#7 - 7 Apr 2026

Episode 7: Punch Cards, Agentic AI & KnowledgeFlow Outperforms Salesforce 🍺

It's Easter week and this week’s recording is in the morning, so too early for beer. Even for us. And that’s in spite of Neil’s granddaughter telling him to toughen up. Welcome to Episode 7.

This one's packed — from a proper history lesson about the early days of computers, to KnowledgeFlow quietly doing something Salesforce couldn't. Pull up a stool.

"There's no future in computers" — a history lesson 🕰️ Neil’s nearly 60 and reflecting like an old pensioner. His school had one computer in the corner and he learned to program with punch cards. His brother wanted to study computing at university, but the teacher told him: “There’s no future in computers son, study Geography instead”. How wrong was she?

Claude Co-work: brilliant, baffling and burning tokens 🔥 Kieron bought a Mac Mini specifically to experiment with Claude Co-work — the agentic AI that can control your computer. He's already burned through half his weekly allowance in half an hour on file reorganisation. One person accidentally spent £27,000 worth of tokens, so be careful out there. The use cases beyond file reorgs? Still figuring it out. The agentic wave is coming — but we're still at the "dial-up modem" stage, and that's fine.

The sales cycle problem — and what to do about it Three meetings with the same prospect. Same demo, slightly improved product each time. Now they're asking for a list of all 300+ AI assistants. Neil's verdict: don't give it to them. Find out what their actual problem is and fix that. Kieron reflects on measuring value versus usage — prompts are easy to count, but impact is what actually matters. Some good news: several renewals this week, which is always a better signal than any number of demos.

KnowledgeFlow outperforms Salesforce 🤯 A customer needed to find out which of their programme participants hadn't logged in since September — across three Salesforce data tables. Salesforce couldn't do it. They extracted the data, dropped it into KnowledgeFlow, and ran the query. 170,000 rows of data across three tables. 45 minutes later: done. Donald (CTO) was delighted. The pricing model, however, took a bit of a hit — roughly £50 of tokens for what counted as two prompts. A fix is incoming.

Product of the week 🎵 (hum your jingle) Smart Targets and Parent Reports — and this one is a genuine game-changer for education. Colleges have to write personalised SMART targets for every student — attendance, achievement, deadlines, the lot. KnowledgeFlow tested with 418 rows of anonymised student data and produced four personalised SMART targets per student in about three minutes. Teachers confirmed them and uploaded them straight back to Pro Monitor. The same approach works for parent reports too — and because it's so fast, you could send them weekly instead of termly. 

Oscar's all-nighter 🎓 Kieron's 19-year-old son Oscar messaged at 3:30am — he'd finally got his agentic AI project working. Messaged again at 7am to say he was going to bed. The next generation isn't waiting for anyone.

Going global, going to a wedding The podcast now has listeners in the UK, Canada, Australia, the Middle East and Sri Lanka (but only till our one listener there gets back from holiday). Kieron is off to his sister's wedding this weekend, where he has been given the role of "crowd control." Given his history, this is deeply alarming. Tune in next week to find out what went wrong 🤣.

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

SPEAKER_01 0:01

Alright, so shall we get this pantomime horse of a podcast underway? So uh good afternoon. Uh is it is it after morning. Oh crikey. All right, we've yeah, good morning. Glad I haven't got the time wrong. Um it is coming up to the Easter holidays though, so uh I'm a little bit discompopulated, but um uh yeah, I've got a few things to talk about this week, Kieran. I've got uh a little bit of a history lesson for you. Um good. Yeah, a little bit of time. I've got a couple of custom and a couple of customer stories that I wanted to share. But let me um let me start with a reflection and a little bit of history. So um I was I've been playing with Claude and Claude Cowork, and um it's really interesting, but also it's making me think I haven't got actual time to bloody keep up with all the changes and all the learning that I want to do. So um, and I was thinking about other people in senior positions in business, you know, CEOs and directors and business leaders, indeed anyone who who's got a full-time job, you know, how the hell do they? I mean, it's our job to keep up with AI and and any and we're struggling. So, how the how the hell we do it? And I was thinking back to um um, I'm nearly 60, you know, and uh I was thinking about 45 years ago, uh, we had a computer at school and it sat in the corner of the room, little pet uh old people will know what that means. And um, in order to program it, we had to punch cards. We literally had to get punch cards, get a little and do and so that we could do the coding, you know. And I was thinking about how that's different now from the agentic that's going on. Wow, and um uh I remember my brother, my brother really enjoyed computers, and uh he said to the teacher, uh, oh I'd I'd really like to go to university to do computing, miss. And she said, Oh, don't bother doing that, son. There's no future in computers, you should do geography instead. So dear. So he did, and he's regretted it ever since. Poor chap. Is he a geographer? Still, is he a geographer today? No, he's not a geographer today. He's uh he's a very hard work. He actually works interestingly for um charity, and um he's working with the lead hospitals on the multineuron stuff, which uh Rob Burroughs and and co so he's uh he he's he's doing he's doing that, which uh is is a really worthwhile and important important thing. But just come back to school.

