It’s time.
It’s sitting across from someone without a screen between you and actually listening.
It’s knowing their history because you’ve supervised them for months, not just skimmed a case note five minutes before the appointment.
It’s noticing when something feels off before it shows up on a risk tool.
It’s being able to challenge them honestly because there’s enough trust for them not to shut down or walk out.
It’s the conversations that don’t fit neatly into drop-down boxes. The bits that never make it onto a dashboard but are exactly where change happens.
It’s also professional judgement. Making decisions based on experience and instinct as well as policy. Having the space to think, not just process.
When I started, engagement meant relationships. Now it too often means logging a contact, updating different systems and proving you spoke to someone for 30 minutes.
That’s not engagement. That’s admin with a pulse.
Real engagement takes continuity, time and trust. Three things the current model systematically squeezes out.
So when management talk about “meaningful engagement”, most of us just hear another slogan for “do more with less”.
Because you can’t build trust at speed, and you certainly can’t automate it.
Anon
"This is so true...sadly the organisation has prioritised other things over this. Over the past ten or so years the organisation has focused on how to complete OASYS, audits of these, data collation, policy dictats, written case records, audits of these records, how things are recorded and measured....meanwhile nothing is really said about how to actually do the job...how to use relationships to good effect and the research which supports this or the skills and professionalism it requires to do it well.
ReplyDeleteAnd what makes me really sick is when finally addressing workloads what was the first thing that was cut? Relationships. RESET and impact pretty much disposed of the relationship, flippantly seeing this as the expendable element. Nothing else was changed...the ridiculous amount of time spent feeding the OASYS machine being the obvious one....huge amounts of money ploughed into data systems like MPOP and the recall portal.
Why was there not a mutiny when they literally cut what matters, left bureaucracy untouched and we all waved this in without blinking an eye."
Helping support the work we have is one thing, but doing ourselves out of a job with Justice Transcribed (Justice denied?) It will soon go the way of the Home Office FNO obligation of going to an anonymous office, with a security guard from G4S, and punching in a code or a piece of cardboard or plastic to show that you've attended. The powers that be should or probably do look at this website, but the heads are in the sand and they'd rather come up with ideas that supposedly stave off being redeployed and give them the integrity that has been completely wiped out from probation as a job, as a profession and a purpose. It's all about prisons because the majority want more put there. But we don't have capacity, but the prisons and national guidelines favour them to the point of not being able to challenge it as community probation. This adversarial, inequitable dynamic is something that needs to be addressed. I'd rather not work in a prison where my office is a makeshift cell and i'm subject to the same rules around mobile phones and where I'm able to instruct, tell or ignore with surety any issues community probation may have. This dynamic is another reason why we're a dumping ground not a service that is valued for our experience and risk assessing- usually far more acute and insightful than the prison's. If you don't have money or don't want to spend it, then at least change the guidelines- that cost relatively little money.
ReplyDeleteThe managed demise of the ‘relationship’ was part of the move towards us segueing into a correctional service in which rules are rigid and relationships are the ‘do as I say’ variety rather than trying to find a collaborative approach to a problem. Justice transcribed is a tool for the SFO inspectors not for officers and penalties for not using will surely follow……..how to do the job is not covered in the training and is part of what you acquire along the way, as is us stop pretending that we know everything and we are not always right! The surpressed glee when a difficult case has been recalled is always a failure , yet I now see managers with little experience applauding this!
ReplyDeleteThe relationship was what separates us from Senior Managers with little experience in the front line and what gave us a unique insight into how and why a criminal act took place…….to first dilute, then sideline, before discarding this relationship is evidence that those decision makes at the top do not understand how, what or why probation is…..
https://www.bbc.co.uk/news/articles/c9wx2dz2v44o
ReplyDelete"The onslaught of AI slop - which he defines as fake, unconvincing videos and pictures, made quickly - is now unstoppable. Tech companies have embraced AI."
