Neil and Kieron cover a busy week in public sector AI.
Neil demoed KnowledgeFlow to occupational health teams from the North of England, and a GP’s feedback showed where AI actually helps: simple, upfront answers to questions like “how do I get my forklift driver with a bad back back to work?”
Kieron explains the next stage of Leading AI’s strategy. That means moving from AI assistants you copy and paste into, to AI-enabled apps that run the whole workflow. He starts with the Education Health and Care Plan (EHCP) consultation tool for colleges: one tester ran 70 cases through it in under two weeks, all with a human in the loop at each stage.
And Mrs Watkins asks “If AI does the work, what happens to the team?”. A very good question indeed.
Kieron and Neil also discuss:
- The human-in-the-loop problem: if AI checks 500 invoices in half an hour, how does a person check the AI?
- AI in housing complaints, as Awaab’s Law moves to its next stage. The aim is a framework for the housing officer’s judgement, not a generic answer.
- End-to-end AI procurement, and why suppliers with their own agents will never miss a deadline
- New UK research on AI tutoring: the same exam results as human tutors, but 30% more of the curriculum covered
- An AI job interviewer meets a candidate with a basketball boot on his head
Next week: A very special guest, Catherine Ritchie, Head of AI Strategy and Innovation at North Yorkshire Council.
Two mates. A bar. Thirty years of business between them. And all they want to talk about is AI.
Pull up a stool, we’ll get the beers in. 🍺
TRANSCRIPT
Podcasting – #39
Neil Watkins
Cool, let’s get going then, Kieron. Happy Friday. You’re looking very smart today. Have you been on an interview or something?
Kieron White
That’s right, yeah. Always Johnny. I think I’m unemployable, so I don’t suspect I’m ever going to go to an interview. I reckon anyone, anyone who would go, what math are you applying for? No, I put a jacket on for our clients that we met just now and for the podcast listener, so that, so that they
Neil Watkins
A.
Haha.
I really.
Kieron White
can admire my slightly smarter look today. How are you doing? What have you got on your list for us to talk about today?
Neil Watkins
Excellent, excellent.
Well, I’ve got a few things, really. One, it’s been a busy old week. I did a demo in the health service, which I will tell you a little bit about what I was doing there earlier this week. I wanted to talk a little bit about housing because I’ve been looking at some of the impact stories. And of course, AWAB’s law starts on
Kieron White
Nice.
Neil Watkins
Those 2 comes into effect on the 30th of November. So that’s coming around for housing and a conversation with somebody about some of that stuff, which was really interesting. And then really the whole how things are changing. And we met on Tuesday with Big Date and looked at his latest stuff.
Kieron White
Yes.
Yeah.
Neil Watkins
And while this isn’t a product pitch, I had a kind of a realisation on Tuesday about how just how powerful this stuff could be for us in a way that I kind of have waxed lyrical to a few people about already this week. So yeah,
really exciting stuff to talk about. What about you? What have you got on your list?
Kieron White
Very good. Well, I have similarly want to talk about our feedback from the strategy session we had with Big D. That’d be great to kick around and to, yeah, share, I guess, summarise it and think it through in our own in our conversation. I want to talk a little bit about some research on AI tutoring.
Neil Watkins
Ios.
Yeah.
Kieron White
which is quite interesting. So that was quite, yeah, quite good. And I’ve just got one other thing and a bit of fun, of course. I wanted to talk, reflect on Mrs. Watkins AI fears that we, yeah, we had that conversation because I stayed, as you know, I stayed at your establishment.
Neil Watkins
OK.
Ohh, no.
Kieron White
on Tuesday night and enjoyed having an evening beverage with you and Mrs Watkins. And yeah, I was quite struck by her reaction to some things, which is not in the usual way. So I’d like to talk about that.
Neil Watkins
Nate.
Okay, interesting. All right. Right, where should we start then? Should we start? Let’s start with my NHS story, because that was really interesting. So for quite a while now, I’ve been doing some work with some people around occupational health. And lots of people think occupational health is really boring, but actually it’s such an important topic because
Kieron White
Yeah.
Yes.