SPEAKER_02 2:25

Uh when I was at school, they um it came to options, and uh a mate of mine at school had said he'd asked to do the typing option, one of the one of the options, um, and he was turned down because apparently it wasn't for kids of his caliber. So, and this will give you an insight for anyone that's listening into me, because because of that alone, I I wrote down typing on my list because I thought, well, I'll just intrigued to see what happens. Um, and I was called up and said it's not for you, you know, this is for the other uh kids. Um, and interestingly, I look back and think, buddy hell, that probably would have been the most useful, single useful thing school could have done for me is have me touch typing. But no, no, there was no future in that for people like me, it turned out.

SPEAKER_00 3:08

Yeah, yeah.

SPEAKER_01 3:09

Was it were you doing it because you thought there were lots of girls in uh the typing class, and that was why you thought you could do it? Is that what you did?

SPEAKER_02 3:16

Oh no, I literally did it was to test the obstinence of teaching. I also write sevens with the dash through, uh, you know, where you do seven with a line halfway through it, simply because, and I can remember very clearly the maths teachers having a go at one of the people in the class and saying, We're this a continental seven, we don't do sevens like that. And from that moment on, I do my sevens with a nine through. So that probably is more of an insight into me as a human. Who knows what it means? You don't do British sevens. Good old British sevens. Exactly. Continental sevens. That's me.

SPEAKER_01 3:50

Never catch on. Anyway, just to finish off my little story, I uh I was thinking about um um I was thinking about my learning and uh how I need to spend more time actually learning new skills. Um uh uh and I was trying to play it again with my six-year-old granddaughter, and it was one of those memory games where you turn it on a record and you put it in, and I couldn't bloody remember where the things were. And she looked across the table and she went, toughen up, granddad! Oh brilliant. Now my granddaughter's giving me advice.

SPEAKER_02 4:20

Very good. Well, she needs a job. Well, yeah, she's six now, isn't she? So you know I think I think it's a bit illegal to have six-year-olds working, but but perhaps soon she can manage you.

SPEAKER_01 4:33

Well, if she could, well, she she clearly is managing me, so yeah, she's got me wrapped around her little finger, I'm afraid.