asked ai to iffer a view on accuracy of ai transcription:
Delete"AI transcription typically achieves 90-95% accuracy in ideal, clear audio conditions, but often
drops to 80% or below in noisy, real-world scenarios with accents, jargon, or multiple speakers. While top AI tools are highly effective for quick drafts and indexing, they rarely match human transcription's 99% accuracy for legal or critical needs"
Has the cobbler or any of the cheerleaders for ai tools discussed if ai transcripts are admissible in legal proceedings? Are the texts in probation systems identified as ai transcribed? If not, that surely has to be problematic?
https://www.farrer.co.uk/news-and-insights/a-critical-eye-on-ai-evidence/
"The occasionally dubious accuracy of AI-generated transcripts creates an evidential difficulty for the courts – to what extent can contemporaneous AI-generated transcripts be assumed to be accurate?
if a separate recording of a conversation or meeting has been made and retained, then the AI-generated transcript can be checked for accuracy against that.
The best way to counter that risk is for parties to check and correct AI-generated transcripts in real time, and to make a record showing that this has been done. Not only is that good practice which ensures that written records are as accurate as possible, but it could also help convince a court later down the line that the transcript is accurate and should be relied upon as good evidence. These checks are particularly important where a transcript records the more important conversations that have taken place."
Can't see evidence of time-saving here. On the contrary, it doubles the time & effort.
Once again, well done hmpps!
consider an interview room at a probation office:
Delete"80% or below in noisy, real-world scenarios with accents, jargon, or multiple speakers."
multiplicity of regional accents
slang terminology
probation jargon
people talking over each other
mumbling
arguments
language translator - or is ai going to do that too?
After parties have checked and corrected AI-generated transcripts in real time, (does that include the interviewee?), do the police sweep up the transcripts as evidence?
Will interviewees, already suspicious of probation, simply refuse to engage? Or lie because they fear their words will be twisted, edited & used out of context to prosecute further offences? Or contribute to children being removed? Or lead to eviction?
Does ai have to swear an oath? Or will its edits/hallucinations be regarded as admissible regardless?
What does this mean for miscarriages of justice?
No mention ai will not need interpreters ai does all the languages needed today.
DeleteIf I were on probation and you sat tapping on their laptop or recording and transcribing me on their phone, I’d be saying very little. Organisations have an obsession with recording everything nowadays. Any meeting I goto and it’s asked if ok to record I say “no, it is not ok”.
ReplyDeleteWhich poses an interesting dilemma re-consent... time was when those with potential for probation supervision (all formats) had to give consent before being sentenced, until Michael Howard's Crime (Sentences) Act 1997 did away with all that 'woke' nonsense:
Delete"Removal of General Consent: The Act, with provisions taking effect around late 1997/early 1998, stipulated that consent was no longer required for a community order to be imposed, except where the order included a requirement for treatment for mental condition.
Previous Law: Before this change, an offender's consent was generally required for the court to pass a community sentence.
Breach and Resentencing: While consent to the order itself was removed, failure to comply with the terms of a community order (a breach) allowed the court to revoke the order and re-sentence the offender, which could result in a custodial sentence.
Psychiatric Treatment: Consent remained necessary if the community order imposed a requirement for treatment for a mental condition."
No doubt lammy et al will bring in some new law or amendment that makes it compulsory to have probation interviews/sessions recorded, stored & available on the hmpps systems (ai transcript, video) and that refusing to co-operate means jailtime.
Every “innovation” seems to move us further away from professional work and closer to a kiosk model. Tag them. Scan them. Log them. Record them. Prove contact happened. Move on.
ReplyDeleteThat isn’t supervision. That’s attendance monitoring.
It might tick a ministerial box, but it does nothing to build trust, challenge behaviour or support change. It just makes probation look like a cheaper, outsourced extension of custody.
We keep being told this is modernisation. It looks a lot more like downgrading.
The AI point worries me for exactly the reasons people are raising.
ReplyDeleteIf it’s 80–90% accurate in real world conditions, that’s not a small margin of error when you’re recording risk disclosures, safeguarding information or potential breaches. That’s the difference between “I did” and “I didn’t”.