Neil Watkins
In the country, we lose, you know, millions, I’m guessing, of days per year to people off sick and, you know, people lifting stuff incorrectly, bad backs and, you know, how do you get people back into work when they’ve been taking time off and, you know, how this all kind of challenge, and especially for employers who have got a duty of care to their staff and all of that good stuff. So how
How does the how does the system support occupational health? Anyway, long story, I was asked if I would share knowledge flow with a bunch of people from the northeast and the northwest of England. And I said, yeah, yeah, no problem. And I thought it was going to be like three or four people.
The invite went out to like 20. So, and I think it was 16 on the course. It was a bit, it was like, whoa, here we go. And what was really interesting, there was a guy on from Sri Lanka who had, bizarrely, he knew somebody and he in
Kieron White
Ha.
Neil Watkins
in the UK working on some of the occupational health stuff. And he created a workflow to help work through the kind of occupational questions that might need, occupational questions that someone might do. And I was like, it was really interesting because he’d clearly done some really
good stuff, but he’d done it in a similar way to us, which was reassuring. You know, he was like, this is built in Azure and it’s built on these tools and, you know, we’re using Chatgpt and it’s only in Microsoft UK South because of GDPR and all that. So actually, he’s a smart guy, lovely chap, and he’d done some really good stuff.
So, he, he, it’s just a bit like Sean Tel, he showed his and I was like, well, just to be clear, I think he’s done a great job, well done.
Kieron White
It must have been a terrifying beginning when someone’s demonstrating another tool, you’re like, oh no.
Neil Watkins
Well, not really actually, because I could see, because immediately it was clear that they had gone in depth and the bits that I wanted to show were the stuff that we’ve done for North East Links on cancer screening. We’ve downloaded the Health and Safety Executive,
Kieron White
Mmh.
Neil Watkins
publicly available PDFs. And we’ve done the same with CQC. And I also showed complaints handling, because complaints is massive in health. So what was really interesting was one of the GPs said the stuff that the other guy showed,
Kieron White
IAM.
Yeah, it is.
Neil Watkins
He said, I could see that, but there’s a big problem beforehand. And it’s the upfront, simple communication, simple guide. What you’ve got is a complex case handling system thing here. And what KnowledgeFlow does beautifully is, I want my forklift driver has got a bad back. How do I get him back into work?
you know, someone’s slipped in the office, what do I need to do? And those kinds of questions are all in the health and safety executive. I showed things from the CQC, like my GP practise is going to be inspected, what do I need to do? And it goes through and it tells you. So they could see there was, and
Kieron White
Nice.
Neil Watkins
and I posted in a complaint and obviously it produced a great response to the complaint letter, including the kind of, you know, you’ve made this up in the nicest possible way, you’ve made this piece of legislation, it doesn’t exist. So that was really helpful because what it was complementary. So it wasn’t scary. It was literally complementary of
Kieron White
Yeah.
Nice.
Neil Watkins
ours does this thing super well. And at the end, the guy said, listen, he said, I’m no, we’re nowhere near where your platform is. It’s far more sophisticated than we are. And so, you know, if there’s a way of us potentially supporting these guys over here by linking into KnowledgeFlow, we’d be delighted to just give it to you and you can
crack on, you know, a bit like ours, he was doing it for free to support them in a way that we’ve done it for some other people in order to get them going. So, no, it was really, but the feedback was super interesting and it made me realise that many people still don’t really understand AI.
Kieron White
Yeah.
Neil Watkins
And even though there were people on the call who formed an AI group on for occupational health on Linkedin, they just, they just so, they just, they so find that kind of far behind where we are. And I wonder, you know, whether even some of the things that we talk about on here, some of the things we’re going to talk about in a minute, you know, that’s just so far ahead of most people’s general kind of
Kieron White
Hmm.
Neil Watkins
dare I say it, co-pilot experience, that it was it was quite humbling to get the feedback and then just reflect on that about how far how far we’ve come in the last three years, Kieron.
Kieron White
Yeah, yeah.
Yeah, no, it’s amazing when you think where we where we started all those time ago. But here’s an interesting piece, and it’s a segue into our Tuesday strategy session, but is I I’ve had three moments, I think, with Leading AI where I’ve genuinely hugely excited. The first was when Donald
Invented Rag, or at least I’m sure he didn’t invent Rag, but he we he stumbled into Rag before we knew what Rag was, and I mate.
Neil Watkins
Yeah.
before anybody in the country knew what it was.
