Listen to the show using this podcast player:
At Forvis Mazars in South Africa, the team is actively working on AI-driven solutions for clients.
Shane Cooper (Head of Digital Advisory) is spearheading this effort, with one of the applications of this technology being in the portfolio management space. Institutional investors with complex structures face multiple challenges in managing their investments. As Shane explains in this podcast, it’s an operating model problem rather than a software problem – but AI can help.
Rishi Juta (Director of Corporate Finance) joined this discussion to deliver insight into real-world applications across due diligence and risk management. It’s all about transforming unstructured data and commentary into useful information for decision-making.
This is an excellent introduction to the technology that Forvis Mazars in South Africa is developing for institutional clients.
Topics covered in this podcast:
- The shift from mechanical work to judgement work – and why it changes the entire process of portfolio oversight.
- Why unstructured data (like management commentary and board‑pack narratives) often tells you more than the numbers.
- How “intelligent ingestion” lets AI chew through PDFs, emails, scans, and commentary like a grown‑up sorting out a toddler’s plate of vegetables.
- Early‑warning risk signals across a portfolio: covenant pressure, reporting behaviour, management tone, governance drift, sector stress and more.
- How this tech is being built specifically for regulated environments – IFRS, GRAP, scenario planning, traceability, explainability and all the governance that institutions actually need.
- Why large‑scale portfolios guarantee that humans will miss something – and how an AI layer can stop the rot early, while still taking advantage of having a human in the loop.
If you would like to learn more about this technology, connect with Shane Cooper or Rishi Juta on LinkedIn. For more information on AI-specific applications, you’ll find Shane’s contact details on the Forvis Mazars website.
Full Transcript:
The Finance Ghost: Welcome to the Ghost Stories podcast. Let me tell you, it is going to be properly interesting today, because we are going to learn all about a real-world AI solution. Drumroll please. Exciting, right?
We keep reading about AI all over the headlines, the internet, and social media, and all over the place. And today we’re going to get close to a homegrown South African solution. Quite exciting.
Let me tell you what it’s trying to do before I introduce you to our guests. So here’s the key issue. Let me paint this picture for you.
You have these huge institutions, and they manage large and complex portfolios in a (usually) quite cumbersome manner. Not because they want to, but that’s just because it’s the way these things practically tend to happen.
Large teams of people running in different directions, having all kinds of conversations with investee companies, obviously of varying quality and on different topics and with different levels of engagement. Not necessarily a huge amount of consistency.
And then it’s quite difficult to pull everything together for credit meetings. If this is starting to resonate with you, then you’ve spent some time in large institutional investing or in advisory. Investment reporting certainly puts pressure on the value chain as well.
I’ve seen this in practice. I know that many listeners to this podcast will be familiar with this world.
And what is interesting is that there is a better way to do it – there’s a way to actually take advantage of some pretty cutting-edge technology. AI, of course. But AI is really just a means to an end, right? What matters is how you use it, what you build and why you’re doing it.
Now, to help us understand that properly today, I’m happy to bring back a familiar voice to Ghost Mail readers and to Ghost Stories listeners. You may remember Shane Cooper from Forvis Mazars in South Africa.
He is the head of digital advisory, and he did a podcast with me last year on international use cases for AI. The man is building what he’s talking about here. He’s not just commenting on what other people have done. He has put something together at Forvis Mazars that we’ll learn about today.
Also joining us today is Rishi Juta. Rishi is the Director of Corporate Finance at Forvis Mazars in South Africa. So he certainly knows his way around due diligences, valuations, and structuring of corporate transactions.
Shane, Rishi, welcome to the show. Thank you for doing this. I’m pretty excited to learn more about it.
Shane Cooper: Thank you, Ghost. Great to be back.
Rishi Juta: Great, Ghost.
The Finance Ghost: All right, Shane, let’s start with you. You call this an operating model problem. Interesting. Not a software problem. I’ve seen this word come up in some of the discussions we’ve had.
Just for the record, I’ve actually been shown the system and it’s pretty interesting.
So, Shane, not a software problem, an operating model problem. What do you mean by that? What are you actually trying to say about the state of play out there, and why do you think this product is important?
Shane Cooper: Most institutions do not actually suffer from a total absence of software. I think that’s clear. They suffer from the way the work gets done across the chain, from data coming in, to decisions going out.
So if you look inside a lot of organisations, especially institutions managing complex portfolios, the real pain is not that there is no reporting tool or no dashboard somewhere. There is a plethora of those out there. The real pain is that the operating chain is messy.
Data arrives late in different formats from different sources with different levels of quality. And we see that teams spend huge amounts of time extracting the numbers, chasing missing submissions, reconciling versions, normalising templates.
Inconsistencies abound!
And of course, the time that you spend assembling packs is crazy, if you think about the cost of the skills that you bring into an organisation relative to what you’re getting out.
And by the time all of that is done, people are moeg. The reporting cycle has consumed a lot of energy. And then of course, the real question is whether everyone has actually had enough time to think properly. That’s why I call it an operating model problem.