So now staff either blindly trust a flawed transcript or spend extra time checking and correcting it line by line. Which means no time saved at all.
We’re being sold this as workload reduction. It looks more like surveillance plus extra admin.
And it absolutely changes the dynamic in the room. If someone feels recorded and monitored, they talk less, not more. That kills “meaningful engagement” before it even starts.
From what i understand Justice Transcribe looks more like AI-assisted dictation than some kind of permanent surveillance. Staff review and edit the transcript before anything goes into NDelius, so liability for the final note still sits with us, exactly as it always has.
ReplyDeleteAnd there are obvious benefits. For accuracy, disputes or breaches, having a clear transcript could actually protect staff and cut down the usual “you said / I said” arguments. Saving hours of typing is no bad thing either.
But let’s not pretend there aren’t downsides.
Recording changes behaviour. People talk differently when they know they’re being recorded. They become more guarded, less open, and less likely to admit risky or embarrassing things. In probation, those off-the-cuff disclosures like “I nearly relapsed”, “I’ve been in contact with my ex”, “I messed up but panicked” are often exactly what help us manage risk early. If sessions start to feel like they’re being formally recorded or could be replayed in court, we may actually get fewer honest conversations, not more.
That’s not theory it’s just human nature.
If it helps with admin, great. But it’s reasonable to ask some basic questions about how it’s used and what’s kept:
Where’s the DPIA?
How long is raw audio stored?
Is it deleted after drafting or retained?
Do staff control what becomes the official record?
Because “AI helps write notes” is one thing. Routine recording of supervision conversations is another.
Im not resistant to change. I just want transparency and safeguards and an honest conversation about the fact that recording might improve accuracy, but it can also reduce candour and trust, which probation work depends on.
If we’re recording supervision sessions at scale, a DPIA (the legally required privacy and risk assessment) must already exist. It should be easy for staff to see. If it isn’t, that’s a problem in itself.
ReplyDeleteIf engagement is not meaningful, then the thousands of people being returned to custody on recall each year for non engagement is pointless.
ReplyDeleteI completely agree with annon 09:51. Probation (whatever format) should be by consent.
When it's enforced it becomes punitive. It creates an imbalance in the relationship. Any relationship where such an imbalance exists (you will do versus I don't want to do) will always be toxic by its very nature.
You can't build a relationship if its based on on one side demanding compliance from the other.
That type of relationship is really not any different from the JFDI relationship people have with management and constantly complain about on this blog.
If it dosent work for the goose, why should it work for the gander?
'Getafix
Probation was built on relationships and consent.
ReplyDeleteThe original model, going right back to John Augustus and later the Probation of Offenders Act 1907, was literally “advise, assist and befriend.” That wasn’t just a slogan. It was the job. People were supervised through trust, support and influence, not threats, breach and recall.
Yes, there was court authority behind it, but day-to-day practice relied far more on cooperation than coercion. The relationship was the intervention.
What we have now is almost the opposite. Surveillance, enforcement, targets, recalls and compliance dressed up as “engagement.” We talk about relationships while structuring the work in a way that actively undermines them.
So when people say probation should be closer to voluntary or consent-based, they’re not being unrealistic. That’s actually what it was designed to be.
Somewhere along the line we stopped asking how to help people change and started asking how to record that we told them to.
And we wonder why “meaningful engagement” feels impossible.
Good history well placed but it cannot exist in today's society not a chance. The variables are too wide and multiculturalism .
DeleteThe NAO report basically confirms what staff already know but management won’t say out loud.
ReplyDeletePerformance down.
Staffing short.
Workloads over capacity.
Risk work getting weaker.
Supervision cut back just to cope.
And Impact and Reset haven’t even been properly tested to see what damage they cause!
So let’s call it what it is because it isn't This innovation and it isn't transformation.
It’s what happens when you can’t afford enough staff, so you quietly lower the service instead.
Reduce contact. Script it. Move people through faster. Call it “meaningful engagement” while replacing relationships with templates and AI transcripts.
