Executive summary
AI automation in professional services should create capacity by removing repetitive work rather than replacing people. Firms should define the business process, underlying data, governance requirements and points of human oversight before selecting technology, so automation supports professional judgement instead of attempting to substitute for it.
About Sean Dendle
Sean Dendle is CEO of Cymax and works with professional services organisations on technology strategy, managed IT, cybersecurity, governance and the practical application of technology to business operations.
The practical insights in this article are drawn from Sean’s experience and observations of AI implementation in professional services.
AI automation is creating significant opportunities for law firms, accounting firms and other professional services businesses.
But the biggest opportunity may not be replacing people.
It may be giving them their time back.
Cymax CEO Sean Dendle has observed professional services firms introducing AI automation, identifying where implementations go wrong and assessing what firms need in place before automating their workflows.
His experience points to an important distinction: the objective of AI automation should not be to remove people from every process. It should be to remove unnecessary work from people.
That changes both what firms should automate and how they should approach AI implementation.
What is AI automation for professional services?
AI automation for professional services is the use of artificial intelligence to perform, support or streamline repeatable business processes that would otherwise require manual effort.
For law firms, accounting practices and other advisory businesses, this could include:
- managing high volumes of information;
- triaging correspondence;
- extracting information from documents;
- creating first drafts;
- checking process completeness; and
- improving consistency around clearly defined repetitive workflows.
However, that does not mean every process should be automated.
Professional services remain heavily dependent on judgement, context, trust and relationships. A lawyer, accountant or adviser is not valuable simply because they can process information. Their value also comes from interpreting that information, understanding context and advising the client.
The starting question should not simply be:
> Where can we use AI?
A better question is:
> Where is repetitive work consuming valuable capacity, and is that work suitable for AI automation?
What should professional services firms automate with AI?
The strongest candidates for AI automation tend to share several characteristics. They are repetitive, clearly defined and supported by accessible information. The desired outcome is understood, and there is a clear point at which human judgement or intervention takes over.
Dendle has seen AI used to help manage correspondence and repetitive processes while leaving professional judgement with the people responsible for the client outcome.
The aim is not automation for automation’s sake.
It is creating capacity.
AI automation examples in professional services
| Professional services application | How AI automation can help | Potential business impact | Where people remain important |
|---|---|---|---|
|
Client correspondence |
Manage and triage high volumes of incoming information |
Reduced manual workload and greater process consistency |
Exceptions, judgement, sensitive matters and client relationships |
|
Defined accounting workflows |
Reduce manual effort involved in repeatable elements of service delivery |
Greater capacity and potentially different service economics |
Professional oversight, interpretation and responsibility for the outcome |
|
Internal repetitive workflows |
Automate clearly defined activities once the process is properly mapped |
Reduced administration and more employee capacity |
Exceptions, accountability and higher-value work |
|
Process management |
Create guardrails around repeatable processes |
Greater consistency in how work is performed |
Process ownership, quality assurance and decision-making |
The business case should measure more than the number of tasks automated.
In one law-firm implementation, Dendle observed benefits extending beyond capacity. The firm improved consistency, created more time for client relationships and changed its view of future staffing requirements.
That is a broader business case than simply asking how many minutes AI can save.
AI workflow automation starts with the process, not the technology
One of the most important lessons from AI implementations is also one of the least glamorous:
AI cannot effectively automate a process the business itself has not properly defined.
In every AI implementation Dendle has seen, the data has needed work. He has also consistently encountered workflows that were not documented sufficiently for an automation engineer to simply arrive and automate them.
A process may have operated successfully for years because an experienced Practice Manager, lawyer, accountant or administrator knows what happens next. Much of that knowledge, however, may exist only in that person’s head.
They know which exceptions matter. They know when something should be escalated. They know when the normal process does not apply. They know which information they need before proceeding.
AI workflow automation forces some of that implicit knowledge to become explicit.