SPEAKER_02 4:39

Very good. Well, I um, as you know, uh this week got a Mac Mini um and installed Claude Co-work on it because I really want to uh play, learn, understand, uh to see what we should know about and what we shouldn't. And um interesting. I've been I I daren't put Claude Co-work on my actual laptop because the big difference with Claude Cowork to the others, as our audience probably knows, but is it can control your computer? That's the point of it, is that it can do stuff, it can go on your web browser and go and go and book flights and things like that, or like quite get to the last stage, but it can do that, it can do all kinds of crazy stuff, including mess with your files, which is what I've been using it for. And here's the interesting thing: it's the use case you see all over YouTube about it is get it to reorganize your files stuff. And I've given it some limited access to some of my uh old files. I'm not that bothered about, um, and it does that really well. It I was I was quite impressed with. I mean, who knows what I've lost. I'll only find out one day, I guess. When I'm going, where's that? I'm not sure how useful it really is because most of the searching I do for files is control F, go find the bloody thing that way, because I can never remember what or what I would have thought about putting it in which folder. But it burns tokens like crazy. And there's an interesting thought for us because as you know, our model, we have chosen a commercial model which is largely sort of focused for public sector, well, any client that needs budget clarity on budgets, and we've chosen a model that says we will cover the token costs, we'll make an estimate of what we think you'll use, and then we have uh token costs, and we run that through a number of prompts, as you know, rather than actual tokens, because I think it's ridiculously unfair to say to a customer you have to have any concept of how many tokens are being burnt or used. I mean, how would you know? Even it's very difficult for us to know, and you know, data burns loads, not structured data burns heaps and heaps compared to a pretty big document which hardly burns anything. Um, but anyway, so we've gone for prompts. So, as in every time you ask it something, we might track how many of those you do, and most of our clients are on somewhere between 5,000 and 10,000 a month, depending on the uh what which level of service they have from us. But interestingly, the Claude thing, I mean, every six hours they give you a sort of limit for each six hourly bit, which is presumably managing their own, sort of throttling their own demand. Um, I burn all mine in half an hour at the beginning of that six hours, and it's happened every single time. I only had it two days, and I'm already halfway through my weekly worth on uh, and that's on the Claude Pro account, so I haven't paid for the Max account because that's loads of money. So I'm really and and here's the thing, and I hate to admit it, is I'm kind of struggling to think what is this useful for, because it's you know, paying the amount of money I have for a Mac Mini and the Claude license in order to organise some folders is kind of not really the you know, I mean that's a good experiment, but I'm like, okay, so what should I do with it now? Yeah, and that's where I'm kind of sitting there going, Well, I could do amazing things, interesting things.

SPEAKER_01 7:49

I could do marvelous things if I can only think of what those marvellous things could be. But oh yeah, just back to the history thing then. I uh I I remember when uh I got a dial up Modem and the internet first came out, and I was uh showing my uh father, uh uh and I was like, Oh look, and you can you can you can download music, and he's like, Go on then, and I was like, and he went made all those black noises. Come back tomorrow. And then 20 minutes later, knocking it at me said, This is rubbish. He won a box. It'll never catch on. Yeah, yeah, it'll never like computers, it'll never catch on. So I yeah, I I hear you, and I think, but I think it is you know, the whole agency, the whole the agentic piece, the uh the whole move towards automating, you know, that is an unstoppable wave, I think. But I just think it's gonna take a long time, a bit like the internet, a bit like the things with AI, you know, there's been a we'll go through the hype cycle. Um, you know, we're we're in we're we're getting towards the peak of uh agentic hype, and we'll go through the trough of despair anytime soon. But I think for us it's important to learn about these things because uh our customers are absolutely going to be reading the same stuff on LinkedIn and other social platforms are available um uh and and wondering the same stuff, and if we can help save them time, money, and effort, and and actually it stop them doing accidentally dumped stuff. I I was chatting to somebody last night who read a thing on the internet about um uh someone had burned 27 grand's worth of tokens using Claude Cowork. Um that's pretty that's pretty dire. Um you don't want that bill in the morning as as they rightly said. So yeah, I just think there's a whole host of um uh areas about it, and it's a bit, I guess, a bit like rag. You know, people s people say to us all the time, oh yeah, co-pilot can do that. Absolutely can't do that. Yeah, just just because you've just because you've swallowed the the Microsoft marketing uh stuff, uh or or you've been forced to buy it, or you've been forced to use it, doesn't mean to say it can actually do the things you want it to do. Um so and that and that takes time and and effort and learning. So I I yeah, I wouldn't give up on it just yet. And by the way, I've heard that they uh those Mac minis are really good at keeping your coffee warm because they run really hot. So if you put your coffee cup on, it'll keep your coffee warm.

SPEAKER_02 10:16

Now, the utility then, it wasn't a waste of money after all. So it was the most expensive coffee warmer you ever bought.

SPEAKER_01 10:24

Indeed, yeah. Very good. So, what else have you been doing this week apart from playing with your new toys?