Justice Transcribe, Impact, Reset… none of it is about quality. It’s about managing shortages. It’s cheaper to record conversations than build trust. Cheaper to automate notes than retain experienced officers. Cheaper to see someone four times and discharge them than actually supervise properly.
That’s not modernisation. It’s rationing.
And then, after hollowing out the job and calling it progress, they offer 4% and expect gratitude.
If this is the “future of probation”, it isn’t reform. It’s shrinkflation.
Same risk. Less time. Less skill. Less pay.
Good post and thanks...I completely agree.
DeleteHas the probation caseload increased to such extenuating levels that rationing is the only choice?
According to our friend AI, not really:
"2015: Approximately 228,844 offenders were supervised by the Probation Service as of 30 June 2015.
2025 (Current): Approximately 244,209 offenders were under probation supervision as of 30 June 2025. "
This then begs the question about why the probation service is currently considered unsustainable? I'm guessing it was massive increases in expectations caused by the "risk management" culture which largely contributed to this..not "over supervision"....they have rationed the service without even attempting to improve service delivery, while leaving bureaucracy untouched....it beggars belief that our beleaguered work force waved this rationing and de professionalisation in as we were too knackered to notice.
https://prisonreformtrust.org.uk/wp-content/uploads/2025/02/Justice-Committee-rehabilitation-PRT-written-evidence-FINAL.pdf
ReplyDeleteGood evening everyone heartfelt thanks for all your contribution's. Thank you 14.38 for your succinct and powerful comments which given the contents over the past 5 months, let alone the preceding 10/11 years will resonate with many. I cant help think ‘dare we believe’ that someone is listening. I do so hope that when the time comes for the Cycles of Re-offending that some of you are able to contribute to that Enquiry on Probation. I will try and track when that might be and certainly let JB know if Im able to find out. Wishing you all a good evening. I valued reading 14.38 and hope that many many more will read this blog iangould5
ReplyDeleteIts ok... jones has spoken:
ReplyDeletehttps://hmiprobation.justiceinspectorates.gov.uk/news/sentencing-act-january-2026/
Published:
23 January 2026
I was pleased to see the Sentencing Act receive royal asset yesterday evening, together with the government’s reassurance that the reforms are supported by plans to rebuild the Probation Service – including investment of up to £700m over the next three years, increased ability to tag offenders, and new technology aimed at reducing administration, so staff can focus on work that reduces reoffending.
It is positive that the Act will not take effect immediately, giving the Probation Service time to prepare for the incoming changes. However, we know that some measures, including the extension of Suspended Sentence Orders and changes to Remand, will come into force in just two months, while implementation planning continues for more complex changes.
I have spoken recently about my support in principle for many elements of the Act and the opportunity it provides to transform the justice system.
However, I have been clear that there will be challenges in ensuring the gap between probation resource and the requirements of the Act are reconciled, and that difficult choices will need to be made around what to prioritise for maximum impact.
I have also warned that there is a danger of the reforms collapsing public confidence in probation unless they are implemented with great care and thought, and the right investment is made in the service in the short to medium term. Failing to address these point risks setting the service up to fail.
I look forward to continuing to work closely with the Department throughout 2026 to advise on how our inspection findings can inform next steps. In particular, I will be sharing the results of our Dynamic Inspection of Public Protection programme, which publishes its first report next week (29 January), with a focus on what improved probation practice can do to make the public safer and reduce harm to victims.
Erm, but the 29 Jan report doesn't appear to be there after all.
DeleteAhhh, I see, its the KSS inspection:
Deletehttps://hmiprobation.justiceinspectorates.gov.uk/document/dynamic-inspection-of-public-protection-in-kent-surrey-and-sussex-2026/
"Difficult choices will need to be made"....scrap OASYS and bring in a one page assessment...scrap data measures, NSI, Delius data....scrap Crissa recording....scrap all those bloody risk register reviews (we never had them before Delius so we can surely do without now)....scrap about 2000 pages of policies that have been brought in over the past 10 years.
DeleteCan't wait to hear what difficult choices will be made....