How to Transform Your Legal Practice with AI and Automation: Lessons from A&O Shearman and Clifford Chance
16 September 2025
Table of Contents
AI is Reshaping Knowledge-Intensive Workflows
Since the public release of ChatGPT 3 in late 2022, the implications of AI for complex, knowledge-intensive tasks have become increasingly clear. As firms continue to invest in AI implementation, we can expect continual improvements in the quality of AI-augmented work, and corresponding improvements in worker productivity. Academics at the Harvard Business School have demonstrated that with current capabilities, augmenting workflows with AI can accelerate knowledge-intensive work by as much as 25%.
AI implementation has been as much a bottom-up as a top-down process. Professionals in a variety of sectors have found new ways to embed AI into their work, and the list of tasks to which AI has successfully been applied grows longer each day. Despite the impressive capabilities of new models, some tasks remain best suited for human reasoning. These limitations, combined with the importance of professional oversight for remaining within ethical and regulatory boundaries, mean that for the foreseeable future AI tools will be implemented most effectively as part of a hybrid workflow - a combination of AI and human reasoning.
Much of the work undertaken by legal professionals is suited to these kinds of hybrid workflows. Already, an increasing number of lawyers have found that implementing AI into their practice has allowed them to spend more time on the judgement-based, expertise-driven work which matters to them the most, while delivering greater value to their clients through the automation of repetitive tasks.
AI in the Workplace: Opportunities for Legal Professionals
Legal professionals are increasingly conscious of the transformative potential of AI - though remain highly optimistic about the impact it might have on their own practices. In recent surveys conducted by Thomson Reuters, 77% of respondents in the legal sector suggested they thought AI would have a transformational impact on their work over the next five years, and 72% of them believed it would ultimately be a force for good in the profession.
But many legal professionals, understandably, remain uncertain about the potential risks associated with AI implementation. Given the sensitive nature of the profession, it is particularly important that lawyers can trust the output of AI tools, which means that they must be trained on authoritative legal data and designed with the input of lawyers and engineers.
The Key Components of a Legal AI Strategy
Professionals generally understand the need to engage with developments in AI technology, and have a sense of the potential for AI to save time and improve the quality of work. But the rapid pace of technological change - and the proliferation of specialised tools and products - has made it increasingly difficult to stay on top of. Data from across sectors suggests that the biggest barrier to AI adoption is a lack of relevant information, with as many as 60% of respondents stating as such.
It is therefore vital that firms invest time and resources into equipping their lawyers with the confidence and know-how to get the most from AI technology. There is a clear return on investment for investing in an AI transition; according to one study, organisations with an AI adoption strategy were around twice as likely to achieve revenue growth in a given year.
So what should an AI adoption strategy look like?
Firstly, it means identifying key use-cases for AI in your practice. AI is suitable for many, but not all tasks, and it is important to separate those where it can help from those where it might only hinder. The list of potential use-cases for legal work is extensive, but some of the most popular include:
- Handling large volumes of legal data - precedents, case law, contracts
- Improving client response times and client intake
- Reducing human error through automated proofreading
- Document review, contract analysis, and automated document generation
- Legal research and case law discovery
- Compliance monitoring and risk assessment
Secondly, it means selecting from a wide range of specialised tools and products which have emerged to meet the demand for reputable legal technology. There are a variety of well-regarded offerings, each with their own advantages and limitations: Westlaw’s CoCounsel offers database integration, Harvey is acclaimed for its intuitive interface, while Spellbook can be integrated directly into Word for easy access. These tools, and many others like them, offer lawyers the potential to accelerate case research, document drafting and communications.
General-purpose models, like ChatGPT and Claude, can also work well for general applications like summarising information for briefs and drafting communications with clients and colleagues, which save time for the more critical legal work. Of course, these models are not trained on authoritative legal data, which limits their utility for sensitive legal work.
Customer surveys and preliminary usage data suggests that AI tools can more than double the rate of document review and contract drafting for individual lawyers. AI tools can also streamline the non-legal functions of law firms: marketing automations, customer service solutions, document management and so on. Saving time on busywork frees up more time for billable work, for working on one’s mental health and wellbeing, or for innovation and strategic planning. Ultimately, effective AI implementation gives lawyers the space to focus on what they think is most valuable.
Three Example Automations to Enhance Legal Practice: Inputs, Prompts and Outputs
Now that we have outlined what an AI adoption strategy requires, it is worth working through some specific examples of how AI and automation can be used to accelerate aspects of legal practice. These are only a few examples, but the aim is to provide a sense of how you might identify use cases for AI in your own practice.
Automations can take various forms depending on the kinds of AI tools you employ, but are built from the same common features. The process of designing a hybrid workflow might look something like this:
- Identify a key use-case (e.g. Contract Review)
- Assign an appropriate tool (e.g. ContractMatrix)
- Provide relevant inputs (e.g. Contracts in PDF form)
- Follow a set process. (e.g. preset prompts, tuned AI agents)
- Set expectations for the output. (e.g. validating data, formatting, routing)
Following this process, we could design the following workflows for a small law firm looking to automate both legal and non-legal sides of the business:
Marketing Automation in Law
This AI Automation helps law firms streamline their marketing efforts by auto-generating personalized emails and brochures based on website lead inquiries, delivering tailored client communication that saves one hour of manual administrative work per lead.
