In this article, we explore elements of the Process & Operations pillar of the Target Operating Model as introduced in AI for in-house legal teams: five pillars for success. 

We discuss how legal teams should strategically design and optimise their workflows to maximise AI's impact while delivering immediate operational improvements that create value regardless of technology adoption.

The challenge

Your legal team has invested in AI technology, but the expected improvements haven't materialised. Turnaround times remain unchanged, errors persist, and reporting is still manual. This common scenario is often the result of applying AI to fragmented or undocumented processes. Across the legal sector, we observe a consistent pattern: AI implementation reveals existing process weaknesses, making inefficiencies more apparent, and sometimes more problematic. Successful AI transformation requires a strategic approach to process design that addresses these operational challenges and considers new AI-era requirements, such as where to embed human oversight within AI-enabled workflows. The central question for legal leaders becomes: How do you design and optimise your processes to create the operational foundation necessary for AI to deliver meaningful value?

When AI meets broken processes 

To understand why process design is so critical for AI success, consider the operational challenges that many legal teams already face. These foundational issues undermine efficiency and effectiveness whether AI is involved or not, but AI implementation makes them impossible to ignore. Common challenges include:

• Fragmented intake: Matter requests / contracts arrive via email, Teams messages, in-person conversations, and phone calls with no central system or tracking.
• A lack of version control: Multiple versions of templates stored across different drives, SharePoint sites, and personal folders with no single source of truth.
• Manual tracking: Matters are tracked in individual spreadsheets that quickly become outdated and can't be consolidated for reporting.
• Invisible workflows: Approval processes that rely on individuals remembering to chase colleagues, with no transparency on status or bottlenecks.
• Inconsistent execution: Five different lawyers handling the same request in five completely different ways.
• Data trapped in documents: Critical information is unstructured and captured in Word files and email threads, impossible to analyse or report on.

Historically, some teams have tried to address these challenges through manual effort and workarounds rather than reviewing and updating these operational elements. However, these challenges are worth addressing regardless of AI aspirations - they represent fundamental inefficiencies that limit team performance and organisational value delivery. It is important for legal teams to consider: "How can we resolve these inefficiencies and unlock AI's full potential?" The solution lies in adopting a strategic approach that prioritises where to begin. 

A strategic approach to process design  

As discussed in our previous articles, successful AI transformation begins with a clear understanding of where AI will deliver the most value for your legal team and the broader organisation. Your strategy should have identified the high-impact use cases - whether that's contract review, matter intake, compliance monitoring, etc. These strategic priorities become your process improvement roadmap. Rather than trying to fix everything at once, it will be important to focus your process redesign efforts on the workflows that support your highest-value AI applications. This targeted approach ensures that your process investment directly supports your strategic objectives and delivers measurable impact where it matters most. 
Once you've identified your starting point, the next critical step is assessing your current process maturity. This evaluation will help you to determine how much foundational work is needed to prepare your chosen workflow for AI integration. The assessment should examine whether your existing process has the standardisation, data quality, structure and accountability necessary for successful AI implementation, or whether more extensive redesign is required before AI can deliver value.

Assessing Process Maturity 

To identify the level of effort required to ensure your processes are AI-ready, we recommend starting with an evaluation of the following elements:

1. Standardisation

• Does everyone on your team handle the same type of work in the same way? Or does each person have their own approach?
• Can you clearly explain what "good work" looks like for this process so that success can be measured?
• Are there clear steps and decision points in your process that could be followed consistently? i.e. could a new team member understand the process without extensive explanation? 

2. Data Quality

• Is your data (particularly the data relevant to your chosen process) stored in an organised way that can be easily searched and analysed? Or is critical information buried in unstructured documents and email chains?
• Do you have centralised document management with robust version control processes, naming conventions, and organised metadata? Or are documents scattered across various systems with no structure?
• Can you track how often errors occur and where they typically happen in your process?

3. Structure and Accountability  

• Do you have established escalation procedures when processes encounter exceptions or delays?
• Do you have clear guidelines about who needs to approve different types of decisions?
• Have you identified which parts of your process would cause the biggest problems if done incorrectly?

Your answers to these questions will help you to determine your path forward. If you answered "no" to most questions, your processes likely need significant overhaul before AI can deliver tangible results. Prior to implementation, at a high level, it will be crucial to:

1. Map the current ad hoc or informal process, identify key pain points, and classify high-risk components of the process. Spend the time breaking down the workflow to understand each activity and individual step (i.e. I receive the contract, then I review our contract database to determine if we have an existing agreement in place, etc). It will be important to have a deep understanding of each component of your current workflow in order to identify both the pain point and the opportunity.
2. Leverage Lean Six Sigma (a data-driven methodology focused on eliminating waste and variation within a process) principles to systematically analyse the process and pinpoint opportunities to improve consistency and efficiency.
3. Design a new process that addresses these pain points, establishes clear approval and escalation procedures, and integrates AI functionality. This involves considering the identified risks and determining where AI enablement will deliver the most value. The redesign may require modifying or eliminating steps in your current process to properly account for AI's new role in the workflow.
4. Document the new process in a comprehensive playbook, standard operating procedure, or guide, and develop the supporting infrastructure required for implementation (such as structured databases, templates, precedents, and other relevant resources).
5. Roll out practical training for your team on this new process and ensure adherence to the new process is modelled by senior leaders in the team. Ensure you have involved the broader team at various points in the redesign to reduce resistance and support buy-in. 

Conversely, if you answered "yes" to most questions, you are likely to have a strong foundation for AI integration and can focus on targeted optimisation to unlock AI's potential. However, regardless of your current process maturity level, a critical consideration in the AI era is determining where human oversight should be embedded within your AI-enabled workflows.

Embedding Human Oversight in AI-Enabled Processes

The steps for assessing and optimising your processes outlined above are likely not new to experienced legal operations professionals. However, an element that is distinctly new in the AI era is determining where to embed "human in the loop" verification within your AI-enabled processes. Australian government guidance emphasises that human oversight of AI systems must be enabled, with those responsible for different phases of the AI system lifecycle remaining identifiable and accountable for outcomes. This requires careful consideration of risk tolerance and the specific AI application being deployed. Here are some key factors to consider:

• Risk Assessment: Consider the potential impact of AI errors in each process step. High-risk decisions (such as contract approvals above certain thresholds or regulatory compliance determinations) should always include human verification, while lower-risk tasks (such as initial document categorisation or routine data extraction) may operate with minimal oversight.
• AI Confidence Levels: Modern AI systems can provide confidence scores for their outputs. Design your process to automatically flag low-confidence results for human review while allowing high-confidence outputs to proceed with minimal intervention.
• Exception Handling: Build in human escalation points for edge cases or unusual scenarios that fall outside the AI's training parameters. This ensures your process remains robust even when encountering novel situations.
• Stakeholder Expectations: Consider the comfort level and expectations of internal clients and external parties. Some stakeholders may require human confirmation even for low-risk AI decisions, particularly during the early stages of AI adoption.

Moving forward 

AI transformation in legal teams isn't just about implementing new technology – it's about rethinking how work gets done. Legal teams need a strategic approach to process design, an honest evaluation of current process maturity, and careful consideration of where human oversight matters most. This creates the foundation AI needs to deliver real value. Whether your processes need fine-tuning or complete overhaul, getting these workflows right determines whether AI becomes a game-changer or another underutilised tool in your technology stack. Legal teams that optimise their workflows now will solve today's operational challenges while building the foundation for tomorrow's AI success.