AI complaints are no longer just a law firm problem. They are now squarely an SRA problem too. The regulator’s own data shows misconduct reports are up 58% in just two years.
For once, the SRA and the profession are dealing with the same problem. The regulator is facing the same operational headache that firms and solicitors have been battling for months.
The SRA’s answer? More money: an eye-watering budget increase, paid for by the profession.
That is the wrong answer. If the SRA expects the profession to adapt to the AI age, it has to do the same. It cannot leave the complaints gateway wide open and then charge firms for the consequences. Yet that is where we are: the profession funds the clean-up, while firms are left to deal with the same tactics in their own complaints processes with little room to push back.
So let’s skip the cash call and get to the real fix:
1. Shut the front door
The SRA’s guidance expressly says it does not investigate everything. It looks for conduct serious enough to justify regulatory action and capable of proof. The real problem now is what hits the inbox before that judgment even starts.
That is where the SRA needs to be much firmer.
Complaints should be built around mandatory fields: who is being complained about, what is alleged, when it happened, what rule or type of misconduct is engaged, what documents support it, whether the issue was first raised with the firm, and whether there is linked litigation or a LeO complaint.
Firms already know they cannot sensibly answer every AI-generated sub-allegation as if it were a standalone point. The SRA should take the same approach: insist on structure and require clarity. And how about a maximum word count? Let’s acknowledge that sympathy for complainants must be balanced with efficiency.
2. Sharpen the triage
Not every complaint deserves the same level of regulatory attention. That should not be controversial. Allegations of dishonesty, misuse of client money, misleading the court, discrimination, exploitation of vulnerable clients or other obvious public protection issues should move fast.
By contrast, complaints about delay, tone, costs or dissatisfaction with an outcome belong in a different category, and the SRA should say so plainly. In our recent webinar on responding to an SRA audit or investigation, many experienced compliance professionals said they were not very familiar with the SRA’s Enforcement Strategy or how it guides decisions on which breaches to investigate. If the profession is unclear on the parameters, the public has little chance. The SRA has some obvious educating to do.
3. Not my job, guv
A good deal of regulatory time is probably being spent on complaints that never belonged with the SRA in the first place.
The sympathetic view is that the public may not naturally distinguish between the regulator, a firm’s internal complaints process and LeO. That is sometimes true, but not always. Firms know that some clients will throw everything they can at a dispute: an internal complaint, a letter before claim, a LeO referral and, for good measure, a misconduct report to the SRA.
The SRA should tackle that head-on. Its own guidance says it is concerned with serious or repeated conduct issues. LeO exists to deal with poor service. So why not say that much more bluntly at the front end and filter out complaints that fall outside its remit?
- Add a short eligibility checker before the complaint form
- Use plain-language examples
- Say clearly that complaints about delay, costs, poor communication or general service dissatisfaction are not for the SRA, while complaints about dishonesty, misuse of money, discrimination or serious conduct failings are
- Ask whether there are parallel complaints with the firm, LeO or the courts and, if so, consider whether SRA triage should wait until those processes finish
That kind of clarity would save time for everyone, including complainants. It would also set expectations from the off. If a complainant can be told, kindly but firmly, that the SRA is simply not the right place, they can at least understand that this avenue isn’t worth pursuing.
4. Stop rolling out the red carpet for vexatious complaints
This is the part that will make some people uncomfortable, but it is hard to avoid. AI makes it easy to produce long, plausible complaints packed with headings, accusations and legal jargon. That does not mean they are false. But it does mean polished presentation is no guarantee of substance.
Like firms, the SRA should accept that traditional complaints handling does not work as well in the age of AI-generated material. The answer is to impose firmer front-end requirements. For example:
- Consider a modest filing fee, waived in hardship cases and refunded if the complaint is upheld
- Impose a maximum word count for the main narrative
- Require a statement of truth
- Require complainants to identify the key documents they rely on
- Set strict response deadlines unless there is a good reason not to
- Set clear rules for how complainants deal with investigators and take a zero-tolerance approach to abuse
- Require an explicit AI declaration confirming that any AI-assisted draft has been checked personally and is not being submitted as an unverified machine account
For measures like these to work, they need teeth. Miss a response deadline without good reason? Close the complaint. Fail to provide clarity or evidence after being asked? Close it. Reopen only in exceptional cases.
I would go further. Complainants should have to say what they actually want: compensation, an apology, an explanation or regulatory action.
Harsh? No. Anyone who runs any sort of organisation will recognise this as basic operational discipline.
None of this means shutting the door on genuine complainants
A firmer process must still be a fair one. The SRA would need sensible exceptions for vulnerability, disability, language barriers and urgent public protection concerns. Any fee would need a real waiver mechanism. Any word limit would need room for annexes and evidence. And nobody should be excluded simply because they are distressed or inarticulate. A better process can still be an accessible one.
The point, however, is that the better operational discipline will pay dividends. It will allow the SRA more time to deal with those complainants who might need reasonable adjustments and more time to deal with the complaints that they really need to focus on.
What does that mean for us?
Law firms are already dealing with AI-generated complaints: longer, more repetitive and often harder to pin down or resolve. Many firms are rethinking their approach, but they can only go so far alone.
As compliance people like to say, it starts with tone from the top. If the SRA gets firmer, firms will feel able to do the same. We may not copy every measure, but we can be clearer with complainants and less willing to indulge sprawling, incoherent or tactical complaints. We can move faster without being less thorough.
And if the SRA runs a tighter ship, the profession gets something back: a leaner, sharper regulator.
More importantly, it would help the SRA cut through the noise and focus on the complaints that really matter. Better triage and better use of resources mean better protection for the public. That, after all, the reason the SRA exists.