Marketing Automation
Google Gemini + Workspace
Website lead
Auto-generate personalised email based on user input or inquiry
Tailored client email and brochure
1 hour saved per lead
Contract Review Automation in Law
This AI Automation helps legal professionals streamline contract reviews by analyzing uploaded documents to flag inconsistencies and errors, delivering a meticulously proof-read contract while saving two hours of manual effort during every legal drafting process.
Contract Review
ContractMatrix
Uploaded contract
Analyse clauses, flag inconsistencies or errors
Proof-read contract
2 hours saved per contract
Case Research Automation in Law
This AI Automation helps legal professionals streamline case research by automatically retrieving relevant statutes and key precedents from legal databases, delivering a comprehensive list of authorities that saves one hour of manual work per contract.
Case Research
CoCounsel
Legal Database
Retrieve statutes, case law, and key precedents
List of potential authorities
1 hour saved per contract
Hybrid workflows like these pose new opportunities for lawyers working in any context to save time on key legal and non-legal tasks. Ensuring that there is no compromise on the quality of the output requires further thinking, however - as AI implementation is not without its limitations.
AI Governance: Managing the Risks of AI Implementation in the Law
An effective AI adoption strategy requires the development of frameworks to mitigate the various risks associated with AI, some of which are particularly pertinent for lawyers. While AI products offer compelling opportunities for legal practice, they also bear certain risks, which have sparked strong reactions from firms concerned about client data security. The UK Court of Appeal is currently dealing with Ayinde v. London Borough of Haringey - concerning the use of unverified outputs from generative AI in litigation. Lawyers who fail to validate the citations made by AI models - or who misrepresent the output of AI as human-written content - could face sanctions in future.
Legal professional privilege makes it difficult for AI services to access high-quality, proprietary legal data for training models within existing ethical and regulatory frameworks. Regulations have begun to emerge across different jurisdictions, but the approaches of legislators have been highly varied and remain largely untested. For now, it is up to firms themselves to ensure that AI implementation falls within acceptable risk parameters.
Large language models (LLMs) are known to occasionally produce ‘hallucinations’ - inaccurate or misleading claims which threaten to undermine the quality and trustworthiness of the response. For lawyers, who depend on their reputations for providing reliable legal advice, such mistakes could come at a serious professional cost. It is often difficult to explain why LLMs produce the responses they do, which poses challenges for accountability and replicability. So long as AI technology remains fallible, there will remain a need for human oversight when it comes to key use-cases such as document drafting and analysis.
These challenges make proper AI implementation all the more important. With the right frameworks in place for vetting the outputs of LLMs, lawyers can work more efficiently without compromising on reliability. New and improved AI models offer clearer insights into the reasoning behind their decisions, and can meet the high standards of accuracy which legal professionals expect. As AI technology continues to mature and more robust procedural checks are implemented, these risks will continue to diminish.
AI in Law Firms: Case Studies from A&O Shearman and Clifford Chance
Let’s now consider how two leading global law firms are integrating AI into their legal practice with substantial benefits. A&O Shearman and Clifford Chance have both utilised external AI services as well as developing AI products in-house.
A&O Shearman: How Big Law is Scaling Generative AI
A&O Shearman was the first firm in the world to deploy generative AI at enterprise level - which speaks to the firm’s long legacy of innovation in legal technology. A&O Shearman had been trialling Harvey - a legal AI based on ChatGPT 4 - since November 2022. In 2023, the firm announced an exclusive partnership with the team behind Harvey, many of whom are themselves former lawyers, working alongside engineers and entrepreneurs who share their vision for transforming the legal industry.
After the successful adoption of Harvey into its global practice, the firm expanded its AI toolkit with ContractMatrix, a proprietary tool for contract drafting and negotiation, built in collaboration with Microsoft - which has the potential to save in-house legal teams hundreds of hours in contract review.
- Building Confidence in Innovation. A&O Shearman assembled an in-house team of lawyers, engineers and entrepreneurs to create the Markets Innovation Group, responsible for searching for innovative ways to improve legal practice. In 2022, the group conducted a large-scale trial of Harvey with 3,500 of A&O Shearman’s lawyers, which provided invaluable evidence for further decisions about AI integration.
- Retaining High Standards. A&O Shearman’s lawyers use Harvey to generate insights and predictions based on ‘benches’ - large volumes of high-quality legal data. These outputs are then reviewed by human lawyers to deliver a result that is accurate and accountable.
- Developing Hybrid Workflows. ContractMatrix was built from the ground-up to be accessible. That’s why the tool was integrated directly into Microsoft Word for easy access to AI assistance. A&O Shearman has explored further integrations with the Microsoft environment to enable lawyers to incorporate AI tools in their own, tried-and-tested workflows.