SPEAKER_02 10:31

Well, I have been obviously out talking to people about knowledge flow and what it can do for them, and uh ultimately trying to sell in that. And uh interesting uh the continued challenge of sales, which it's worth I think pausing on as part of really for our own benefit, if not for our audience. Um, which is we get we our sales cycle is incredibly long for most people. Most organisations, I mean, three months would be quite normal for us from first contact to having something in their hands, and longer. We had one, I think was nine months, I think it was, from a from from agreed yes, we want it to actually getting down to doing it. And it's really interesting to kind of think about what's going on here, and obviously, from a business perspective, how we can increase the pace of that. I've got, for example, this week I met for the third time, having met for a first time at a uh at a conference event where I spoke for about half an hour to this chap, um, then met him, then met him and a colleague, and then yesterday met him and a colleague and two other colleagues, showed them largely the same thing. Knowledge flow obviously improves and increases what it can do every week with our new functionality, so a few bits that are new, and now they're saying, Okay, brilliant, thank you. Could you give us a list of all of the AI assistants that you have? Because as you know, we tend to build quite unique things for people. So I was showing them some things that I think they might be interested in, and then by the end of it, they were like, Okay, that's really interesting. Could you give us a full list? So I'm like, why are you gathering that? Yeah, what's going on in your mind? Well, partly, well, there you go, there may be, but it's like, what's going on in your thinking? Is this literally just a next step? Which is why obviously a thing people do, isn't it? Your point, uh, you've talked about before is if we offer free demos, everyone will say yes, because it's the easy thing to say to yes to, and then we have to sort of set people up and do some work and then barely use things some of the time. It's literally just a way of politely ending a call. And so I'm kind of like, what is going on in the decision making? And in this particular example, that you seem you think it'll be useful to you to know all of the sort of 30, 40 different AI assistant solutions that we could we have currently. Because you you need to what is your problem is the big challenge, but that's what you should be coming out of this from. It's like we've got this challenge, this fixes it, and then we've got potential to do other things. Great. Let's focus on the fix.

SPEAKER_01 13:03

Yeah, if it if it was me, I'd go back and say, I'm not giving you the full list because uh it's pointless. Because actually, it tell me what your problem is, and tell me what your three biggest problems are and we'll we'll work on those. So just give me one, you know, just looking down the list because it doesn't matter. You know, it's it's like uh on the website we've got I don't know how many, 15 or 16 of them. And the last one says, if you don't see what you what you want there, tell us what it is you want and we'll we'll make you one. So it's that it's it yeah, that seems really odd. Oh yeah, yeah. I know if I was you I'd be pushing back harder and going, yeah, maybe. No, I'm not I'm not giving you a full list because uh we've got I don't know how many we've got in the wild, 250, 300. I don't know what the number is.

SPEAKER_02 13:43

Probably something like that, yeah.

SPEAKER_01 13:45

We're not gonna list every single bloody one just because someone is it because is it because they've got a lack of imagination or they just want to find something, they go that's the one, or they're trying to create a business case to do it. But yeah, no, that feels really odd. It doesn't feel that indeed.

SPEAKER_02 14:01

But but it's often that's one example of what is a common process of I think it's it is a if we could crack it and help customers understand what they should be assessing, what they should be thinking about, what's actually important, and of course, you know, we're sales, we're on the sales side, so it's very difficult for and I always say to people, you've got to take a pinch of salt with everything I say, because obviously I'm a I'm immersed in a particular world of AI, the way we look at the world, but also I want you to buy our product, so it is a a a challenge to but I do try to be honest with people. I've sent many people away and said that you should just use go use chat GPT, it's nothing private, typically a lot a lot of marketing stuff like that. If you're going to produce an image that you're then gonna shove up on uh you know loads of websites, then you probably don't need a private model to do it. Yeah, interesting.