Kieron White
Yeah, and we were the first people doing RAG on, you know, like the policy document, policy buddy sort of thing, which is really good. Now, that was hugely exciting and we, you know, created a load of noise and hopefully helped a lot of people be better in their work and improve outcomes for citizens as a result. The second big moment was just over a year ago when we created KnowledgeFlow.
which was originally conceived as to bringing, because what happened was people wanted the rag and then they wanted another rag for something and another one and another one. And then they wanted to upload their documents and then they wanted this, that and the other. So KnowledgeFlow was born to be the kind of platform that could bring it all together in the harness and governance layer and security.
And my next biggest excitement is our latest evolution into what we’re calling apps, applications that are AI enabled, and which really does a disservice to them. But it’s absolutely tremendous. And I want to briefly talk about the EHCP tool. So I’ve touched on it on here before, but we’ve
Neil Watkins
I think it does.
That.
Kieron White
created. So if you are a college, most of our customers in education are colleges. If you’re a college, you will have to respond to a consultation, an EHCP consultation, and you’ll get many of those per year. The lowest number I’ve heard about is 600 a year, and the highest I’ve so far heard about is 1,200.
Each one of those requires about two hours minimum work just to read it, comprehend it and work out. And what’s going on there is a local authority is putting together a consultation letter and an EHCP and saying to a college, this is Minnie Mouse. We would love you to take Minnie Mouse and as a student,
Let us know if you can. And the consultation response is, firstly, can we meet their needs? And there’ll be quite often many, many special needs. And can we, and is there any additional cost associated with that? You know, if it’s like a full-time support worker with you at all times, etc.
Neil Watkins
Yeah.
Yeah.
Kieron White
and they have to work all that out and then write back to it within 15 days. So you’ve got, it’s all a bit of a challenge when you’re getting 1000 of these and they don’t all come at once, but you get lots. And so we,
Neil Watkins
I heard of a I heard of a college that yesterday that had to do 60, and they got 60 yesterday, just like, wow.
Kieron White
Yeah, yeah. Well, the test person, get this, this is the best thing. So we built a first version of this in KnowledgeFlow, which is what we call a step runner prompt. So it’s, I’m not going to bore anybody with what that means, but it’s a way of getting Gen AI to be diligent. So we gave it to one of our colleges to test.
Get this, she has run 70 through it. That is how useful, that’s how useful, that’s in just under two weeks. So she’s had it. And yeah, I mean, normally people test, don’t they? They do one or two, maybe five and go, right, there’s all this feedback I need changing. And she’s, we’ve got some feedback, of course, which is great. But yeah, 70 have gone through and it’s saving app.
Neil Watkins
What happened?
Wow.
Yeah.
Kieron White
absolutely aeons of time, very accurate. So that’s great. Where it gets exciting, though, is to take that away from KnowledgeFlow into an app. And what we’ve built is a case management system. So you can now upload your EHCP. That becomes permanently in your
Neil Watkins
Wow.
Kieron White
Case log now, until you decide you know, obviously you want to purge the data. The AI processing runs in the background, so that when you, so you can have all of your EHCPs looking in front of you in this case system, they’ll track the dates they came in, and when the response is due, it gives you an immediate, when you click on anyone, you get an immediate.
Neil Watkins
Yeah.
Yep.
Kieron White
Summary of what it is and whether you can meet the needs or not and the likely costs. That’s immediate. And then you go into the consultation response where the AI has been through and looked at all of the sections and given you the summaries, checked it against your own college’s provision map, as we call it, the things you can provide, things
Neil Watkins
Mm.
Kieron White
you can’t, cheques it against the law, cheques it against all the kind of rules that you want to introduce, and then writes a letter for the local authority, yes or no. All of it is human in the loop. You cannot make it write the letter until you have ticked that you have reviewed each section. So it’s really, really lovely. And then it’s got a sign off process where it will send it
Neil Watkins
Yeah.
Kieron White
to a signer offeror, an approved signatory. There you go. An approved signatory in your organisation who then can do the final sign up. That’s an option, don’t have to have that, but that is a nice belt and braces. All of it is tracked with evidence file. All of it exists in the case system as an audit trail.
Neil Watkins
That was a new title.
Kieron White
So when you come back to it in six months, it’s still there. It will also produce learner support plans or success plans, which is really for the teachers and teaching assistants to give them a one page or very short summary of the needs of the learner and the teaching strategies that you should employ. And those teaching strategies can come from the college’s own versions if they have them,
Neil Watkins
Yeah.
