Summary (Key Takeaways)
- AI can cut the time it takes to produce a first draft, but in a regulated firm every output still needs checking, so the real saving is usually smaller than a demonstration suggests.
- The true cost of AI includes review time, overlapping subscriptions, training and governance, not just the licence fee.
- When staff choose their own AI tools, the firm can lose track of where client information is going, which turns a cost question into a compliance and cyber insurance question.
- The clearest returns come from applying AI to a specific, measurable bottleneck rather than adopting it because competitors appear to be.
- A short review every six months, focused on outcomes such as turnaround times, rework and client experience rather than usage figures, shows whether AI is earning its place.
Why Time Saved Doesn’t Automatically Mean Money Saved
Nearly every conversation about AI in professional services ends up at the same place: productivity. Partners and directors have seen the demonstrations. A client letter drafted in seconds, a set of meeting notes summarised before the call has finished, a lengthy document reduced to a single page of key points. It is natural to assume that if the work takes less time, the savings will follow.
Sometimes they do. When a fee earner or adviser spends less time on routine drafting, they have more time for the work clients actually value: judgement, advice and relationships. That is a large part of why AI has spread so quickly through accountancy practices, law firms, IFAs and brokers.
The difficulty is that the financial benefit is harder to pin down than the time benefit. A task that used to take thirty minutes and now takes one minute has not necessarily saved twenty-nine minutes of cost. Something else usually fills part of that gap, and in a regulated firm the biggest of those things is checking.
In a Regulated Firm, the Review Step Is Not Optional
Consider a typical example. A paraplanner uses an AI assistant to produce a first draft of a suitability report, or a solicitor uses it to prepare an initial version of a client care letter. The draft appears almost instantly. It reads well and looks professional.
But it still has to be reviewed. Figures must be confirmed against the client file. Product details, dates and fees need checking. The wording may need adjusting so it reflects the firm’s agreed approach and the client’s actual circumstances. If the AI has misunderstood something or filled a gap with a plausible guess, someone has to spot it. For firms regulated by the FCA or the SRA, responsibility for that document sits with the firm and the individual who signs it off, not with the software that produced the first draft.
None of this is a criticism of AI. Most useful tools work this way. Accounting software does not remove the need for an accountant, and a case management system does not replace legal knowledge. Good technology helps skilled people work more effectively; it does not remove the need for them. We made a similar point when we looked at why AI can save your firm hours but should never replace the human voice clients trust. The professional judgement is what clients are paying for.
The practical consequence is that any honest assessment of AI has to count review time. If a draft takes one minute to generate and fifteen minutes to check and correct, the saving is real but much smaller than it first appeared. And if the review is rushed because the output looks polished, the firm may be swapping a time cost for a risk cost.
Where the Hidden Costs of AI Build Up
Beyond review time, several other costs tend to accumulate once AI starts spreading through a firm. Individually they look small. Together they can change the picture considerably.
- Overlapping subscriptions: One team adopts a tool for drafting, another uses a different service for research, and someone else signs up for a meeting assistant. Before long the firm is paying for several products that do broadly similar jobs, often on separate cards or expense claims.
- Licences that are not being used: Seats bought during an initial burst of enthusiasm are not always reviewed later. Some users rely on the tool daily; others tried it twice and went back to their old way of working.
- Training and change time: Getting good results from AI takes practice. Time spent learning, experimenting and sharing techniques is worthwhile, but it is still time.
- Governance and administration: Someone needs to decide which tools are approved, what data they may be used with, how outputs are checked and how the policy is kept up to date. In a regulated firm this work cannot be skipped.
- Rework: When AI output is used without enough care, mistakes can surface later, sometimes in front of a client. Correcting them costs more than getting them right first time.
None of these is a reason to avoid AI. They are simply part of the true cost, and a firm that ignores them will find it difficult to tell whether its investment is paying off.
When Unapproved AI Turns a Cost Issue Into a Compliance Issue
There is a further pattern that deserves particular attention in professional services. When the approved tools do not quite fit what people are trying to achieve, staff tend to find their own workarounds. A personal account gets used for a work task. A free version of a tool fills a gap. Text from a client file is pasted into a service the firm has never assessed.
At that point the question is no longer just about wasted spend. Client information may be moving into systems that were never designed to hold it, under terms the firm has not reviewed. That raises questions about data protection, confidentiality and the firm’s ability to show a regulator, an auditor or a cyber insurer that it knows where sensitive information is held and who can access it.
Most firms underestimate how much of this is already happening. If you have not checked recently, it is worth reading our guide on whether you really know which AI tools your team is using, which sets out practical ways to find out without creating a culture of blame.