As a result of successful AI implementation, lawyers at A&O Shearman have reported significant time savings - while maintaining a high standard for accuracy. Over 1,000 of the firm’s lawyers now use ContractMatrix, which saves each one about seven hours per contract review, on average - which marks a productivity gain of about 30%.
“No one went to law school for manual process exercises – this is about enabling lawyers to focus on tasks that require strategic thinking and decision-making,”
- David Wakeling, Head of A&O Shearman’s Markets Innovation Group
This success has not been unwarranted. A&O Shearman prides itself on being a pioneer in the legal industry, and has invested considerable resources into finding innovative ways to improve existing legal processes, for which it was named the world’s most innovative law firm in the Financial Times Innovative Lawyer Global Summit, 2025.
Through effective partnerships and successful implementation, A&O Shearman has demonstrated the value of investing into an AI adoption strategy.
Clifford Chance: A Comprehensive Strategy for Legal Innovation
Clifford Chance has adopted machine-learning technologies in the past for a variety of cases. In one instance, a team of employment litigation lawyers were faced with a short deadline to review over a thousand employment contracts, written in several languages. Working together with the Innovation team, the team developed a machine-learning tool to assist in analysing the contracts using keywords, which cut the workload down by around 55%.
- Utilising AI Safely and Ethically. Clifford Chance developed an ethical framework for using AI responsibly, the AI Principles and Policy framework, which ensures that clients consent to the use of AI solutions for critical work. All AI-generated outputs are tagged and validated by qualified lawyers in the relevant jurisdiction, which mitigates concerns about hallucinations.
- Piloting Further Innovation. Clifford Chance is investing in future technologies with new partnerships. In March, 2025, the firm announced that it would extend its partnership with Wexler, the legal technology innovator. Wexler’s specialised legal tool can process up to 500,000 documents at a time, uncovering previously unseen connections.
With the power of generative AI, the Innovation team at Clifford Chance has pursued further productivity gains for both legal and non-legal functions. In early 2024, Clifford Chance announced that it would roll out Copilot for Microsoft 365 across its entire workforce, using AI and natural language processing to enhance productivity by automating a variety of daily tasks - drafting emails, managing inboxes, and scheduling meetings. For legal work, the firm built its own private AI tool, Clifford Chance Assist, developed on Microsoft's Azure OpenAI platform. After extensive trials, which involved around 1,800 users, the tool was rolled out across the entire firm.
Like A&O Shearman, Clifford Chance has demonstrated that successful AI implementation is no accident, but results from investment and clear strategy.
How to Implement AI in Law Firms
Even as major firms adopt AI technologies and productivity gains are realised in law and other sectors, the transformative potential of AI remains widely overlooked. AI will do more than bring new efficiencies to existing legal processes. It is likely to bring about substantial, structural changes to the business model of many law firms, and create space for new and more innovative firms to emerge.
In recent surveys conducted by Thomson Reuters, as many as 43% of lawyers said that they expected the adoption of AI to undermine the traditional hourly rate billing model on which most firms operate - as the number of hours required to complete legal work continues to fall. But as AI challenges existing models, there is an opportunity for a paradigm shift towards more sustainable practices. Here, lawyers could emulate the example of others in the professional services sphere, such as in management consulting - where firms have explored novel subscription-based and outcome-based revenue models.
AI implementation will require new skills from legal professionals, and a willingness to adapt to change. It will put a premium on creativity, communication skills and problem-solving, while potentially carving out wholly new positions for AI-specialists and implementation managers. These shifts represent opportunities, not threats. Freeing up time previously spent on repetitive tasks means that legal professionals can spend more time on activities which strengthen their practices, including building relationships with clients and strategic planning for their firms.
AI implementation has the potential to benefit practitioners - not just their clients. Lawyers who use AI effectively are lawyers who can spend more of their time on the expertise-driven and judgement-based activities where they can add the most value, and from which they derive the most personal satisfaction.
“The role of a good lawyer is as a ‘trusted advisor,’ not as a producer of documents . . . breadth of experience is where a lawyer’s true value lies.”
- Interview with an attorney in the Future of Professionals Report
There are promising directions for further innovation in legal AI technology. Specialised, agentic models could act as ‘legal personas’ that can approximate the expertise of human lawyers for more basic applications. Improvements in the reasoning of LLMs could expand the role of AI in drafting legal arguments, while continuing to decrease the prevalence of hallucinations and informational biases.
As the benefits of existing AI technology become increasingly clear, clients will expect their lawyers to be well-versed in using it. Those who prove unwilling or unable to make use of the best available technology risk falling behind their competitors. Even the largest, most well-established firms risk losing ground to new, more innovative firms unless they can successfully navigate fundamental changes to the legal industry.
It has never been more important for lawyers to develop a clear, coherent strategy for AI adoption. The successful implementation of legal AI solutions will require management of industry-specific dynamics, organisational structures and data privacy concerns. Done properly, AI implementation into existing workflows offers substantial benefits to both clients and practitioners, in both the quality of output and the time taken to complete various legal tasks. Equally, AI technology offers the potential to create entirely new ways of working, as it continues to reshape knowledge-intensive industries.
Exactimo helps ambitious companies make the most out of AI and Business Automation. Find out more here.