Kieron White
or we load in the background the Education Endowment Foundation’s research-backed evidence of what works for particular special needs. So you’ve got those produced. And the final part is it will also do an annual review, which is the process of reloading the EHCP, uploading any new evidence,
Neil Watkins
Then.
Hmm.
Kieron White
progress reports, whatever else, and then making it look through there to give you your quick summary. So it’s taking out a load of what is pure admin in a really important area, and we’re super excited about it.
Neil Watkins
Ai.
Yeah, no, that is, I mean, it does sound brilliant. I pitched it yesterday to somebody who traditionally doesn’t get very excited about anything at all, but even they could see, they could see the benefit of it. So that was fantastic. The other thing going through my mind is we should, I’ll contact Sue Balagodu because she’ll be delighted that we’ve listened to her about Human in the Loop and the
In the process that we go through, and there’s a signer offerer at each at each stage.
Kieron White
Yeah, absolutely.
Exactly. There is. I interviewed people I was interviewing yesterday and for a client of ours helping them out ones of AI side of it. And interestingly, there was a very knowledgeable chap we were interviewing. He knows lots about AI. And I asked him about, because he talked about human in the loop, the important thing. And I
And I asked him, I said, this is a really difficult question and I’m not actually expecting you to have an answer, but how do you manage the human in the loop problem where effectively the AI has done all of the, checking of the invoices was one of the things they did, cheque all the invoices and then write a summary e-mail and they do 500 of them in half an hour, what used to be 6 people for a week.
but then the human has to cheque. And I’m like, the real challenge is there, unless you open the invoice and open everything else, you can’t do it, can you? And yeah, so we had, it was quite an interesting discussion about how do you do it. And he was like, yeah, you’re right, it is a very difficult one. But we’ve got that somewhat cracked in this EHCP, very important area that the human judgement is there.
Neil Watkins
Mm.
Kieron White
Is it enough? Probably. Well, I hope so. I hope we’ve got close enough, but yeah, we’ll keep looking and seeing how we can improve. But the whole…
Neil Watkins
Mm.
The the human, sorry, go ahead, Kieron.
Kieron White
No, go on. I was going to talk about a bit about our strategic direction, but…
Neil Watkins
Well, I was going to talk, I was just going to touch on complaints because I was looking at some of the success stories of AI in complaints handling in Housing Associations. And indeed, I posted it on Linkedin just now, but it was really about that whole kind of
helping people make judgments within a framework and giving them the information, you know, pulling in the previous complaints letters, the history of repairs, you know, other information that might be relevant and
germane to the complaint. And how do you stop the doom loop of, you know, complaint, fix complaint and it going around in circles. And it just struck me that it is, it’s, we’re not giving generic answers because a generic AI doesn’t know the context of the organisation.
Kieron White
Mm.
Neil Watkins
What we’re giving them is a framework in which to make judgments. You need to be aware of, we, you need to be aware of all of these things. And if you take them into account, this is the suggested response, but you need to, or housing officer, you’re the hero in this story, you need to make the decision and you need to produce, you know, use your skill and judgement in order to make the correct decision and help the tenant. So
Kieron White
Yeah, it’s nice.
Neil Watkins
Similar to the EHCP things in some ways, that fits that kind of judgement, a framework in which to make judgments was the way that I kind of was thinking about it. And that goes back to my Every Child Matters days, and we did the outcomes framework, if you remember that piece of paper that we printed off. And I remember,
Kieron White
Knives.
I do. Yes, I do.
Neil Watkins
This is for Kevin Yong. When Kevin was asked to produce 14,000 copies of it, and he’s like, the DFA won’t like us printing that many off on the 2nd floor. So here’s the PDF, you can do it yourself. Back to you, back to you, you’re going to talk about strategy.
Kieron White
Haha.
Yeah.
Spending a long time by the printer that day, I imagine.
Neil Watkins
He was, yeah, he was fed up with it.