SPEAKER_01 14:53

Um we do lots of people who give us advice, and I'm using the term loosely, about where we should take our business and what we should do with it. That whole kind of you know, you should go deep in one section. We've we've chosen not to do that, and and it's and and that implies creating a killer product, you know. Here's the killer product which will solve your XYZ32B problem. And um and and actually that's not that's not the approach we've taken, and actually uh the bit with the customer who is looking if they're looking for a silver bullet, um, as a wise man once said to me, there are no silver bullets, only lead ones. You gotta fire a lot of them. So get on with it. And and and it is what's the biggest problem you've got? Let's help you fix that. So maybe maybe you need to have a more direct conversation. Well, not telling how to do your job, but you know, maybe a normal direct conversation, which is what you know, what's your problem? You know, why what what benefit would this give you? Because I don't think it's gonna give you any benefit. All it's gonna do is slow things down. If you don't want to go ahead, crack on. That's fine. We don't, yeah, we we'll help you if you do, but if you know if you don't want to work with us, that's fine. But actually coming back and asking for another bloody demo after all of that, it's like I'm not gonna show you all 300 um rags that we've got running in the wild, and some of them are confidential, obviously, to to other customers, so you definitely can't know about these ones. So yeah, that's really that's really strange. The um there is a couple of um uh yeah, I know it does take a long time, but once we do get customers, we've had a few renewals this week, and I was thinking about um, which is great, and I was thinking about what you said last week about it's not necessarily the amount of prompts that people put in, it's the value that is added. So, you know, they might have only had eight users and they might have only used used it 250 times this month, but actually, if they've got if those eight users got real value out of it, then um that's uh and that's really hard to measure, isn't it? And lots of people try and do it on things like time. And I've seen lots of people saying, Oh, yeah, forecasted savings are gonna be excellent. And I've always thought that was consultancy nonsense of of uh uh people uh trying to make stuff up to to justify it. But I think yeah, one of the things we should spend more time thinking about, and if our audience has got any suggestions, we'd have to take them on how we how we measure value, then that would be uh be really helpful.

SPEAKER_02 17:24

Yeah, and it's very interesting exactly that challenge as you and I think the um because trying to decide for people, send them an insights report. We we can run prompts. We have, as you know, the full audit logs of of everything that goes through our systems, which we can get out of a very few limited people in our organization can get to. We can, and which is the only way we really look at them is to run them through an AI tool and say, give me a summary of the frequent things, or um which is good and gives us that stuff. And I think that that route into what appears to be high value. If we can build a sensible prompt which knows enough about the kind of thing that business does, and then say, What tell me in this 10,000 prompts that I've run through this month, which things appear to be high value? Because I think that is giving those insights reports back. I'm I'm uh we touched on this at stand-up today, but I think if you think back to our consulting days, you know, one of the things you'd always be doing is ensuring that though when we're trying to make change management program a change program happen, quite often we would like try and, as I described this morning, rain down praise on those teams, those people that are doing the right thing, and you get the CEO to be aware, and the senior team get some thank you letters or something you know, mention in a in an all-staff meeting or whatever. Um, we are not enabling our customers to do that because invariably our kind of key product owner in an organization is not the CEO. Um, and we're not enabling those people to be able to go to their CEO and say, look, what a great decision I made, getting knowledge flow. It appears we've done these amazing things this month. Yeah, and that's we need to do that, and um, yeah, we can do it on time savings as we've done before, but as you say, it's the it time saving is fictitious nonsense, in my opinion, because it's not cashable, and uh you know, yes, it's nice, but hopefully it's good well-being stuff for staff, which is not unimportant, so um, that's good. But ultimately, if you're the CFO, you go and go, whatever, that was some more coffees they had that week. What what more did they do? Show me the profit on the bottom line from that savings. So I think being able to show some of the sort of more important and impactful things people are doing with knowledge flow, I think it's is where we've got ahead.

SPEAKER_01 19:46

Yeah, I think so. Um, and that's uh that presents a whole set of different kinds of challenges, doesn't it? Um but uh but good ones. It's like understanding our customers' businesses more, becoming more embedded with them, helping them more. Um, and we've got some really good examples of that where you know we've uh definitely seen a huge increase in in usage, but also therefore uh you know, we're equiting usage with value. Um uh, but we do need to spend more time working with them on on what that actually means in practice.

