal teams. In-house lawyers will finally be empowered internally to access these budgets. measurable pass-through benefits in the form of reduced time to focus on more strategic and interesting work, to provide and cost, or increased value, quality and insight. greater data-driven insight, and drive business growth through 5 The impact on external providers and law accelerated and more agile support. Together, we see these ALSPs are particularly ripe for Generative AI automation given the firms changes bringing about a cost efficiency opportunity in the typical nature of scaled, repeatable, standardized work. ALSPs will region of 20-30%. We explore what this theme means for General New competitive advantage is set to shake up the need to adopt this technology quickly to remain competitive, with Counsel and CXO leaders in our deep dive: the CXO choice. landscape, for the better the potential for new technology-focused entrants to quickly gain Law firm and client relationships are set to be redrawn as ground on established players. 4 The adoption journey Generative AI becomes more widely adopted. Our survey shows Clients anticipate that external partners will play a critical role Despite some hurdles to overcome, Legal should be that corporate legal departments are not currently getting value in Generative AI adoption: almost 7 in 10 are expecting a priority area for Generative AI investment from their external law firms’ use of Generative AI: only 6% of to partner with external firms in the next 6 months to respondents report direct benefits, such as cost savings, The most significant barriers to adoption of Generative AI run GenAI proofs of concept. First movers who partner to from law firm use of the technology. Law firms should identified in our survey are investment related: financial pass on the value Generative AI brings will benefit from real take note: clients have high expectations of increased value if and resource limitations, and challenges with developing a competitive advantage. Generative AI is adopted. 73% of respondents expect to see compelling business case. The role of corporate legal teams in a reduction in cost from law firm use of Generative AI, and advising the business on how to adopt Generative AI safely and 70% expect faster turnaround time. 6 What are the practical next steps for should not be exempt from this and must visibly lead from the strategic repositories, changing processes so that key data front. Whilst they don’t need to be experts, they should have a and documents are captured and stored in the right place legal departments? foundational understanding of the technology and its impact, every time, and data clean up to make data useable. Use proactively sharing learnings. This enables leadership to set a Generative AI to make data clean up exercises easier and 76% of our respondents report no current adoption of the positive vision for the future and more effectively support their more cost effective. technology, but 86% expect targeted or broad adoption teams to navigate through inevitable change. within 2–3 years. To achieve this level of adoption in this + Plan for the longer-term timeframe, legal leaders need to take swift positive action. + Evaluate and prioritize use cases, and then develop/ With Generative AI evolving at such a rapid pace, long-term refresh your digital roadmap planning can feel very challenging. However, it is important We see five key steps that every proactive legal department Legal teams should start to identify and evaluate use cases for leaders to consider the strategic implications of the should take now: with light touch governance. As priority use cases emerge, legal technology on the wider business, the internal workings + Put governance in place for the use of AI departments should dispassionately re-appraise (or develop) their digital roadmaps, assessing the extent to which current of your legal department, and on your external partners. We suggest focusing, initially, on: The General Counsel has a unique dual role in relation to Generative AI; as the guardian of its safe introduction into the technologies have the capabilities to meet their needs. To do – The business: actively seek out and engage with CXOs business, and in using the technology to improve legal service this effectively, legal departments must understand both the and other executive leadership to shape the strategic delivery. The legal department can play a leading role, working organizational technology landscape and what enterprise direction of the business in the age of Generative AI, closely with risk, IT, data, HR and other corporate functions, tooling may be available, as well as the legal technology enabling commercial benefits to be realized safely. to establish governance and risk frameworks for Generative market to understand the art of the possible. Regular market General Counsels should use this understanding to AI’s safe and ethical use. This is an opportunity for the legal scans have never been more important given the step- inform their long-term departmental planning. department to both enable board-level strategic priorities change in capability that Generative AI provides and the for the business and accelerate business growth, whilst also pace of development. Lean on external partners to support – Your legal team: map the skills you have now, the building confidence in adoption within the legal department. these endeavors: they bring vital external perspectives and skills you may need in the future and your plan to train methodologies that will be increasingly important as corporate or recruit to close the gap. Make sensible assumptions + Demystify Generative AI and drive buy-in by putting it functions compete for investment. about the impact of Generative AI on resource mix and into lawyers’ hands