Kieron White
No, I said, so I was just going to share a little bit just to kind of put it on record for our benefit, as much as anything else, is getting to the new strategy that we kind of have been talking about playing with and Tuesday was our kind of right, we’re definitely doing this. And my excitement is
So KnowledgeFlow is brilliant as an evolution of what were individual rags, rag assistants into a platform that is now super capable, has contextual stuff, memory in it, has history and all the other kind of tools and things that it does amazingly. And then the challenge of KnowledgeFlow is still this copy and paste thing.
which is an AI problem, isn’t it? Is that we all have, and you know, and that’s where, if I’m honest, the co-pilot bit, I’m slightly jealous of because it exists inside all of your Microsoft stuff, it can interact with your apps directly. And that is a challenge because nobody wants a world of copying and pasting for the rest of their life. So this move to an app-based model, which is more traditional tech, the mix of
Neil Watkins
Yeah.
Kieron White
what is really just a database system with AI enablement in the back using the KnowledgeFlow platform for all of its security, its incredible accuracy, its ability to handle unbelievable amounts of data. So the whole of these apps are built on exactly the same platform.
Neil Watkins
Yeah.
Kieron White
but exist in a more traditional way and hold the information and manage the entire workflow. So your full end to end. I’m super excited about it. And it gives us a much clearer focus. We have in colleges, we have 23 different AI agents that we’ve created because we love to work with colleges on what is the thing they need.
Neil Watkins
Yeah.
Kieron White
And if the thing they need is something all colleges need, then we will build it with them and involve other colleges too, and put it on KnowledgeFlow to colleges for free. It’s not, we don’t even charge extra for them. We’re delighted to be kind of in the sector helping shape, you know, the sector helping itself really.
Neil Watkins
Yeah.
Kieron White
So, but this, the 23 is a problem because training, it’s difficult, having usage tracking on number of users and prompts when actually they’re really specialist tools. And you, I mean, the EHCP is probably only going to be used by four or five people in any college. So there’s lots of things that make noise very slightly problematic.
Neil Watkins
Yeah.
Kieron White
in the success story. And when you come to reflect on a renewal, if you’re a college or housing, and you see that, you know, usage isn’t that massive, it’s a lot of copying and pasting, it appears I could do some of that in co-pilot, which it appears isn’t necessarily true. So the app based
future is, I think, resolves a whole bunch of that and is so much more useful to the humans, to the workforce, and ultimately to the students and tenants that will benefit from it. It’s tremendous, really very exciting.
Neil Watkins
Yeah.
Yeah.
Yeah. The other thing, I mean, we’ve been talking about it, but we’ve never been able to crystallise it before Tuesday, I don’t think. But that is that whole kind of evolution of indexing, RAG indexing, RAG, and then how do you bring those together in KnowledgeFlow, and then how do you overlay the workflow?
Kieron White
Mm.
Neil Watkins
application on top. And the bit that I got super excited about was the whole procurement piece for education and for any particular organisation doing procurement. But that whole kind of, you know, inquiry comes into the inbox, it’s triaged into the relevant group, they get it, you track how long it’s been sat there and you know what the responses are.
in many instances, if it’s something relatively small or more information is required, the AI goes out and does it itself. You know, you kind of, I need 5 laptops, great. What do you need? You know, what size, you know, what colour, what make, what are generally preferences. And then being able to go out to the suppliers and get quotes and then bring those back in and give them to the customer.
all of that stuff. And so we’ve been working on that simple thing. But then the next stage around mini competitions where you’ve got an ITT to create and, you know, we’ve got a library of hundreds of successful ITTs that we can draw on, you know, what makes a good one and how do you incorporate that? How do we help this particular time?
Kieron White
Mm.
Neil Watkins
other type of the similar schools have done similar things. So you can just see that whole kind of end to end workflow of inquiry coming in the door right the way through to a successful and compliant procurement, which has saved massive amounts of time and money being just really super exciting. And as I’ve mentioned to lots of people, you know, nobody ever wakes up at the age of 10.
wanting to work in procurement, but they should, Kieron, they should, because it’s really exciting.
Kieron White
I want to be a Procurement Officer, Dad. No, I don’t think anyone’s.
Neil Watkins
Yeah, that’s never going to happen. Even my grandkids are not going to say that to you. Granddad, I want to be like you, I want to be a procurement officer.
Kieron White
Huh.
Well, it probably won’t exist in a few in 10 years, in my opinion, but…
Neil Watkins
Well, indeed, yeah, yeah, we’ve talked about that too. But actually, this is part of the step towards that, isn’t it? And really interesting, because I got pushback from somebody who said, well, what happens if your suppliers don’t respond in times? Like, well, we give them a deadline, they’ve got 24 hours to respond. But if they, but what if one of them’s got an agent? I was like,
Well, if one of them’s got an agent that can respond to these things, then guess what’s going to happen? They’re never going to miss one. They’re never, you know, they’re not going to miss one because someone’s on holiday or someone’s out. They’re always going to be fast. They’re always going to be first in the queue. You know, and actually what we want to drive is
Kieron White
Yeah.