SPEAKER_02 20:20

I think it's loosely related usage to value, I think. It's uh probably it's an indicator, but nothing more. Yeah. I should share with you something that happened this week, which was great. I don't think we we in fact I'm almost certain we didn't talk about it. Did you get lost on a golf course again, did you? No, no, this is totally related to uh to knowledge flow. Um, and in fact, as you know, I have put all of the transcripts of our conversations we have in these podcasts into a rag model, the AI that we build. Um, and therefore I've asked it, have we talked about this before? So there's a good test to see whether, and it said you have not mentioned 170,000 as what I searched for um rows of data. So we had a uh customer contact us to say, uh, I might have messed up. Um they have a they are a Salesforce platform of Salesforce House, and this uh chap was trying to do some analysis. Um, I think I can probably say one step further. So he was trying to find out which of their many participants on their programs had not logged in since September. So um he got Salesforce, three different data tables were needed to answer the query, tried to do that in Salesforce and wasn't able to. So he extracted the data, put it into Knowledge Flow, and ran the same prompt, and Knowledge Flow got to work and it gave him 45. Here is 45 uh example answers, the first sort of sample, not example, sample. Um, do you want me to do the lot? And he went, yes. And so it got to work on 170,000 rows of data across three tables. Um, and 45 minutes later it was finished, and and he said. Suddenly realized that that probably burnt quite a lot of tokens. So we had a look and it certainly did. It looks like it burnt about£50 worth of tokens, which is pretty excessive. Um, and that's kind of fine as long as the client was because of course we measure on prompts back to the previous model. So that would have been two prompts, I think. It would have been one do the thing to the sample, second prompt, do the lot. So we would have we wouldn't, you know, charge by prompt, but we would attract two prompts. Uh so uh what I said to Donald when I had a chat with him, is that if if everyone starts doing that, we don't have a business in about a week and a half, I reckon. It's only when Mike's will send the end of the month bill.

SPEAKER_01 22:40

Oblivion 170,000 rules of data.

SPEAKER_02 22:44

The most amazing thing is that Knowledge Flow was able to outperform Salesforce, it did the job, and yeah, it took a bit of time doing it, but there's a huge amount of ask that he was doing to to uh try and create it. And when I told Donald about this, our CTO, uh, he had a big smile on his face and he said, That's exactly what I've coded it to do. I'm delighted. And I was like, Yeah, no, I get it, Donald, I get it, but there's this whole problem of how we so we're gonna build in what he said he's gonna look at is building in a more stringent warning, firstly, possibly a setting you've got to turn on to do mass data above whatever thousand records or something, um, and counting it in more prompts, finding a way of saying, well, that you know, if you want to do that, that's like a thousand prompts or five hundred prompt or whatever there's a sensible uh level. So that if if an organization wants to do that, then they'll need to just move to our different levels of our service because we obviously can't pay for them at a loss to do it.

SPEAKER_01 23:45

Yeah, that uh what's really cool about that is we've we've talked probably for a year, a year and a half, about people wanting to be able to query structured data and it being super problematic because of the way the models in the past have you know that they disconnect the the rows and so they they the way they chunk them uh uh and and they they miscalculate or they misread and uh get the if the headings aren't right, etc. etc. So the fact that it's it's done that and it's taken it and it's done something that Salesforce couldn't do is even is even better. That's brilliant, really good.

SPEAKER_02 24:22

Yeah, very impressive. Which um if I if I may hog the airwaves for a moment longer, is brings me to our product of the week. Hang on a sec. So yeah, playing your jingle in your head. I should remind people that might have only watched the podcast now. That comes from if you I think we talked about it in podcast one when we um I saw Stuart Francis put doing a gig, uh comedy, comedy gig, and he was on stage saying, I'm filming today. Um and he said he said, right, so at this stage I would normally play some music behind the scenes, but because of Brexit, I can't do that because it means I have to go and get the IP rights or whatever it is, the rights to have that music in so many different places, it's a nightmare. So he said, Can you please just imagine some music now? And of course he laboured it in his comedic style. Kept stopping and going, hang on, I just want to check the kind of music you've got. I don't want it too uplifting. Anyway, so we uh for our product of the week, you have your own jingle at this point.

SPEAKER_01 25:21

I do, it is quite uplifting and it's quite it's quite um it's quite playful, let's call it that.