Before automating a workflow, firms need to understand:
- What triggers the process?
- What information does it require?
- What happens at each stage?
- What are the exceptions?
- Where are approvals required?
- Where is professional judgement required?
- When should the AI stop and a person take over?
An automation specialist may understand the technology but not the nuances of legal or accounting work. The professional understands the work but may not understand what needs to be defined for automation.
Neither perspective is sufficient on its own.
Is your firm ready for AI automation?
Before recommending an AI automation solution, firms should examine three areas.
1. Data
AI is dependent on the information available to it.
If information is fragmented, inaccessible, inconsistent or poor quality, that becomes part of the implementation problem. Firms should understand where relevant information resides, who can access it, how reliable it is and whether it is suitable for the proposed use.
Australian firms should also consider privacy obligations when personal information is entered into or produced by AI systems. The Office of the Australian Information Commissioner’s guidance on generative AI in the workplace highlights the importance of privacy impact assessments, appropriate settings, staff education, accuracy controls and clear retention practices.
2. Processes
A process that works because one experienced employee knows how to navigate it is not necessarily automation-ready.
The workflow needs to be understood and documented. That includes the normal process as well as exceptions, approvals, escalation points and areas requiring judgement.
3. Budget
There can be a gap between what organisations expect AI to achieve and what they expect to invest.
AI can already do an enormous amount, but successful implementation is not necessarily a matter of plugging in a product and expecting it to operate effectively within a few weeks.
The real cost may include:
- process mapping;
- data preparation;
- integration;
- testing;
- governance;
- employee training; and
- ongoing improvement.
The Australian Government’s Voluntary AI Safety Standard provides a useful governance reference point, particularly around accountability, risk management, data quality, testing, monitoring and meaningful human control.
What should professional services firms not automate?
Not every task should be handed entirely to AI.
There is an important distinction between using AI within a process and giving AI responsibility for the entire process.
Professional judgement, sensitive decisions, unusual exceptions and areas involving significant accountability may still require human oversight. That is particularly important in law, accounting and other professions where clients are relying on expertise rather than simply purchasing an administrative output.
The right question is not:
> Can AI do this?
Increasingly, the answer may be yes.
The better question is:
> Should AI do this without human oversight?
Those are very different decisions.
How should firms measure AI automation ROI?
Hours saved are useful, but they should not be the only measure of AI automation ROI.
Depending on the business problem, firms could measure:
- reduction in manual effort;
- employee capacity created;
- process consistency;
- reduction in delays or bottlenecks;
- service-delivery costs;
- client responsiveness;
- quality or error rates;
- staffing requirements; and
- additional time available for higher-value professional work.
Measurement should follow the original business problem.
If the problem is slow client response times, measure response times. If the problem is overloaded employees, measure the capacity returned. If the problem is inconsistent execution, measure consistency.
The ROI measure should ideally be established before implementation rather than selected afterwards to justify the project. The KPMG guidance on AI ROI measurement also emphasises the importance of including the full cost base, including licences, integration, training, workflow redesign, governance and monitoring.
Why change management matters in AI automation
The person who understands a process best may also be the person with the least capacity to help automate it.
Dendle encountered precisely this situation when discussing an AI project with the owner of a professional services firm. The person best placed to help map the processes was the Practice Manager.
The problem was that it had taken that person three weeks just to respond to an email because they were already overloaded.
Adding a large AI transformation project to that person’s workload would create another problem rather than immediately solving one.
A more practical approach is to start with a repetitive process consuming a meaningful amount of that person’s time. Use AI to create some breathing room. Then use some of the capacity created to improve processes, information and subsequent automation.
This can create a different employee experience. Instead of AI being something management imposes on employees, the employee experiences AI as something that made their job easier.
That person can then become an internal advocate for what happens next.
Can AI automation make professional services more human?
It sounds contradictory, but AI automation could make some professional services firms more human rather than less.