The bottleneck is not the absence of software. The bottleneck is the way the organisations are structured, which we accept. Too much effort goes into the administrative motion – classic corporate friction – and not enough into interpretation, judgement, challenge or intervention. Doing stuff with the data and creating value out of it.
For me, for us at Forvis Mazars, we really think that matters more today – because the world has become more complex. Portfolios are data-heavy, your governance expectations are much higher, committees want faster answers.
And the date itself is not arriving in lovely neat rows and columns. It’s arriving through emails, PDFs, Excel files, commentary packs, scanned documents, classic unstructured formats – and if you’re lucky, you have an API link.
So if the operating model is still based on people manually stitching stuff together, the reality is that every month or quarter, you’re going to hit a ceiling.
The Finance Ghost: Yeah, I love that. It describes so much experience that I’ve had in my own life in corporate advisory, and all these inefficiencies. And the other point I want to touch on there, that I think is going to come up later in the show and is so important, is the data that comes through in natural language conversations and commentary packs.
You learn more from that, in my opinion, than you do from digging through some of the numbers.
Certainly, a lot of the numbers are critical, and Rishi is going to tell us about that shortly. But you certainly learn a lot from what management is saying. That’s why investors refer to transcripts, and they try to learn something from that. In private company land, that transcript is a board pack, or it’s a meeting between investors and management. It’s really valuable stuff.
So Rishi, let’s then move on to how you essentially spend a lot of your time: combing through this private company data, doing due diligence, doing valuations. This is technical work; it’s difficult work. And if the data is not structured nicely or if it comes to you in very poor shape, I imagine it just makes your job that much harder, right?
So what are some of the challenges that you end up seeing in practice, as you’re doing your day-to-day stuff in the corporate finance team there at Forvis Mazars?
Rishi Juta: Thanks, Ghost. Well, I think it’s a completely different world and I don’t think people fully appreciate just how different it is until they’ve actually worked in it.
When you’re dealing with public company data, you at least have some scaffolding around you. There’s a market price, there are standard disclosures, there is analyst coverage. There are public filings and there’s broader news flow. And there’s actually some degree of structure in the information released. It may not be perfect, but it is a framework.
In private company work, especially in due diligence and valuation environments, much of that scaffolding disappears. What you get instead is a much more fragmented information landscape.
You may get management accounts, but they come in different formats with different definitions. You may get boardbacks, commentary notes, customer concentration, schedules, budget files, cabinet packs, legal papers, strategy decks, email explanations and operational updates. All sitting in different worlds and often speaking different languages, even though they are supposed to be describing the same business.
That is where unstructured information becomes extremely important. A lot of the real story of a business sits outside the financial statements. It sits in the management commentary in the way issues are described, in how risks are framed, in what gets emphasised, in what keeps recurring, in how operational setbacks are explained, and where the strategic story is consistent over time.
You can learn a lot from how management talks about working capital, customers, delays, margin pressure, leadership changes, or refinancing conversations. And it’s really interesting what happens over time. One month of commentary is just a piece of narrative. Four or five years of commentary starts to become a data set. Then you can begin to see patterns.
Is the tone changing? Are there some issues being dressed up in different languages? Are strategic priorities shifting too often? Is management becoming more vague or defensive? Are explanations becoming less specific as pressure builds? Is there a growing gap between the story and the numbers?
That is why unstructured data should not be treated as background noise. In private company environments, it is often one of the richest sources of early insight, especially when combined with structured financial data.
The challenge, of course, is that most traditional tools are not particularly good at handling that. They are often fine with tables and figures, but not with years of qualitative material that may hold clues about leadership quality, strategic consistency, accountability and emerging risk.
So the opportunity now is to bring those two worlds together, the structured and unstructured, in a much more useful and governed way.
Shane Cooper: If I could jump in there, Ghost. What we think is particularly useful today is dealing with unstructured data, where portfolio oversight becomes really interesting. It’s not just about the numbers anymore. We often make investments as a broader community in the quality of the leadership. But seldom do you interrogate that pattern of language over time.
The ability now to keep (as part of your corpus) a history of language, creates some really interesting insight on your investments over time. And this is where unstructured data can become incredibly powerful.
The Finance Ghost: I love the reference to that history of language. That’s exactly right. You can even see it in my world when I’m looking at listed companies and the transcripts. You can see how the management tone moves over time, and it changes. And like you say, Rishi, it becomes more defensive or more vague.
And obviously, in your day-to-day, you are applying professional judgements. That’s why people are hiring you to say, “Listen, please do a due diligence, come back to us with some kind of recommendation based on how you see this thing”. And so you’ve got to do that anyway.
But you’ve got to do that whether the data arrives and it’s very messy (and it’s a huge menial task to actually get on top of it) or whether it’s a bit easier to get to that point, and then you can actually spend more time thinking, which I guess is the overarching point of AI at the end of the day.
But Shane, maybe coming back to you on this, the model is only as good as the data that it gets, right? And that’s the reality. And the way it’s been built, and the way it’s been trained, and all those things about AI that are interesting.
