How Law Firms Can Use AI for Intake, Client Communication, and Operations | Alejo Pijuan

How Law Firms Can Use AI for Intake, Client Communication, and Operations

Most law firm owners thinking about AI are asking the wrong first question.

The question most owners ask is: which AI tool should we use? The question that actually leads somewhere useful is: where is our firm losing time or clients right now, and could AI help fix that specific problem?

Alejo Pijuan is a fractional AI officer who works with law firms on AI strategy and implementation. Through his company AmplifyVoice, he helps law firms stop losing clients to voicemail, handle intake more efficiently, and build the internal AI infrastructure that matches how the firm actually operates. In a recent conversation on Your Profitable Law Firm, Alejo walked through the people, process, tools framework he uses with every client, how attorney-client privilege intersects with AI, and the one AI tip that changes how accurate your results will be.

Why 60% of Law Firm Calls Go to Voicemail (and What It Costs)

When Alejo started working with law firms on AI voice agents, one of the first firms he analyzed had 60% of incoming calls going to voicemail.

Not because the staff was incompetent. Because the volume of calls exceeded the firm's capacity to answer them consistently, especially during busy periods and after hours.

Every call that went to voicemail represented a potential client who, in most cases, called the next firm on their list. They did not leave a message and wait. They moved on.

This is the intake problem that AI can solve most directly and most immediately for law firms. An AI voice agent can answer calls when no human is available, conduct a natural intake conversation, collect the information the firm needs, qualify the lead, and schedule a consultation, all without the caller being sent to a generic voicemail box.

Alejo is careful about what he promises here. The AI is not perfect from day one. Every firm he works with goes through a two-week hyper-care period after launch where his team monitors every call, identifies where the AI struggled, and makes adjustments. The AI gets significantly better over that period, and the process of training it to handle a specific firm's nuances is what separates a working system from one that frustrates callers.

The People, Process, Tools Framework for AI Implementation

The most common AI implementation mistake Alejo sees, regardless of firm size, is starting with the tool.

An executive team decides the firm needs to do something with AI. They evaluate tools. They pick one. They tell the team to start using it. It fails or underperforms, and everyone concludes that AI is not ready yet.

The problem is not the tool. It is the sequence.

Alejo's framework starts with people. What is the biggest pain point for the people inside the firm? Who experiences it, and at what stage of the client lifecycle does it appear? Only after mapping the people and the pain does the conversation move to process. What is the specific process that is broken, missing, or creating that pain? And only after understanding the process does it make sense to ask which tool could help fix it.

People, process, tools. In that order. Every time.

The reverse sequence, tool, process, people, is how firms end up with expensive software that nobody uses and problems that remain exactly as large as before.

Attorney-Client Privilege and AI: What Law Firms Need to Know

Attorney-client privilege is the concern Alejo hears most often from attorneys who are curious about AI but hesitant to move forward.

His answer is honest: the law is still being worked out.

Judges have ruled on both sides. Some rulings have found that AI-processed information constitutes work product or falls within privilege protections. Others have found it does not. The legal landscape is actively evolving, and any firm relying on a blanket assurance that AI use is fully protected would be taking on more risk than they realize.

What firms can control is where the data goes and what AI providers do with it. Every AI provider has terms of service that specify whether they store conversation data, whether they use it to train future models, and what security standards they maintain. Reading those terms, or having someone like Alejo read them on the firm's behalf, is the starting point for making an informed decision about which tools are appropriate for which use cases.

The practical principle Alejo applies: treat AI data handling the way you would treat any other third-party vendor handling sensitive client information. Ask where the data is stored, who can access it, whether it is used for model training, and whether the provider can execute a Business Associate Agreement or equivalent data protection agreement if needed.

Alejo has produced a dedicated 45-minute webinar on this topic specifically for law firms. Link: [PENDING — add once received from Alejo]

The Most Useful AI Tip in This Episode

Alejo shared one piece of practical advice that applies to every AI interaction, regardless of tool, task, or firm size.

After you ask AI a question and receive an answer, ask it a follow-up question:

"How confident are you in that answer, and what might you have missed?"

This single follow-up changes the quality of AI output significantly. It surfaces the AI's own uncertainty, flags the areas where its answer may be incomplete or based on limited information, and gives the human reviewer a specific signal of where to focus additional attention.

Without this prompt, most people receive an AI answer that sounds complete and confident and accept it as such. With it, the human stays in the loop in a meaningful way rather than simply handing work off and hoping the output is accurate.

Alejo described this as the difference between using AI to accelerate yourself versus using AI to replace your judgment. The goal is always the former.

How the Right AI Strategy Changes by Firm Size

One of the most useful frameworks Alejo shared is that the right AI approach looks meaningfully different depending on how many lawyers a firm has.

Small Firms: 5 Lawyers or Fewer

For small firms, Alejo's primary recommendation is to start with education rather than implementation. Get every person in the firm using AI individually so they develop their own intuitions about what it can and cannot do well. The Claude Teams plan is his recommended starting point because it provides governance over what data the team is putting into the system.

He has also created a seven-minute quick-start guide specifically for setting up the Claude for Legal plugin. Link: [PENDING — add once received from Alejo]

The goal at this stage is not to build a system. It is to get brains moving. Once attorneys have hands-on experience with AI tools, they have an informed basis for a conversation about where AI could actually help.