numbers, guided by the wider business strategy. It is only through using the technology that lawyers can + Focus on data hygiene and clean up – Your external providers: Engage with law firms and identify how it can best be applied to meet the needs of their Identify the data that is critical for solving your business ALSPs on how they plan to use Generative AI and pass on legal team and wider business. Legal teams should be given problems. This could be data in your document repositories, these benefits. Build this into commercial decision making access to Generative AI tools to experiment and learn through from the wider business, external market data or data from and preferred panel requirements. Consider how your in targeted proofs of concept. Ongoing focus is needed on external law firms. Test Generative AI’s ability to extract source/external mix will change over time and how you training, feedback, best practice sharing and adoption to move insight from this data and identify what data improvements can adjust your mix of external providers to realize GenAI the department forward collectively. Legal leadership teams are needed to leverage Generative AI in the future and benefits earlier. begin actioning them e.g., storing of key documents in 7 Section 1: The rise of Generative AI and its impact on the legal department The launch of ChatGPT in November 2022 organizations within three years.3 Business leaders versatility means that it has application across a range of legal activities, unlocking potential across all areas of legal work. fundamentally changed our relationship with across industries recognize that the technology Fourth, beyond the use of AI by lawyers, the legal department Artificial Intelligence. The application went viral, will have a profound strategic impact, and leading has a critical role in introducing Generative AI into the business becoming the fastest growing consumer application organizations are already responding. C-Suite and ensuring that it is used safely, ethically and responsibly. in history 2 and introducing us to Generative AI, a priorities reflect this, with AI and Cloud Computing This view is supported emphatically by our survey findings. 79% technology with the ability to generate new content ranked as the two technologies that will most likely of our respondents believe that Generative AI is not just hype, that is indistinguishable from—even superior to— drive growth over the next year.4 and will have a moderate to significant long-term impact on how legal work is performed. Further, 49% expect more than just human output. Since that date, Generative AI has What does this mean for legal work? Despite the transformative transformation, with some tasks rendered entirely obsolete by developed at a dizzying pace, extending beyond Generative AI. promise of ‘legal tech’ over recent years, the work of the legal text to generation of detailed images, production of function has remained largely unchanged. Legal technologies entire video sequences and creation of new musical like document management, eBilling and IP management have Figure 1: Overall long-term impact of Generative AI on how typically had narrow application, driving only incremental legal work is performed compositions. Many people now routinely use the efficiency gains. We believe that Generative AI is different, and technology to enhance their daily lives, for tasks that 79% that this technology will drive genuine, sustainable change in say that Generative AI will have moderate range from the administrative (arranging a holiday the way legal services are delivered. We also believe that the to significant long-term impact on how schedule) to the deeply personal (coaching through impact on legal teams will be far reaching, for four key reasons. legal work is performed a difficult personal challenge). First, the legal function often lags other enabling functions in What option best describes your legal department’s perception of digital maturity, which means that Generative AI can deliver the long-term impact of Generative AI on the delivery of legal work? The impact of this transformative technology is significant and measurable improvement. Second, Generative AI surpasses the capabilities of traditional technologies, including Moderate impact Significant impact also being felt in our professional lives. A recent 36% 43% AI technologies, in analyzing documents and unstructured data. Deloitte survey of CEOs found that a great majority This makes the technology well adapted to legal work, with its 79% (79%) expect Generative AI to transform their focus on text-heavy, unstructured data. Third, Generative AI’s Limited impact No or negligible impact 3 State of Generative AI in the Enterprise 2024 | Deloitte US 2 ChatGPT sets record for fastest-growing user base—analyst note | Reuters 4 Generative AI adoption | Deloitte Insights 8 Of course, Generative AI’s impact will not be felt equally across the M&A teams can use dynamic, natural language querying to Generative AI on both legal functional areas (see Figure 2) and a legal department. Some functional areas will be impacted more search huge swathes of documents in multiple languages to range of legal tasks (see Figure 3). than others. Generative AI’s core capabilities (intelligent search/ identify potential risks and upsides. analysis, summarization, and content generation) are very effective According to our survey respondents, the ‘contracts and Legal and regulatory compliance teams can rapidly compare commercial’ functional area is expected by some margin to be when applied to large volumes of text-heavy documents: content with emerging or existing rules and regulations, most transformed, followed by ‘legal operations’, ‘corporate Generative AI can help contract and commercial teams to identifying areas of potential non-compliance