The role of in-house counsel, especially at small to medium-sized businesses (SMBs), demands agility and frugality. For solo in-house lawyers, you are the contract reviewer, compliance advisor, policy drafter, risk manager, dispute resolver, legal operations expert, and head of legal — all in one. If you are part of a small team, the experience is similar, though you tend to have more resources to share the workload. Even in small teams, each member often serves as a generalist.

Unlike large legal departments with deep benches and big budgets, SMB legal teams must do more with less. The good news for small or solo legal departments is that generative AI (Gen AI) tools present a significant opportunity to level the playing field.

Over the past 18 months, GenAI tools like ChatGPT, Claude, Copilot, and others have matured from novelties to powerful productivity boosters. Beyond these general-purpose AI tools, legal tech providers have created solutions specifically tailored for the legal profession, with a particular focus on in-house counsel — for example, Thomson Reuters CoCounsel and the suite of related products. GenAI can streamline how in-house lawyers work, reducing bottlenecks, increasing impact, and enhancing business support.

However, adopting these technologies is not without challenges. In-house legal teams are understandably cautious due to concerns about confidentiality, trustworthiness of AI outputs, and drafting prompts that get beneficial results. These are all valid considerations that should not be taken lightly or dismissed out of hand.

Fortunately, you can manage these challenges. This white paper explores practical, high-value ways for SMB in-house counsel to simply and easily integrate GenAI into their daily workflows while safely navigating key risks.

Pulling back the curtain: What GenAI is and isn’t

To start, there is no big mystery here — generative AI uses advanced computer algorithms and large-language models to produce human-like work based on user-generated prompts. In particular, GenAI tools are very good at generating text from user inputs, including document summaries, reviews, analysis, and even first-draft work product.

So, what does this mean? Think of GenAI as a sophisticated chatbot, one that enables incredibly human-like conversations and responses to questions. Unlike traditional search engines, GenAI interacts with you in a dialogue format, allowing for follow-up questions and iterative responses. However, it does not think or understand like a human. It works through pattern recognition, like a sophisticated autocomplete, synthesizing and generating responses based on patterns in the data it was trained on.

GenAI solutions offer legal professionals a range of capabilities designed to streamline their workflow. These tools can:

• Draft documents such as contracts, memos, and emails
• Summarize documents
• Simplify legalese into plain English for business briefings
• Help brainstorm or outline
• Identify issues or risks based on input

What AI cannot do is provide guaranteed accurate legal advice, replace human judgment, or understand the nuances of a client’s question with the context of the situation — only a human can do that. Consider these tools as capable interns or summer associates; they provide valuable work, but require your oversight, and they cannot replace your experience and legal acumen.

Choosing the right tools

Choosing the right version of GenAI depends on your professional needs. There are two main levels of GenAI available today. First, there are consumer-focused tools like ChatGPT, Claude, Gemini, and others. These tools are flexible, customizable, and good for general brainstorming and business productivity uses, but they typically lack the data security and legal foundation required for more critical work. Second, there are professional solutions, such as sophisticated legal tech, including legal AI assistants, legal research, e-discovery, compliance, and legal-focused drafting tools. Some examples are Thomson Reuters CoCounsel Core, CoCounsel Drafting, and Westlaw Precision with CoCounsel.

These generative AI tools are much safer for legal teams since they are domain-specific and built with legal guardrails around databases and the use of confidential data. Regardless of which GenAI tools you choose, strongly consider enterprise or professional versions of these tools for enhanced data security and privacy. Doing so ensures that your business’s data and information will not be retained or used for training purposes.

Concerns with using consumer-grade GenAI and how to solve them

When using consumer-grade GenAI, exercise caution when inputting information. Any data entered may be used for training, shared with others, or potentially exposed in the event of a data breach. Any one of these can put lawyers in a tough spot when it comes to protecting confidentiality and preventing waiver of attorney-client privilege. Here are two ways you can mitigate these concerns:

1. Avoid entering confidential or proprietary information. Refrain from using client names or details that could compromise privilege. Many queries do not require confidential information, like “Explain the insider trading laws in the United States.” For requests involving specific names or confidential data, perform a “find and replace” to substitute generic terms like “Big Co” or “Vendor.” Be vigilant about other identifiers, such as addresses or proprietary formulas, that may inadvertently disclose confidential information. This approach requires thorough review and sanitization of documents.

2. Leverage GenAI tool settings to protect your data. Take a few minutes to utilize available controls to enhance data security:

• Use a strong password and turn on multifactor authentication
• Don’t give the tool login information to other accounts
• Use “temporary chat” features to prevent conversations from training the model
• Regularly clear the memory
• Turn off “improve model for everyone”
• Delete all chats daily or on a set schedule

While these measures are not foolproof, they substantially reduce risks.

Trusting generative AI results can present complex challenges, particularly when not using embedded GenAI. Although instances of generating inaccurate information are limited, they do occur. As noted above, treat GenAI like a valuable intern or summer associate. Provide background and context for each assignment, follow up to refine outputs, and independently verify uncertain results. This diligence doesn’t detract from the tool’s value, as it typically delivers a solid first draft.

If your budget and priorities allow, implementing a GenAI system specifically designed for legal is a much better solution, and most of the above controls can be handled during implementation and relied on thereafter.