SPEAKER_02 25:27

Good. Well, so the product of the week is uses the same kind of uh knowledge flow capabilities as that data task. It uh smart targets or parent reports, so education focused, but I think this has lots of legs in all of the sectors in which we're working, you know, housing particularly. But for for colleges, so colleges all have to write smart targets for their students, and that's uh the tutors that have to do that. And typically a student will have multiple smart targets, maybe one about attendance, maybe one about handing in on time, maybe one about overall achievement and how they can improve things. Um and it's a headache to do well. Writing a smart target is quite challenging, actually. One's probably all right when you've got to write sort of 40 in a row. I imagine it's quite irritating. We have built in Knowledge Flow uh an ability where you can load the structured data set that relates to each student, and you can do that as a big batch. So colleges often use a thing called Pro Monitor, which is the MIS system kind of that tracks all of that data. You can load that, and Knowledge Flow will then analyse that and provide the smart targets per student, totally personalized to their own attendance, the course they're studying, etc. etc. So it does it did four. We've tested it with 418 rows of data that a college sends us, anonymized, um, and we uh tested it with four smart targets each. It takes about three minutes to do the lot, and they have said that's brilliant, and they've literally taken them and uploaded them back to Pro Monitor for checking. So there is uh and you can do exactly the same thing with parent reports. We've done that with it as well. You take the same data and you write it now for a parent to give them a summary of how things are going. And and my excitement in that is one, it is a task every education organisation has to do, and it's frustrating, takes huge amounts of time, and most parents probably have a cursory glance and give their son or daughter a pat on the head or a slap around the ear or whatever. Not that I would do that anymore, Kim. No, indeed, yeah, yeah, back in the days when you were punch carding computers. Exactly, exactly. And um, but interestingly, the parent thing particularly is the ability to now do that every week if you wanted to. No longer is this because it's so laborious. You couldn't do it, you could do it every week on whatever data sets you wanted. You just our world would be telling Knowledge Flow what data it's got. It actually does a pretty good job of writing its own data dictionary, but if they are acronyms and weird column headings, then it will try and you know it will do its best, but better if you tell it. But you could then say this is you know, this is the CPOMS behavior deck record, this is the achievement record, this is the sports stuff, whatever. And you could send that every week. Every day if you wanted to, and it would literally press a button to do.

SPEAKER_01 28:22

Wow, that is cool. Uh I'm reminded of two stories. Uh one is uh when my boy was uh secondary school every Sunday evening, I'd grill him he hit Sunday dinners uh on because he he'd get another question about uh how are you doing on this and why aren't you doing that and da-da-da. And the second thing is uh we went to uh one of my daughter's parents' evenings, and um uh the report was really um how'd she say, inadequate. It was uh and and Anna's like, that's not hang on a minute, hang on a minute. And the teacher went, Oh yeah, no, that's not you.

SPEAKER_02 28:54

She read us the right same thing happened to my mum, she told me, and she said that doesn't sound like him at all. And they went, Oh no, actually, you're right. That's the teachers really knowing their pupils really well there.

SPEAKER_01 29:08

Which one are you again? Well, I think it's part of the challenge, you know. It is the whole teaching challenge. You've got so many kids you've got to write if you've got to write these things out. I mean, in the olden days, we used to get them, they have to write them out by hand. I mean, I remember books of things. Actually, creating something which is bespoke, specific, and targeted is I think is brilliant. It's that's got to be a God sent for them.

SPEAKER_02 29:32

And I think you can make it really clear that that's what you're doing, because as I mean, you can't shirk responsibility or accountability really for the teacher, at least confirming it's accurate, but it's going to be accurate by the data that's already in there. And to say to a parent, you know, we're going to run AI report, so you'll get a weekly personalised report. We get a newsletter from our school weekly, and it's lovely. I mean, it's lovely that they do it, but frankly, it's you know, you have a quick skim and go, right, okay, thanks. But if you've got a weekly and here's how your child's got on. Here's how Finn's doing on his quadratic equations. I think I know the answer to that one already. You don't need to write to me about that.

SPEAKER_01 30:10

For transparency, I think Finn is Finn six, yeah.