with tailored transactions & M&A’, and ‘regulatory & compliance’. Our work with review and summarize contracts, assess contract risks, suggestions for remediation. clients supports this; in contracting and M&A, in particular, we are compare contracts to playbooks, and suggest how to amend Our survey results reflect the opportunity we see in these areas. seeing some truly impressive results (see Case study 2). contracts to bring them into compliance with the playbook. We asked participants for their view on the likely impact of Figure 2: Impact of Generative AI on practice or functional area of the legal department. Figure 3: Impact of Generative AI on specific legal department tasks or activities Document review Contracts & Commercial 73% 19% 78% 19% and summary Legal Operations 57% 30% Data extraction, 75% 22% management and analysis Corporate Transactions & M&A 37% 31% Translation 65% 24% Regulatory & Compliance 29% 43% Corporate Governance & Document drafting 47% 39% 29% 31% Company Secretariat Knowledge searching General Legal Advisory 24% 49% 44% 46% and retrieval Data Protection & Privacy 24% 51% Legal research 42% 42% Disputes & Litigation 26% 40% Handling business queries 37% 37% Intellectual Property 17% 40% Monitoring compliance with 36% 33% Antitrust & Competition 11% 23% regulations/laws Human Resources Decision support and 9% 37% 19% 24% & Employment strategic advising Significant impact Moderate impact Limited impact No or negligible impact Don't know 9 The task-based view is consistent with the view by functional area. For example, many activities that underpin the ‘contracts Case study 1: Generation of service schedules within contracts and commercial’ functional area also score highly in our task- We ran a pilot with a telecommunications multinational, leveraging their own instance of Generative AI to apply its capabilities in based view: ‘document review and summary’, ‘data extraction, the creation of contract service schedules. management and analysis’, ‘translation’, and ‘document drafting’, Our own lawyers and contract management experts worked in close collaboration with the client to develop and fine-tune a set of to give some examples. This trend extends to other areas. In legal prompts to generate first drafts of the service schedules, replacing what had been a manual, labor-intensive process. operations, tasks like ‘data extraction, management and analysis’ score highly, which we see as particularly powerful in enabling Our findings show that by leveraging Generative AI to support with the service schedule generation, we were able to reduce the more effective knowledge management. The same is true for cost of production per contract by up to 60% as compared to a traditional, manual approach. document review and data extraction in the context of M&A. An example of a high impact capability that intersects multiple tasks is translation, which, when combined with other Generative Case study 2: M&A due diligence AI capabilities (such as comparing contracts in several languages We have compared Generative AI enabled legal due diligence to both traditional AI-enabled due diligence (which uses machine against a regulatory standard, or searching and analyzing learning assisted contract review solutions) and manual review. precedents across a multi-language knowledge repository), We have found an additional 25% efficiency saving by using Generative AI compared to traditional AI enabled due goes beyond incremental gains and delivers a step change in diligence, and a 75% efficiency saving compared to manual review. efficiency and quality. Effective translation is a powerful asset for global organizations, allowing for better collaboration and Generative AI has brought additional benefits—both qualitative and quantitative. For example: efficiency across linguistic and cultural boundaries. By using the – Generative AI’s language translation abilities has improved the speed and quality of reviewing multi-language document sets. translation capabilities of Generative AI, organizations will not be – Summarization capabilities go beyond traditional extraction of clauses, enabling more succinct summaries of content across tied to recruiting in particular jurisdictions for access to language documents (including complex contractual provisions involving mathematical formulae). capabilities, resulting in a workforce that is more flexible and agile. – Contextual understanding and ‘comprehension’ of documents has enabled nuanced, risk-based scoring, saving lawyers’ It is interesting to note that our respondents generally assess review time. the impact of Generative AI on individual tasks (such as document review) to be higher than the impact on functional areas (such as corporate transactions and M&A). This may be explained by Generative AI’s relative immaturity, and suggests that most corporate legal departments are currently experimenting at the individual task level. This is a sensible starting point; the best way to learn about Generative AI is to apply it to individual everyday tasks, building familiarity and confidence with the technology. However, there is a bigger prize to be won for legal departments that take a more strategic approach. By applying Generative AI to an entire legal or 10 business process, using it to improve each individual step in Interconnecting multiple use cases can combine in a ‘string of pearls’ to collectively enhance end-to-end business processes the chain, the benefits of the technology can be compounded. Imagine a ‘string of pearls’, with each step in the process an Organize, manage and Analyze regulatory individual pearl, linked together to create a chain of value. Analyze market data and analyze large amounts of requirements to identify industry trends to identify data and identify potential potential issues or risks