Professional services are fundamentally human businesses. Clients value expertise, but they also value responsiveness, judgement, trust and relationships.
All of those things require time.
If a lawyer spends less time managing repetitive correspondence, what could they do with those hours? If a Practice Manager spends less time completing repetitive administrative processes, what operational problems could they solve? If an accountant can deliver part of an engagement with significantly less manual effort, how could the firm improve the client experience?
This is the opportunity Dendle sees in AI: removing repetitive work so skilled professionals have more time to apply judgement, solve difficult problems, speak with clients and build relationships.
A useful leadership question is:
> What do we want our people to do with the time AI gives back to them?
The answer should form part of the AI business case.
When considering AI automation services, look beyond the technology
The implementation sequence matters.
A more useful order is:
Business problem → Process → Data → Governance → Human oversight → Technology → Measurement
Starting with a particular AI platform reverses that logic. Technology should be selected because it can help solve a clearly defined business problem, not because the organisation feels it needs to “do something with AI”.
This is particularly important while AI capabilities are evolving rapidly. A product can change. The underlying business problem remains.
Before implementing AI automation, professional services firms should be able to answer:
- What business problem are we solving?
- Where is repetitive work consuming capacity?
- Is the process sufficiently defined?
- Is the underlying data usable?
- Where is human judgement required?
- What governance is needed?
- What outcome will justify the investment?
- How will we measure whether it worked?
If those questions are not clear, the firm probably is not ready to select the technology.
AI automation should create capacity, not simply reduce work
Much of the public conversation about AI still revolves around job replacement. That framing is too narrow for professional services.
Yes, AI will change how work gets done. But a more useful strategic question for many firms is what happens to the capacity AI creates.
AI automation does not need to begin as a headcount-reduction strategy. It can be a capacity strategy:
- remove repetitive work;
- improve processes;
- create headspace; and
- redirect capacity towards the work where people create the greatest value.
For professional services, that means judgement, expertise, relationships, problem-solving and client experience.
That is a more useful definition of successful AI automation.
Not sure whether your firm is ready for AI automation?
Before investing in another AI platform, it is worth understanding whether your data, processes and technology environment are ready to support automation.
Cymax works with professional services firms to identify where AI can create practical value, what needs to be addressed first and where human oversight should remain. Our broader perspective on AI governance, digital transformation for legal firms and client experience and operational capability reflects the same principle: technology should support a clearly defined business objective.
The objective is not to automate because AI is available.
It is to identify a worthwhile business problem and determine the right way to solve it.
Talk to Cymax about AI automation.
> When repetitive work consumes professional capacity, automate the process, not the judgement.
Frequently Asked Questions
What is AI automation for professional services?
AI automation for professional services is the use of artificial intelligence to automate or support repeatable workflows within businesses such as law firms, accounting practices and other advisory organisations. The strongest opportunities generally involve repetitive work where the process can be clearly defined while appropriate human oversight remains.
What processes should a professional services firm automate first?
Start with repetitive processes consuming significant employee capacity. Before automating them, ensure the workflow, data requirements, exceptions and points requiring human judgement are understood.
Can law firms use AI automation?
Yes. AI can support law-firm processes such as managing and triaging high volumes of information and correspondence. Human oversight remains important where professional judgement, sensitive information or client-specific context is involved.
What should a firm do before implementing AI automation?
Assess the business problem, existing workflow, data quality, governance requirements, human oversight and realistic implementation budget before choosing the technology.
How much does AI automation cost?
There is no single reliable figure. Cost depends on the complexity of the process, quality and accessibility of the data, existing systems and the amount of implementation required. Firms should consider process mapping, data preparation, integration, testing, governance and change management as well as software costs.
What are the benefits of AI automation for professional services?
Potential benefits include reduced repetitive work, greater employee capacity, improved process consistency, faster service delivery and more time for professional judgement, client relationships and higher-value work.