SPEAKER_02 30:14

He is six, yeah, Jeff. But here you go. Let me, whilst we're talking about children, so uh my eldest boy Oscar, who I know is an avid listener, he probably is our only listener, to uh to our previous audience still asleep from the last last. So I got a message from him at 3:30 a.m. this morning, which woke me up and I was like, uh-oh, what's up? Um and he said, I finally got my agents doing this thing on whatever he's trying to play, whatever he's created using uh some pretty crazy AI stuff. So he's wiped his laptop completely and started afresh, so he's got like a clean thing that he can play with and trying to build stuff. And I was like, God, I remember pulling all nighters at university, and he's at uni, and I just wish he would spend some of that same time, maybe focused on some of his work. But I mean, what a wonderful experience that he's actually going through. I'm super supportive of it. I think it's brilliant, and you know, hopefully it might turn into something, or he'll learn enough from it, and that will turn the next thing will turn into something, who knows? But um, yeah, he uh he's been at 7 a.m. this morning he messaged me again, saying, I'm just going to bed now. So uh that's a joy of being 19 years old, I guess. But it was but yeah, he's been like smashing through trying to, and he's got a couple of mates that that do it with him on different they're all sort of trying to build different things out. Obviously, all gonna make millions of pounds, that's their main focus. If I could get him a bit more focused in the public sector ethos, a bit more, let's make a difference to people's lives rather than make a difference to your bank account. But um, yeah, really cool that he's doing it.

SPEAKER_01 31:47

Yeah, very cool. The um well comes back to a couple of things, doesn't it? You know, he's got he's got time on his hands to invest in the learning piece. Um, brilliant, which is where we can we we we kind of started. Um uh and uh the other thing that kind of flashed through my mind, I think we we talked about it last week. I haven't used your rag tool because you haven't shared it, so I don't know whether we did talk about this or not. I'll check for you live if you like. Yeah, yeah, yeah. Uh we talked about social workers turning down jobs because they didn't the councils didn't have an AI. And you know, Oscar's that kind of guy. He's gonna turn up at a place for a job interview and he's gonna go, Well, I've done this and I want, you know, how am I gonna do that? And they'll say, Right, here's some here's some bits of card and a little punch machine, and you have to code using the punch cut, he's just not gonna bother, is he? So really interesting that the the younger people who are doing that stuff, they're they're brilliant. I really, really enjoy it. It's good. Yeah, it's very good. Cool. All right, fella, is there anything else? It's coming up to uh the uh Easter weekend. I hope you've got lots of um interesting things to do. I know you're a bit of a chocolate fan, so you will be chocolate fan.

SPEAKER_02 32:56

I will be hunting for the Easter Bunnies deliveries, but actually I'm going to my sister's wedding, which is very exciting. It's been a very long time coming, and uh they're getting married on Saturday. So uh going to that, where I have a small role to play, as I'm not really sure what they're calling it because it's quite an informal sort of approach, they're doing it their way, which I fully support and appreciate. Um, so I've got a kind of I think generally crowd control, I think is probably the best way to describe it. You crowd control.

SPEAKER_01 33:25

You're the you're the inciter, you're the when it comes to when it comes to bad behaviour, you're the common denominator.

SPEAKER_02 33:32

I know. I remember I was thinking, I think about this too many times. Um, do you remember the Folly Inn in the Isle of Wight that we used to visit quite frequently? And there was the landlord there was a guy called Andy for a long time, and he was really, really good, and I knew him really well because I was there quite a lot. Um, and I remember he had always had music on, and they would in fact encourage dancing on the tables, which was always a great fun, and everyone was there uh rather drunkenly. Anyway, the music stopped, and they was uh trying to get everyone out, and I led the entire pub in a sing song of the Wild Rover. And he I remember he came up to me and he said, Kieran, please stop. And I and I did uh even through my drunken enthusiasm, I thought, oh yeah, I've really overstepped the mark here, haven't I? But I was just singing it and got everyone going back on the tables singing the Wild Rover.

SPEAKER_01 34:23

Well, I look forward to next week hearing about how you've disrupted your sister's wedding.

SPEAKER_02 34:28

Hope not. I'm hoping I behave myself, but the joy of being me is you never really know.

SPEAKER_01 34:36

On that note, you have a great weekend, fella, and I'll be right back.

SPEAKER_02 34:39

Thank you, and you enjoy the chocolate. Alright, see ya.