The answer is rarely to ban AI outright. Bans tend to push usage further out of sight. A better approach is to provide approved tools that genuinely meet people’s needs, set clear rules about client data, and make it easy for staff to ask before they try something new. For many firms already on Microsoft 365, that conversation starts with understanding what their existing licences include before adding anything else.
Start With the Bottleneck, Not the Tool
The firms that see the clearest returns from AI tend to approach it from the opposite direction. Rather than asking which AI product they should buy, they ask where work is slowing down and what that delay is costing them.
In a professional firm, those bottlenecks are often easy to name once you look for them:
- Client onboarding that stalls while documents are chased, checked and filed.
- File notes and meeting summaries that are written up late, or inconsistently, because advisers are back to back with appointments.
- Monthly or quarterly reporting that takes hours of manual collation every cycle.
- Staff copying the same client details between a CRM, a back-office system and a spreadsheet, with every re-key creating a chance for error.
That last example is particularly common. We explored it in detail when we looked at the hidden compliance cost of people acting as the glue between systems. In many cases the best fix is a straightforward integration or automation rather than AI at all, and it is worth being honest about that.
When AI is applied to a clearly defined problem like these, it becomes far easier to see whether it is delivering value. You know how long the task took before, you can measure how long it takes now, and you can see whether quality has held up. When AI is introduced simply because everyone else seems to be using it, the technology arrives first and the business case gets worked out afterwards. That can still produce useful results, but it makes the return very hard to judge.
A Simple Way to Build the Business Case
You do not need a complicated financial model to assess an AI initiative. A one-page business case, completed before you commit, will tell you most of what you need to know.
- Define the problem precisely: Name the process, the people involved and what is going wrong. “Annual review packs take each adviser around half a day to prepare” gives you something to measure; “we want to be more efficient” does not.
- Record a baseline: Measure how long the task takes today, how often it happens and how often errors or rework occur. Without a baseline there is nothing to compare against.
- Estimate the net saving: Work out the time AI could save, then subtract the time needed to review and correct its output. Be realistic rather than optimistic.
- Count the full cost: Include licences, any implementation or configuration work, training time and the ongoing governance effort.
- Identify the risks: Note what client data the tool will touch, how outputs will be checked and who is accountable for them.
- Set a review date: Agree in advance when you will look at the results and what would count as success.
This discipline is also what separates AI projects that deliver from those that fizzle out. Our article on why so many AI projects go nowhere describes a risk-first approach that suits regulated firms particularly well.
How to Tell Whether AI Is Paying Off Six Months On
The encouraging news is that most firms are still early in their AI journey. There is time to step back and look at what has actually changed. Usage reports, which show how many people opened a tool and how often, are a weak measure on their own. Heavy use does not prove value, and light use does not prove failure.
More revealing questions focus on outcomes:
- Are people getting through their work more efficiently, and has that freed time gone into higher-value activity?
- Have turnaround times for clients improved, for example on reports, letters or responses to enquiries?
- Has the volume of errors, corrections or rework gone down, stayed the same or increased?
- Is the firm operating more smoothly than it was six months ago, or have new frustrations appeared?
- Do you know which AI tools are in use, what data they touch, and whether your policies are being followed?
If the answers are positive, you have evidence to support further investment. If they are mixed, you have a clear starting point for adjusting how AI is used, which tools you keep and where you focus next.
A Practical Checklist for Partners and Directors
If you want a quick sense of where your firm stands, these points make a useful agenda item for your next management meeting:
- List every AI tool currently paid for, including those on individual cards or expense claims, and identify any overlap.
- Check how many licences are actively used and remove or reassign those that are not.
- Confirm that you have a written AI policy covering approved tools, client data and the review of outputs.
- Pick one or two clearly defined processes where AI should make a measurable difference, and record a baseline for each.
- Make sure every client-facing or regulatory output produced with AI has a named person responsible for checking it.
- Diary a review in six months and decide now what success would look like.
Getting an Honest View of Your AI Spend
AI can be a genuinely valuable tool for professional firms. Used well, it reduces routine effort, speeds up client work and gives skilled people more time for the things only they can do. But the value does not arrive automatically. It comes from applying AI to the right problems, counting the full cost, keeping client data under control and checking the results with the same rigour you would apply to any other business investment.
If you would like help reviewing where AI is helping your firm, where it is creating extra work or risk, and whether it is delivering real value, the team at Absolutely PC would be happy to talk it through with you.
What next?
One of my passions is helping businesses to succeed and if I can help you save some money as well – even better. You can fill out our contact form, phone us or click on the appointment button below and let’s start a conversation to see if I can help your business. Our guarantee:
- There are no hidden charges – this is a 100% free 15 minute consultation with no hidden charges.
- We will never spam you or sell on your contact details.
- We will treat your information with absolute confidentiality.
Other Posts for you to Enjoy
