7. Informed examples of how law firms are adopting generative AI
Generative AI is being adopted by law firms to try and improve how they work. One of the main ways they are doing this is by using AI for tasks that take a lot of time. For example, in big commercial transactions or litigation, lawyers have to do due diligence or disclosure, which means reviewing thousands of documents. AI tools can scan these documents much faster than a human can, looking for key clauses or specific information. This offers better value for clients because the firm can charge lower fees for these tasks, or get them done quicker (Legal IT Insider, 2023). This can make a firm more competitive when trying to win new work.
Firms are also using AI to increase what they can do for clients. By automating some of the more basic work, lawyers have more time to spend on giving strategic advice, which is what clients value most. This also helps with innovation. Firms that are seen to be using the latest technology, like generative AI, build a reputation for being modern and forward-thinking. This can attract both new clients and also the best lawyers who want to work at a firm that is investing in the future. For instance, the use of AI in contract drafting can create first drafts of standard agreements, which a lawyer then checks and customises. This speeds up the process and can ensure consistency across the firm’s documents, which improves quality.
This adoption is also a response to client demand. Corporate clients are under pressure to reduce their legal spend, so they want their law firms to be more efficient. A firm that can show it is using AI to reduce costs has a significant competitive advantage. For example, Allen & Overy’s use of the AI platform ‘Harvey’ was widely publicised, and it created the impression that the firm was at the forefront of legal technology, enhancing its reputation and competitive position in the market (Financial Times, 2023). By using AI, firms are not just cutting costs; they are trying to change their business model to deliver legal services in a new way that is more aligned with what modern clients expect.
8. How two UK law firms are using generative AI
The two firms I will discuss is Clifford Chance and Keoghs. These firms are very different, which shows how AI can be used in different parts of the legal market. Clifford Chance is a large international ‘Magic Circle’ firm that works on big corporate deals, whereas Keoghs is a specialist insurance law firm that handles a high volume of claims.
Clifford Chance has developed its own private generative AI tool called ‘Clifford Chance Assist’ (Clifford Chance, 2023). This tool is built using Microsoft’s Azure OpenAI service, which means it operates in a secure environment, which is very important for protecting client confidentiality. The firm is using this AI to help its lawyers with tasks like legal research, summarising long documents, and even drafting clauses for contracts. The impact of this is that lawyers can get through their work more quickly. For a firm like Clifford Chance, where deals are very time-sensitive, this speed is a major benefit. It allows the firm to handle more work and to provide a more responsive service to clients.
This strategy has an impact on the firm and its lawyers. For its overall strategy, it positions Clifford Chance as a leader in innovation. In terms of career development, junior lawyers may now spend less time on document review and more time learning how to use these AI tools effectively. They are being trained in ‘prompt engineering’, which is the skill of asking the AI the right questions to get the best results (The Lawyer, 2023). This is a new skill for lawyers. A challenge here is making sure that junior lawyers still get the foundational knowledge that used to come from doing the detailed document review work themselves. The firm needs to make sure its training programmes are updated to account for this.
Keoghs operates in a different market. As a defendant insurance law firm, a lot of its work involves handling a large number of similar cases, such as motor insurance claims. Efficiency and data analysis are key. Keoghs, which is part of the Davies Group, has invested heavily in technology and has an AI platform called ‘Ki’ (Keoghs, 2024). This tool is used to analyse data from thousands of insurance claims to spot patterns, identify fraudulent claims, and predict the likely cost of a claim. This is very different from Clifford Chance’s use of AI for bespoke corporate work.
The impact for Keoghs is that it can manage claims much more efficiently for its insurer clients. By using AI to triage cases, the firm can quickly identify which claims are straightforward and can be settled quickly, and which are complex and need an experienced lawyer’s attention. This saves their clients money. The opportunity for Keoghs is that this data-driven approach gives them a strong competitive advantage in the insurance market. It changes the firm’s strategy from just providing legal advice to providing data analytics and risk management services. For lawyers’ training, it means they need to be comfortable with data and statistics, not just the law. A challenge is the risk of the AI making mistakes in analysing claims, which could lead to a genuine claim being wrongly flagged as fraudulent, so human oversight is still very important.
9. Potential risks in using generative AI and how firms can mitigate them
There are several potential risks when using generative AI in legal work. The first and most obvious is accuracy. Generative AI models can ‘hallucinate’, which means they can create information that is completely false, including fake case law or legal principles (SRA, 2023). If a lawyer relies on this false information in advice to a client or in a court document, the consequences could be very serious, including professional negligence claims.
Another major risk is client confidentiality. Many commercial generative AI tools are public, and any information entered into them could be used to train the model and may not be secure. This would be a clear breach of a lawyer’s duty of confidentiality to their client. This is why firms like Clifford Chance are building their own private, or ‘walled-garden’, versions of these tools.