Crafting precise prompts for optimal GenAI results

Utilizing generative AI effectively begins with crafting precise prompts. Prompts are simply the questions or requests you enter into the tool. The better the prompt, the better the results. Here is a quick summary of the basics, complete with examples:

1. Be specific. Detail exactly what you need: 

• Simple – “Explain GDPR.”
• Detailed – “Explain the duties of a data processor under GDPR.”
• Comparative – “Explain the differences between a data processor and a data controller under GDPR.”

2. Specify the response format. Define how you want the information presented:

• “Explain GDPR in a memorandum.”
• “Explain GDPR by setting out each provision in table form by numbered section.”

3. Give it a persona. Tailor the response by defining the perspective:

• “You are an experienced U.S.-based data privacy lawyer. Explain GDPR by setting out each provision in table form by numbered section.”

4. Identify the audience. Ensure clarity for the intended readers: 

• “You are an experienced U.S.-based data privacy lawyer. Explain GDPR by setting out each provision in table form by numbered sections. Write your explanations at a level an average business executive can understand.”

5. Provide context. Offer background to refine the response:

• “You are an experienced in-house commercial lawyer based in the U.S. You have received a contract from a vendor to review. Writing from the company’s point of view, prepare a table setting out each section of the contract and why that section is good or bad for the company. Add a column to set forth how you would change the wording to be more favorable to the company and how you would justify the edit to the lawyers for the vendor.”

6. Consider both perspectives. Explore all angles of an issue:

• “You are an experienced in-house commercial lawyer based in the U.S. You have received a contract from a vendor to review. Prepare a table setting out each section of the contract and, in separate boxes, why that section is: good or bad for the company, how you would change the wording to be more favorable to the company and how you would justify the edit to the lawyers for the vendor, what arguments the lawyers for the vendor will make opposing the changes, and how you would respond to those arguments.”

7. Break it down into steps. Structure complex tasks into manageable parts:

• “Draft an email supporting the legal department’s need for a contract management system using the following steps:
  • Introduction. Draft an introduction discussing the adoption of contract management systems by legal departments over the past 10 years.
  • Benefits. Describe five major benefits to the company and the legal department of using a contract management system, including examples. Include examples of how other parts of the company can use the system.
  • Challenges. Outline the key challenges to adopting and using a contract management system and how best to overcome each.
  • Future trends. Predict the savings to the company in dollars and time over the next five years if it adopts a contract management system today.
  • Conclusion. Summarize the main points and provide a call to action for purchasing a contract management system for the legal department.”

8. Follow up. Refine the output: 

• “You are an experienced contract lawyer and negotiator. You represent the vendor in a potential sale to a customer. The customer has sent you its standard procurement agreement. Summarize this contract with a table that outlines each party’s duties, fees, deadlines, rights granted, and representations.”
• Follow-up 1 – “Expand on the rights granted section.”
• Follow-up 2 – “Redraft this contract in a manner most favorable to the vendor.”

9. Modify. Adjust as needed:

• “You are an expert U.S.-based competition lawyer. Prepare a summary of Section 2 of the Sherman Act.”
• Modify 1 – “Add 10 examples to this summary of how product manufacturers have run afoul of Section 2 of the Sherman Act.”
• Modify 2 – “Turn this into an email to senior executives. Use words appropriate for business executives lacking much experience with legal issues or antitrust laws.”

These prompt guidelines will help you utilize GenAI’s capabilities for practical applications. By crafting well-defined prompts, you can generate more precise results.

High-impact GenAI use cases for SMB in-house lawyers

With some basic prompt engineering skills in hand, it’s time to dive into some practical generative AI use cases and corresponding prompts. These use cases are designed to show a handful of practical ways you can leverage GenAI to be more efficient:

• Drafting and revising contracts. Create base templates or clauses from scratch and modify based on your input. For example, “Make this clause more tenant-friendly.” While it won’t replace your expertise, it can save you hours getting to a first or second draft.
• Redlining. Input two contract versions and prompt GenAI to summarize the changes or highlight non-market terms. Although not a substitute for professional redlining tools, it provides useful high-level comparisons.
• Contract summarization. Brief executives on lengthy contracts using GenAI to extract key commercial, legal, and risk terms into a memo or table format.
• Policy drafting. Draft, localize, and simplify policies from employee handbooks, FCPA policies, and website privacy notices for different audiences, and translate legal policies into plain English for greater clarity and understanding.
• Legal research and issue spotting. Brainstorm potential legal issues for new products, summarize areas of law, and provide starting points for deeper research.
• Drafting internal communication. Write or polish stakeholder updates, compliance alerts, or board summaries, tailoring the tone to suit the audience. For example, “Make this more executive-friendly.”
• Vendor due diligence support. Generate due diligence checklists, draft or review questionnaires, or summarize a supplier’s data processing policy. These capabilities are invaluable in procurement-heavy industries.
• Training materials and FAQs. Develop legal FAQs or create training materials. For example, “Explain anti-bribery rules for our international sales team in a presentation.”
• Project management and checklists. Create checklists, such as “Create a checklist for reviewing a SaaS vendor agreement.” 
• Brainstorming. Explore new ideas or directions with GenAI as your brainstorming partner to serve as a catalyst for innovation.

A dedicated legal GenAI tool may incorporate many of these skills, with dedicated workflows and legal guardrails ensuring the model uses the correct law and practice for your situation. As with all things GenAI, you get more out if you put more in — both in the choice of a robust, domain-specific tool and in your time and effort in crafting a detailed prompt.