related Legal teams should look to identify ‘strings of pearls’ for their own potential targets risks and opportunities to the transaction organizations, with areas like contracting, M&A, litigation and regulatory compliance holding particular promise. For General Counsel, the impact of Generative AI extends beyond Strategic using the technology to improve legal and business processes. Regulatory planning Due approval and Legal must play a critical role in introducing the technology safely and market diligence closing into the business, with legal teams leading the way in establishing research the right guardrails and governance for Generative AI alongside the CIO, CTO and others. By establishing the right safeguards, GCs can actually accelerate time to value by encouraging safe and responsible experimentation and usage of GenAI. In playing this role, GCs are faced with the dual challenge of Preliminary Negotiations Integration due diligence recruitment (do the lawyers in the legal department have the required specialist legal expertise?) and mindset (is the department equipped to navigate the tension between accelerating and protecting the business?). Generate specifically Analyze previous SPAs to Analyze organizational To address this, legal teams will likely require new technical legal tailored due diligence identify potential issues data to identify skills with data privacy, intellectual property, regulatory and checklists and reports or risks associated with integration-related risks technology contracting skills particularly important. It will also specific clauses Analyze policies for require a new, enabling mindset across the legal department, Drafting and compliance-related as the teams will be under pressure to move quickly to support negotiation support remediation Generative AI use. Upskilling and training in these areas will be critical, as will partnering with external legal providers with deep expertise and practical experience of applying technical legal skills Value is compounded as outputs from one use case act as inputs into the next. in practice. When one stage improves, the next stage benefits. 11 Section 2: The benefits of Generative AI for legal departments We believe that Generative AI can create Figure 4: what benefits do you expect Generative AI to deliver? significant and sustained value for corporate legal departments, and that those who embrace change head on will more rapidly reap greater rewards. Efficiency Experience Capability Improve operational performance Provide fit-for-purpose and Develop and enhance digital We see three core ways in which Generative AI will through improved productivity by customized experiences to lawyers, and data capabilities that are bring value to the corporate legal function: through doing more with less and the business enabled by GenAI enhancing efficiency, experience and capability. We outline these in Figure 4, along with our Gains in productivity and 88% Improved business self- 62% Improvements in quality of 60% efficiency service and enablement legal work respondents’ assessment of their likely impact. Acceleration in the performance Improved ability of the business Enhanced quality and of legal tasks and activities to self-serve and handle consistency in the outputs from + Efficiency defined activities with reduced legal intervention legal work and advice ‘Efficiency’ relates to Generative AI’s ability to accelerate outputs with reduced effort. Various studies show that workers who Improvements in employee 60% Improved, data-led 45% leverage Generative AI perform their roles more efficiently than satisfaction decision making those that do not.5 Small efficiency gains can add up to significant Replacing low value, repetitive Improved access to data to legal activities and tasks with drive greater data and insight wins: over the next decade productivity gains resulting from Reduced costs associated 67% more varied/interesting/ driven decision making Generative AI are expected to increase global GDP by $7 trillion.6 with performing legal work challenging work Reduced costs associated with Legal work will similarly enjoy efficiency gains. Internal Deloitte performing legal tasks and studies anticipate that Generative AI can help lawyers become activities Enhanced service from 40% More proactive management 43% law firms and legal service of risk 25% more efficient on average.7 Depending on the nature of the providers Using insight from data task, this percentage can increase significantly. For high volume Improved quality, speed and/or proactively to prevent or cost of Generative AI enabled minimize legal risk 5 Navigating the Jagged Technological Frontier: Field Experimental Evidence of services the Effects of AI on Knowledge Worker Productivity and Quality 6 Generative AI | Deloitte UK 7 Internal Deloitte study on Generative AI’s impact on legal work 12 and labor-intensive activities we can expect to see efficiency gains of up to five times; even more in some cases. Our survey Case study 3: Efficiency—eDiscovery and investigations respondents overwhelmingly agree that efficiency is likely to be We have worked with multiple clients across industries using our proprietary software, ‘NavigAite’, a Generative AI solution built on the key short-term benefit associated with Generative AI. Our Relativity, one of the leading eDiscovery and legal search software platforms. work with clients also supports this view (see Case study 3). With extensive backend engineering to create guardrails to minimize ‘hallucination’ risk (i.e., the risk that the model simply makes As adoption increases, we expect to see more of these examples. up an answer to a question, without any factual basis) and promote transparency (showing the reasoning behind the output Broader deployment of Generative AI will drive efficiency