OK.
Yeah, accurate.
Neil Watkins
And one of the whole purposes of public sector procurement is to drive efficiency and effectiveness. And, you know, if people are going to reduce their prices because they can, they know they’ll get it faster, cheaper, better than, brilliant. What’s wrong? What is wrong with that as a picture? So, yeah, somebody, and I think it, you know,
I obviously won’t say who it was, but just because they as a supplier aren’t ready for the change that’s coming, that’s not my problem. But they can’t say I didn’t warn them. So
Kieron White
Yeah, there you go. They should be listening to this podcast so they know what’s going.
Neil Watkins
Yeah, that would make two people listen to this podcast, Kieron.
Kieron White
Exactly. Well, they go, we double our audience overnight.
Neil Watkins
Yeah.
Indeed, indeed.
Kieron White
Very good. Let me share with you quickly the tutoring research, AI tutor. So this was some research using Google Gemini in the UK. There were quite a few, I think it was 2000 odd students, split into three groups. One group had the usual experience, no tutors.
Neil Watkins
Oh, yes. Yeah.
Kieron White
One group had human tutors and one group had AI tutors. So interestingly, the bit that interested me most in it is that at the end of 10 weeks when they tested all three groups, there’s basically no difference between them. They all scored about the same in the exams. But here was the interesting part.
is that the human tutors slowed down those students, but gave greater depth.
Neil Watkins
Kane.
Kieron White
The AI so slowed down significantly, like the AI was able to do 30% more of the curriculum because it’s instant on demand, who knows exactly, but the assumption is because you can get an instant answer now, you haven’t got a queue up and wait and you don’t have to be kind of listening to a tutor going back too far, you get specifically to the bit you want to.
Neil Watkins
Mm.
Kieron White
clarify. But interestingly, and yeah, again, the kind of supposition here is maybe the human tutors are better, it turns out, but that pace is significant because the AI tutors covered the whole curriculum, therefore you can score more results, even though you might get 3/4 of the marks for everything.
Neil Watkins
Yeah.
Yeah.
Kieron White
Whereas the human tutor, you might be getting 100%, but you missed 1/3 of the curriculum. So it’s really quite interesting because I always, I worry a lot when I see the human versus AI. There’s a lot of that, as we know. It’s like, let’s compare them. And normally AI is about the same or slightly better, but nobody seems to talk about time in that.
Neil Watkins
Yeah.
Kieron White
And then right back to the thing we talked about a long time ago, the medical scanner, which is I think, I think was Northumberland Hospital, maybe Hospital Trust, maybe somewhere anyway, was that has an AI scanner and it can read your scan after a CT scan and in seven seconds you have your answer rather than two weeks of waiting or whatever long it takes for someone, the specialist to read it and come back to you.
Neil Watkins
Well, I had to go for a scan and as you know, I’ve got a problem in my eyes, so I had to go for a head scan. And they said, right, it’ll be two weeks. I was like, you already know, don’t you? I mean, I got off the machine and you already know, you’ve said, it’ll tell you, you already know, just tell me, just tell me. And they refused, I absolutely refused. And I was just like,
Kieron White
Is that right? It’s got to be the GP he goes back to.
Neil Watkins
Oh, no, there’s got to be the consultant, you know, someone’s got to just…
Kieron White
Yeah, back to the specialist. That is a load of process as well, isn’t it, that goes on.
Neil Watkins
Yeah, yeah, yeah, but listen, you already know, you already know, so just tell me. And they just absolutely would not. So, yeah.
Kieron White
You should have taken a photo of what they were looking at on the screen and given it to Claude or something.
Neil Watkins
But.
That’s right, yeah. I didn’t want to get into that much trouble.
Kieron White
No, indeed.
Neil Watkins
Now then, we were on a time limit today because you’ve got another important meeting coming up, but you said you had a funny to finish on, so what’s your funny to finish on?
Kieron White
I do.