Mid-Size Firms: 6 to 50 Lawyers

For firms with multiple departments and specializations, Alejo recommends thinking vertically rather than firm-wide. How is the marketing function using AI? How is the personal injury practice group using AI? Think department by department and client lifecycle stage by stage rather than trying to implement something across the whole firm at once.

The highest-return AI investments for mid-size firms, in Alejo's experience, are intake automation and marketing. These are the areas where the firm has the most to gain from faster, more consistent processes and where AI can be deployed without touching privileged client work directly.

Large Firms: 50 or More Lawyers

For larger firms, the biggest risk is the top-down implementation. An executive team gets excited about AI, selects a tool, and mandates adoption across the organization. The tool underperforms, adoption is poor, and the AI initiative is quietly shelved.

Alejo's recommended approach for large firms is to start with a small contained pilot of five to ten people focused on a specific, well-defined problem. Run the pilot, measure the impact, and only then conduct interviews with the broader group of people who would eventually be affected by a larger rollout. The interviews surface the real pain points at the ground level, which are often different from what the executive team assumed.

When the people feel heard and their specific problems are part of the solution design, adoption is substantially higher and the implementation is far more likely to succeed.

Where AI Should Not Go in a Law Firm

Alejo is equally direct about where AI should not be deployed, at least not yet.

The pattern he sees most often in failed implementations is firms that connect AI to systems without fully understanding what the AI will do with access to those systems. QuickBooks AI is one example that came up in the conversation: a widely used tool whose internal AI produces enough errors that the time spent correcting its output exceeds the time the AI was supposed to save.

His general principle: if a wrong AI output in this system would be difficult to detect, expensive to fix, or harmful to a client, that is not where AI should be operating autonomously. It can assist. It can draft. It can summarize. But a human needs to remain the decision-maker in any process where an error carries meaningful consequences.

Key Takeaway

AI is not a tool problem for law firms. It is a sequencing problem.

Firms that start by asking which tool to use will consistently underperform firms that start by asking where the pain is, whose problem it is, and what process is failing. The tool comes last, not first.

The firms that get AI right are the ones that treat it the same way they would treat any other significant operational change: understand the people, map the process, then select and deploy the tool.

Free Resource for Your Profitable Law Firm Listeners

Alejo is offering a free 30-minute consultation for podcast listeners. In that time he can assess where AI can help your firm most and where it should not go. Link: [PENDING — add once received from Alejo]

Connect with Alejo:

Free 30-minute consultation: https://cal.com/alejo-amplify-voice/30min-w-alejo 

Claude for legal quick start guide: https://www.youtube.com/watch?v=db6i06tTIAs 

"Is Claude For Legal Safe?" Webinar (aka Attorney-client privilege in AI webinar): https://www.youtube.com/watch?v=Lg7NWXUZzRM 

My LinkedIn profile if they want to get in touch: https://www.linkedin.com/in/alejopijuan/

Frequently Asked Questions

What is a fractional AI officer, and does a law firm need one?

A fractional AI officer is an external specialist who embeds into a firm part-time to develop and implement AI strategy, train teams, and oversee the people, process, and tool decisions that determine whether AI actually works. Small firms may not need one initially, but mid-size and larger firms benefit from having someone who can map the firm's specific pain points to the right AI solutions rather than deploying tools and hoping for the best.

How does AI handle attorney-client privilege during intake?

This is an area where the law is still evolving. Courts have ruled on both sides of whether AI-processed intake information falls within privilege protections. The practical approach is to evaluate each AI provider's terms of service carefully, focusing on where data is stored, whether it is used to train AI models, and what security agreements the provider will execute. Alejo has produced a 45-minute webinar specifically on this topic for law firms: [PENDING — add link once received].

What AI tools should a small law firm start with?

Alejo recommends the Claude Teams plan as the best starting point for firms with five lawyers or fewer. It provides governance over what data team members are putting into the system and supports the Claude for Legal plugin, which can be set up in approximately seven minutes. The goal at this stage is experimentation and education, not full-scale deployment.

How do you train an AI voice agent for law firm intake?

The process starts with mapping the firm's current intake workflow: what calls come in, what information needs to be collected, how calls are routed, and what happens after the initial contact. The AI is then built around that existing process rather than replacing it. After a soft launch, a two-week hyper-care period involves monitoring every call and making adjustments. The system improves significantly over that period as edge cases are identified and handled.

What is the single most useful thing you can do to improve AI accuracy?

After receiving any AI answer, ask the AI a follow-up: how confident are you in that answer, and what might you have missed? This surfaces the AI's own uncertainty, flags areas where the answer may be incomplete, and gives the human reviewer a clear signal of where to focus additional scrutiny. It keeps the human in the loop as the decision-maker rather than simply accepting AI output at face value.

Related Reading on Your Profitable Law Firm

If this episode connected with where your firm is right now, these posts go deeper on related topics:

• How Law Firms Can Use AI to Fix Intake, Compress the Sales Cycle, and Stop Losing Clients

• How Law Firms Can Use AI to 2X to 10X Their Productivity 

• Plugging the Leaks: Strategies for Law Firm Profitability