gains generated), NavigAite uses Generative AI to accelerate outputs across a range of legal activities including but not limited to across a wide range of legal activities: document summarization document review and summaries, first pass reviews, privilege reviews, translation and personal information identifier reviews with and review, document drafting, legal research, advice memos, automated redaction. project management, and due diligence, amongst others. Any By leveraging NavigAite, we have achieved some truly impressive results that have improved quality (in turn reducing the need for calculation of efficiency gains must, of course, factor in the time manual review), and slashed both time and resource demands to complete historically highly manual, labor-intensive tasks. required for human validation of outputs (at this early stage in the The table below outlines indicative review metrics, staff, and timing across methods: development of the technology, human review remains a critical step in the process). Even taking this into account, we anticipate Please note, the different review approaches require different numbers of documents to be manually reviewed. Adding Generative AI and that greater efficiency will unlock significant levels of capacity using appropriate industry—standard statistical sampling techniques increases the speed of getting to the key information, and reduces the across the legal team, presenting General Counsel and the C-Suite need to look at false-positive documents. with strategic choices. See separate section—Deep dive: the CXO choice, p.16. Approach Documents Metrics Staff Timing Linear 1M reviewed, 1M “Eyes on” 106 4 months + Experience Technology assisted review 5M 100k reviewed, 100k “Eyes on” 25 2 months ‘Experience’ relates to Generative AI’s ability to deliver qualitative Generative AI enabled review 5M reviewed, 5k “Eyes on” quality control only 5 1 month benefits: typically, an improved or more tailored experience for both lawyers and their business counterparts. Further, we expect Generative AI to streamline and improve the is required. In addition, human-like conversation will improve the Perhaps the most far-reaching qualitative benefit will be felt by legal department’s interactions with the business. Whilst lawyers customer experience and encourage users to engage—rather the lawyers themselves. As we described above, Generative AI may instinctively feel uncomfortable using Generative AI as a first than circumvent—the legal department. We expect this to improve has the potential to relieve lawyers from the more repetitive contact point for business queries, the reality is that it represents overall risk management. elements of their role and allow them to focus on more fulfilling, an improvement on the position today. Through Generative AI strategic work. In turn, this should improve legal team morale and The ability to personalize responses to the business and enabled chat bots and legal front doors, business users can engagement scores. Our survey corroborates this view, with 60% streamline the customer experience elevates Generative AI to an ask plain language questions and receive quick, tailored, and of respondents expecting improvements in employee satisfaction effective first point of contact, fielding routine queries and further contextualized responses. Only the most relevant knowledge is and engagement. freeing up lawyer time. Our survey respondents agree, with 60% surfaced, with suggested outreach to legal teams if more detail expecting improved benefits from self-service. 13 + Capability Case study 4: Efficiency—Regulatory compliance review The last of the three ways in which Generative AI can deliver value Our proof of concept used Generative AI to compare a technology company’s customer Terms & Conditions (T&Cs) with a new is by enhancing capability. This refers to Generative AI’s ability to piece of regulation to identify any changes required to bring the T&Cs into compliance. improve legal service delivery, either by elevating the quality of legal work or by enabling the legal department to offer entirely The analysis produced a table outlining the regulatory requirement, compliance status of the T&Cs, explanation for the assessment new services and insights to the business. including source reference points and recommended changes to bring the T&Cs into full compliance. Our approach required 3 iterations of prompting between our Deloitte Legal regulatory lawyers and Generative AI engineers to get Whilst efficiency and improved lawyer experience can be quick to a good first draft of the analysis (rated 4 out of 5 by our lawyers for accuracy and completeness). wins, augmenting the capabilities of the legal department with Generative AI requires more strategic thought. The key to We saw a 50% reduction in both the trainee and experienced lawyer time required to complete this analysis. enhancing capability lies in capturing and benefiting from the data that flows through the legal department. Much of that data is unstructured, housed in legal documents and contracts, and Case study 5: Customer experience—Contractual simplification and alignment difficult to query and analyze using more traditional approaches. Consequently, legal departments have tended to be less data- Our client had a significant number of different versions of customer facing Terms & Conditions (T&Cs), creating complexity in their driven than other enabling functions, with decisions based on customer relationships and exposing the organization to risk. institutional knowledge and professional experience—or simply In what would have historically been a labor-intensive manual process, our client used Generative AI to compare the T&Cs and on gut instinct. In our work with clients, we have found that identify key differences. Generative AI