There is also the risk of bias. AI models are trained on vast amounts of text from the internet, which can contain historical biases. The AI could therefore produce outputs that are discriminatory. In a legal context, this could influence decisions about recruitment, or even legal arguments, in a way that is not fair.
To mitigate these risks, firms must have strong governance policies in place. The most important mitigation is to always have a ‘human in the loop’. This means that any work produced by an AI must be carefully checked, verified, and approved by a qualified lawyer before it is used. The lawyer remains professionally responsible for the work. The Solicitors Regulation Authority (SRA) has emphasised that existing professional standards apply to work produced with the help of AI (SRA, 2023).
Firms can also mitigate risks through technical solutions. As mentioned, using private, secure AI platforms is essential to protect confidentiality. Firms must also provide extensive training to all staff on the limitations of AI and the firm’s policies for using it. This includes training on how to check and verify AI outputs and how to spot potential hallucinations or bias.
10. Suitability of generative AI for certain types of firms and tasks
Generative AI is not equally suitable for all parts of legal practice. Its suitability depends on the type of task, the area of law, and the type of firm.
AI is particularly suitable for tasks that are repetitive, structured, and involve processing large amounts of information. This includes tasks like document review in e-disclosure for litigation, due diligence in corporate mergers and acquisitions, and reviewing and summarising standard form contracts. These are often called ‘low-value’ tasks, and using AI for them frees up lawyers for more complex work. Practice areas like conveyancing, personal injury, and insurance claims, which often have high volumes of cases that follow a similar process, are well-suited for AI automation. Keoghs’ use of AI in the insurance sector is a good example of this.
However, AI is currently less suitable for tasks that require deep human judgment, strategic thinking, empathy, and persuasion. For example, it is not suitable for conducting a cross-examination in court, advising a client on a highly sensitive family law matter, or negotiating the key terms of a unique, high-stakes commercial deal. These tasks rely on a lawyer’s experience, intuition, and ability to understand human relationships, which an AI cannot replicate. Bespoke advisory work in areas like complex tax law or regulatory investigations still requires human expertise at its core.
The type of firm also matters. Large, wealthy firms like those in the Magic Circle have the resources to invest millions of pounds in developing their own custom AI tools or paying for expensive subscriptions. This gives them an advantage. Smaller high street firms may not have the budget for this and might have to wait for cheaper, off-the-shelf products to become available. This could create a ‘tech divide’ in the legal market, where large firms pull even further ahead of smaller ones.
11. References to theories of leadership, strategy and business development
The adoption of AI by law firms can be understood through theories of business strategy and leadership. Michael Porter’s theories on competitive strategy are relevant here (Porter, 1985). By using AI to lower the cost of their services or to offer a new, differentiated service (like Keoghs’ data analytics), firms are trying to gain a competitive advantage. AI can affect several of Porter’s ‘Five Forces’. It can lower the threat of new entrants by creating high investment costs, and it can increase the bargaining power of firms over their clients if they offer a unique and valuable AI-driven service.
Implementing a new technology like AI across a whole firm is a major change management challenge. John Kotter’s 8-Step Process for Leading Change provides a useful framework for this (Kotter, 1996). A firm’s leadership needs to create a sense of urgency about why adopting AI is necessary (e.g., to stay competitive). They need to build a guiding coalition of enthusiastic partners and staff to lead the project. They also have to communicate a clear vision for how AI will be used and empower employees to act on that vision by providing training and resources. Finally, they need to generate short-term wins (e.g., a successful pilot project) to build momentum and anchor the new approaches in the firm’s culture. Without this kind of strategic leadership, an expensive investment in AI could fail because lawyers may be resistant to changing the way they have always worked. The success of AI in a law firm is as much about people and culture as it is about the technology itself.
—
**Note on Reflective Statement:**
I am unable to prepare the Reflective Statement as the instructions for it, mentioned as being on page 1 of the document, were not provided.
References
- Clifford Chance. (2023) Clifford Chance launches ‘Clifford Chance Assist’, its own secure generative AI tool. Clifford Chance News.
- Financial Times. (2023) ‘Magic circle’ law firm Allen & Overy signs deal with OpenAI-backed start-up. Financial Times, 15 February.
- Keoghs. (2024) ki: augmenting our intelligence. Keoghs.co.uk.
- Kotter, J.P. (1996) Leading Change. Harvard Business School Press.
- Legal IT Insider. (2023) Macfarlanes enters Gen AI partnership with Luminance. Legal IT Insider, 20 November.
- Porter, M.E. (1985) Competitive Advantage: Creating and Sustaining Superior Performance. Free Press.
- Solicitors Regulation Authority (SRA). (2023) Artificial intelligence. sra.org.uk.
- The Lawyer. (2023) Squire Patton Boggs rolls out Co-Counsel to all lawyers globally. The Lawyer, 26 October.