Well, I was going to talk about Mrs Watkins. Well, well, the interesting thing was we were talking about AI and how it, you know, might come for lots of jobs and particularly in, you know, in all kinds of worlds, but we were talking about finance at the time. And interestingly, because that interview I then did, it was exactly the use case we were talking about. This, the interview I did, the chat was involved in this AI project where they had three people doing
Neil Watkins
Ohh, go on then.
Mm.
Interesting.
Kieron White
finance reconciliations, reconciliation against the bank account. And they had three people who would spend like 5 days off, I can’t remember what it was, AI did it all in 5 minutes.
Neil Watkins
Yeah, yeah.
Yeah.
Kieron White
and done and perfect. It’s like, just like, wow, that’s quite impressive. But anyway, we’re talking about that. And interestingly, Mrs Watkins reflection was, but what about the kind of the team? What about, and it was, and she was really talking about the kind of community of, you know, that kind of, well, yeah, community, I guess, of people that exist in organisations, the team spirit and
Neil Watkins
Yeah.
Yeah.
Kieron White
all the kind of, you know, that family type stuff. And it really interested me because, you know, on a cold-hearted shareholder level, it’s like, well, does this business exist to create communities of people to have to give them something nice to talk about, or does it exist to fulfil its ambition? But, you know, the short answer is, it’s not just one or the other, is it? And it did, I was quite struck.
And I have reflected on it quite a lot, and I haven’t got any answer that one.
Neil Watkins
Well, you and I definitely need to go and sit in a pub and talk about it, because I’ve been reflecting on that as well, because, you know, there are some challenges. We’ve always said, we’re not producing this AI to get rid of jobs. It’s to help people do a better job and deliver better services, whether that’s public or private sector, but mainly public sector for us. So I agree. And there’s been a couple of use cases.
including with the occupational health people that we started talking about, you know, where I’ve been very, very much pushing me. This isn’t about getting rid of people. This is about helping people do a better, more effective job. And, you know, I’m not being funny. I wouldn’t want to do five days of invoice reconciling. That would drive me. I would have a big pencil stuck in my leg all the time, be stabbing myself with it.
Kieron White
Just keeps him awake.
Neil Watkins
just to keep me awake. So yeah, I wouldn’t, yeah, interesting. We will, yeah, we’ll come out to that topic because I think it is a genuinely interesting one. So yeah, yeah.
Kieron White
It is, I think so. And the amusing anecdote to finish is I saw anything on Instagram, and I’ve no idea if it was really true or not, but it was pushed as one, was Burger King using an AI avatar to do interviews. And it’s really stilted and rubbish. But hilariously, this kid really was being interviewed. He had a basketball boot on his head.
Neil Watkins
Okay.
Kieron White
and no top on just to see what happened. And it and it and eventually it said, I think it’s quite inappropriate that you’ve got a shoe on your head. Anyway, tell me about how would you get to the local Burger King? And he goes, I’d get there, I’d get there because I’ve got feet. And then there’s this sort of awkward pause and it’s going, you’ve got feet.
Neil Watkins
Huh.
Yeah.
Kieron White
But that doesn’t explain how you’re going to get there. Please answer the question. How will you get to her? And he said, I’ve got feet, so I will walk. And he’s like, and then she get, he didn’t swear and said, that’s very inappropriate thing to say to me. What do your colleagues say about you at work? It was such a weird, stilted thing. It would tell him off and then.
Neil Watkins
Yes.
But.
Should.
Kieron White
immediately asked the question in its normal way. Very, very odd indeed, but it was quite amusing.
Neil Watkins
Mm.
Well, I’ll finish on with one amazing, final amazing thing because I know you’ve got to go, but I was told this week by two people who I’ve interviewed and bought and employed in other areas that one of them said I made their palms sweaty when I interviewed them. I was like, oh, crikey, I didn’t think I was that.
Kieron White
Blimey.
Neil Watkins
I know.
Kieron White
Well, you should have you should have delved into finding out more about what they mean.
Neil Watkins
I will when I see them, but that’s not for today. So right, you go do your next session, fella. It was lovely to talk to you. Have a great weekend when it comes and I’ll catch you next week for a very exciting episode with a special guest, Catherine Ritchie from North Yorkshire County Council. So looking forward to that.
Kieron White
Yeah.
Yes.
Right. Nice one. Take care. See ya.
Neil Watkins
Cheers, fella. Bye bye.
Kieron White
Bye.