can support more data-driven decision making The client then used Generative AI to extract the best examples of compliant provisions from each version of the T&Cs to create a across legal teams. The technology makes combining and master, ‘best practice’ version. analyzing data from disparate sources easier to do, so that data from legal operations, from the wider business, external market The client reported that using a traditional manual approach, the exercise would have taken several weeks. Using Generative AI, data or data from external law firms can be brought together to the task was reduced to less than 4 days (2 days of prompt engineering and refinement, and 2 further days of human analysis and surface new insights. This, in turn, can help legal departments review). The client estimated that this represented a five-fold acceleration. to identify the root cause of legal and business issues, and work This project resulted in a simpler, more accessible set of T&Cs for the end customer, and also improved the experience for the proactively to prevent issues from occurring—or, at the every legal team on this project by focusing them on more strategic review and decision-making activities. least, to minimize their impact. To share some examples, we have seen Generative AI being used to analyze litigation and investigation outcomes, surfacing patterns and allowing for root cause analysis of the reasons that gave rise to the claim or investigation. We have also seen the technology used to prevent unnecessary contract spend (a common problem with very large portfolios of contracts): Generative AI is used to extract data from documents in the contract repository and combine it with invoice and spend data to make sure that 14 erroneous payments are not made. Generative AI can also who has access to it. Now is a good time for legal department to combine internal institutional knowledge (about a customer, a start to address these issues. market, or a competitor) with external data sources, helping legal departments to gain a richer understanding of a particular issue 60% of our survey respondents agree that Generative AI and shape a strategic response. should improve the quality of legal work, whilst 45% and 43% expect improved data-led decision making and more proactive This is not to suggest that this technology is the silver bullet management of risk, respectively. These percentages represent a to resolve all data challenges. To extract the most value using slightly lower level of confidence than in other survey areas. This Generative AI, it should be applied to data that is centralized, supports our view that enhancement of capability is a benefit that cleansed and appropriately structured. Many of our clients tell us will take longer to emerge, requiring the technology to develop that data management is an issue for the legal department—many and legal functions’ approach to mature before the benefits can do not know where their data is located, how it is being stored or be fully realized. Case study 6: Capability—IP value leakage prevention In a large-scale carve out transaction, the data separation and transfer process required our client to review the data for transferring employees whose laptops had not been wiped or refreshed when the transfer was made. This exercise required us to identify data requiring segregation, in order to mitigate the risk of Intellectual Property (IP) leakage and misuse of information, which could result in reputational damage or breach of separation agreements. Using Generative AI augmented document review, we were able to review 1.2m documents in 3 weeks, across a variety of document types and languages. This would not have been feasible in the client’s timescales without using the technology. This approach increased speed, reduced error and mitigated the risk of IP leakage by rapidly and accurately identifying documents containing critical data. 15 Deep dive: the CXO choice strategic activities will find themselves reaping dividends. Legal teams will be able to provide closer business partnering, supporting activities that drive top-line growth (IP strategy, new product As teams deliver more with less, CXOs face a strategic choice as to how to reinvest a legal team’s and service launches, strategic alliances) as well as those that avoid cost (improved regulatory unlocked capacity. compliance, improved contract portfolio management, and others). On the one hand, time saved in performing legal work can be reinvested into more strategic On the other hand, increased productivity can enable cost reduction. This is a view shared by two activities. Many legal teams describe themselves as capacity constrained. They tell us that they thirds of our survey respondents, who expect the cost of performing legal work to reduce as a have limited time for strategic activities because they are constantly ‘fire-fighting’. This can create consequence of Generative AI adoption. From the work we have done with our clients, we expect a reactive culture which contributes to the perception of the legal function as a cost centre. Generative AI to enable a reduction in the overall cost base of a legal team of up to 30%. We However, those organizations that choose to invest a legal team’s unlocked capacity back into provide an illustrative view of the cost savings and associated levers in Figure 5. Figure 5: Potential cost savings across legal departments using Generative AI Business self-service In-house resource Changed sourcing External spend Technology and enablement productivity and mix reduction automation investment optimisation through use of GenAI 100% 3–5% Legal virtual assistant Legal Front Door build out + 10–15% ~20-30% GenAI cost saving opportunity Productivity gains 5–10% Reduction in paralegal/junior headcount/admin 3–5% 3–5% Insource work currently Simplify spans and layers delivered externally that can Push to use GenAI in service Additional investment now be delivered more provision/pass on benefits required in technology, e.g., efficiently with GenAI, e.g., Pilot with alternative GenAI platforms, foundational low-mid volume eDiscovery/ providers to gather data tech like document document review points on services with GenAI management and legal front Outsource work currently door vs. traditional services delivered in house that could Incorporate into RFP Offset by tech rationalisation, be more efficiently delivered requirements and evaluation e.g., removal of expensive AI on an outsourced basis criteria modules for CLM, simplified leveraging GenAI Illustrative Build data points/case document management baseline spend studies/industry benchmarks solutions Key Savings (lower end) Savings (upper end) Cost increase (lower end) Cost increase (upper end) 16 Deep dive: the CXO choice (cont.) We explore the cost savings levers in more detail: + Improved business self-service through sophisticated and hyper-personalized human-like chatbots. A conservative estimate suggests that up to 5% of a legal team’s effort can be saved by generating FAQs, and surfacing legal knowledge from knowledge bases at the point of need, for the business to self-serve in an intuitive and highly personalized way. + Increased productivity and better use of in-house resources (already discussed in this section at length, and highlighted through our case studies). + Changes to the sourcing mix (i.e., who performs what work) both internally and externally. Generative AI should give corporate legal departments more flexibility in their resourcing models. Work that had previously been outsourced to law firms or alternative legal service providers (because of capability or capacity constraints) can be brought back in-house as Generative AI frees up capacity and enhances capability. The inverse may also be true; the technology should enable legal departments to outsource more activities to alternative legal service providers and managed service providers, who will become more cost competitive and deliver greater service innovation by using Generative AI. The choice that the legal department makes will depend on its organizational and economic model; either way, using Generative AI offers more flexibility. + Reduction in spend on external law firms, as a result of law firms’ ability to deliver legal services more efficiently. See Section 5: External law firms and legal service providers, p.23. The most likely outcome is that legal departments will combine these two approaches, with legal teams using Generative AI as an enabler to meet future growth in business demand whilst maintaining a similarly sized legal team and cost base. Regardless of how efficiency gains are realised, the good news for legal departments that embrace Generative AI is that it will enable them to do more with less—and deliver greater value to the business. 17 Section 3: The adoption journey Generative AI’s full potential will only be realized if it is successfully integrated into Figure 7: Current Generative AI adoption levels by practice area and activity type legal workflows and adopted across the legal department. Adoption by practice area Contracts and commercial 24% 73% If our respondents’ predictions about the speed and scale of impact are correct, Legal operations 27% 55% we can expect to see rapid adoption of Generative AI over the next 12 months. Regulatory and compliance 9% 52% Our survey supports this. The vast majority (87%) of respondents expect some Data protection and privacy 9% 48% adoption within 2–3 years, with just under 4 in 10 anticipating that the technology Corporate transactions and M&A 6% 45% will be broadly adopted in the same time frame (see Figure 6). This is ambitious: General legal advisory 12% 45% adoption at this scale would represent a significant increase in uptake when Corporate governance and company secretariat 9% 42% compared to existing technologies used within the legal department. Intellectual property 3% 39% Disputes and litigation 6% 36% Figure 6: Current and planned adoption levels of Generative AI Antitrust and competition 3% 27% Human resources and employment 6% 24% Today 76% 87% expecting at least targeted Adoption by activity type Next 2–3 years no adoption currently adoption in 2–3 years Document review and summary 23% 74% Document drafting 23% 67% 56% 49% Data extraction, management and analysis 15% 64% 38% Monitoring compliance with regulations/laws 10% 54% 24% Legal research 23% 51% 15% Knowledge searching and retrieval 23% 51% 5% 8% 3% 3% Translation 36% 41% Not started/ Cautiously Evaluating Targeted Broad Handling business queries 21% 38% no plans observant impact adoption adoption Decision support and strategic advising 8% 33% Today Next 2–3 years 18 Legal teams have a significant gap to close if they are to meet use relatively easy to identify. Scores for all other functional areas Beyond financial challenges, our respondents also highlight these adoption ambitions. Currently, 76% of respondents report are low, with only one other area exceeding 10%. The task-based a significant ‘people’ hurdle to developing the right skills and no adoption of Generative AI, indicating that they are at the “wait view aligns with this—activities associated with contracts work capabilities, as well as lack of data readiness, and legal and and see” or early impact evaluation stages. Positively, however, the score highly. Translation, not strictly a ‘legal’ task—scores highest, regulatory concerns. We share more in the following sections on majority of our respondents are either currently undertaking or perhaps indicative of a reticence to use Generative AI for legal how legal teams can overcome these challenges. planning to undertake in the next 6 months readiness activities. tasks, and a bias towards starting with ‘safer’, non-legal work. This includes over two thirds undertaking internal initiatives Legal teams can address this head on through foundational Figure 9: Barriers to Generative AI adoption around training and awareness, as well as external initiatives education programs that give confidence to lawyers on how best 57% Financial/resource limitations to partner with providers on targeted use cases. 66% plan to to use Generative AI’s capabilities. additionally give legal teams access to an LLM, in support of both (see Figure 8). Both the functional and task-based views also demonstrate 51% Business case challenges the significant gap that legal teams will need to close. Even Consistent with our use of functional and task-based views in areas of higher adoption, levels are expected to double (in to assess the impact of Generative AI, we asked our survey the case of legal operations) or treble (in the case of contracts 51% Skillset barriers and driving adoption respondents in what practice areas/functions and on what tasks and commercial). In areas with low adoption, our respondents they were using Generative AI. From a functional perspective, are expecting five, sometimes six-fold adoption increases in 43% Data availability and quality concerns ‘legal operations’ leads with 27% adoption (perhaps unsurprising, 2–3 years. This presents a significant and pressing challenge, given that it is often the team tasked with exploring new and legal teams will need to overcome a number of barriers 38% Legal and regulatory concerns technologies). Second is ‘Contracts and commercial’ with 24%, for Generative AI to buck the historic trend of slow technology again in line with expectations given this area’s high potential for adoption within legal departments. automation, which makes opportunities for practical everyday 35% Technical and data challenges Investment-related barriers ranked highest for our respondents Figure 8: GenAI readiness activities currently being (see Figure 9). These include financial and resource limitations, and undertaken/planned for the next 6 months challenges with developing business cases. A slim majority of our Figure 10: The source of budget for Generative AI initiatives respondents report no budget having yet been made available 68% for Generative AI initiatives, restricting the ability of legal teams 60% Training and awareness 68% to meet their objectives and timelines. Whilst there is confidence on GenAI that budgets will soon begin to materialize, those legal teams 40% 33% Pilot with external providers on that are able to make a compelling business case that clearly 25% 68% targeted use cases outlines the investment ask, projected benefits (both quantitative and qualitative), and the anticipated return on investment will be Individual employees exploring 3% 66% most successful in securing investment. This is not an easy task, public LLMs, e.g., ChatGPT particularly as the technology is still untested. Legal departments Not funded/no Funded by legal Funded by broader Data cleansing and should consider partnering with external providers, who have budget made department budget organization 21% available budget aggregation exercise experience of working with clients in their industry, to provide context and benchmarks to help with business case development. Today Next 2–3 years 19 Section 4: The legal department’s workforce We have already touched on the impact of Underpinning our predicted legal ecosystem changes are four Figure 11: Expected changes to the legal ecosystem over time key trends: Generative AI on the overall size of the legal function. Generative AI will also impact who delivers the legal 1. Increased self-service and automation. Demand on department’s work, and the range of skills required As we set out in Section 2, we expect Generative AI to deliver legal significant productivity benefits and to improve the lawyer to support that delivery. Currently, legal work is and customer experience. This will lead to an increase in self- delivered by a combination of the legal department service and automation. As adoption increases, we expect itself (lawyers and allied professionals), external legal teams to be able to meet increasing business demand without an increase in headcount. 30% law firms, alternative legal service providers, the 40% 35% business (through self-service) and automation. 2. More strategic deployment of in-house lawyers driving reduced reliance on external counsel. 10% 10% Figure 11 presents our view of how work is delivered Greater automation and business self-service should unlock 5% today, and how we expect this to evolve over time as lawyer capacity, meaning that work currently sent to external 40% Generative AI adoption increases. counsel because of capacity constraints can be serviced 50% 45% in-house. In addition, Generative AI should augment the capabilities of in-house lawyers, enabling them to do more Who delivers legal services: compositional strategic and challenging work within the department, rather 10% 20% 5% changes than sending it to specialist law firms. Of course, reduced Typical legal Targeted GenAI Broad GenAI reliance on external counsel means a reduction in overall ecosystem today adoption adoption Figure 11 shows that as business demand increases, so will self- legal spend. service and automation, enabling legal teams meet increased Self-service ALSP/shared services demand without increasing lawyer headcount. We also expect 3. Short to medium term increase in the use of shared the roles and skills in the legal department to change, with more service capabilities/ALSPs. Automation Lawyers operational and technology-focused staff. Currently we believe that ‘alternative’ legal delivery models are External legal