There’s something that I read in a piece you wrote that talks about “intelligent ingestion”, which feels like something I’d like you to teach my toddlers about their vegetables. But I think that’s a very different context. This is intelligent ingestion of data by these models.
What does that actually practically mean? Because this is where the rubber starts to hit the road, right? On what these models actually do. AI, it’s just two letters. It’s a small little term, and yet it’s this incredible technology that has all these elements to it, and this is one of them.
Shane Cooper: The simple idea is this. “Intelligent ingestion” is the process of taking very messy incoming information and turning it into something usable, something that’s structured, that’s governed, and that’s even decision-ready. And you don’t have to rely on armies of people to get it manually done every time.
So in a normal institutional environment, your data doesn’t arrive in one perfect stream. As we’ve already said, between Rishi and I in the early part of this discussion, it arrives in various forms. Typically email, and normally in the email, the attachments of PDF, Excel, whatever. These inconsistent formats create a huge challenge for data ingestion.
If you want a coherent view of your portfolio, the first problem is not the final dashboard. The problem is getting all of that raw data into a state where it can be trusted and used.
So intelligent ingestion is about using technology like document processing, parsing, classification, automation; some RPA built in if you really want to. And then you apply machine learning rules, engines and AI to recognise what has come in.
Step number one: you need to recognise what comes in, identify what kind of document it is, and then extract the relevant information. And then you map it into a canonical structure. So there is some upfront work that’s required to ensure that the mapping is correct.
But relative to time that you’d spend over a longer period, this is an investment well worth going through, and then capturing the narrative parts as well as the numerical part. So normally with ingestion, historically you just capture stuff that could be deemed to be structured, and normally that’s numerical. Now you have this amazing ability with AIs to fully ingest unstructured information.
You do apply some validation checks. We still like to refer to “human in the loop”, just to make sure that you’re 100% certain that the data is transferring correctly. And then for anything that is uncertain or anomalous, you can request a human review as part of your process.
Now, the last point is important, because we don’t want to pretend that everything comes through magically clean. But from what I’ve seen over the last three months in particular, the technology has moved on significantly. So the point really is to automate the repetitive grind. And then you want to let people focus on the exceptions where judgement is required, and of course, interpretation.
The analogy around a toddler eating vegetables is something like this. You’re faced with a plate of peas, carrots, broccoli and chaos. The toddler doesn’t see nutrition; it sees some level of betrayal, if you’re anything like me.
The Finance Ghost: Poison and toxins!
Shane Cooper: Exactly. The intelligent ingestion is basically the grown-up equivalent of taking the vegetables, chopping them properly and sorting them, mixing them to something more useful, and then removing the obviously terrible bits. And deciding what can be eaten safely, and what still needs adult intervention.
Now, in the portfolio world, which is really why we are here talking about this, instead of broccoli in the trauma, it’s the management accounts, the commentary, the covenant packs, the risk notes. Those things traditionally were all very manual in terms of how you would handle them.
Now what the system does is (in terms of what we’ve designed)(especially in the unlisted environment, because your information is not that well organised) it helps sort the mess. It extracts what matters, it flags what looks wrong, and then it creates a consistent information base that can actually support the analysis.
And where this becomes economically important is that if you don’t solve for ingestion, then all of your expensive people spend their time opening files, moving numbers, fixing mappings, and checking versions. It’s not anywhere near high-value work.
So once you industrialise this ingestion, you properly unlock the rest of the chain. Better reporting, faster ratio calculations, earlier risk identifications, essentially the stuff that you really want to be focusing on.
The Finance Ghost: That makes a world of sense.I do enjoy the leaning on the toddler example. So thank you. I think that actually makes it clearer. Sometimes you just need to get this very visual representation to be like, “Okay, that actually does make sense”. So thank you. I like that.
So far, we’ve talked about how, in private company land, the data is going to come in a way that is messy. We accept this. A lot of it is going to be natural language, a lot of it is going to be discussions, maybe even be emails. Who knows? Anything that is just natural language.
That, for me, in my very layman’s understanding of AI, is where AI is so different to the world we’ve come from. It can actually read and then summarise properly, in theory at least, and come up with some pretty cool insights.
So I can understand how this machine is just creating more data across a portfolio, and feeding it into this AI model. But of course what really counts is what comes out the other side and what we can do with it.
So Rishi, I’m going to bring it back to you here because now we’re stepping back into your world.
You’ve already mentioned some of the issues around the data that comes in a private company. I think as Shane has explained to us, maybe it is really just chopping up the carrots and the peas, and just getting it to a point where this toddler’s going to eat it. Now it’s your job to eat it, right?
Sorry for using you as the toddler in this analogy, but unfortunately, it’s about the closest we can really get here [laughs]. Maybe we’ll make you the parent; we’ll think of an innovative way.
But the underlying principle is that there needs to be a point to all of this, and that point needs to address a client’s need. Now we’ve got financial investors, we’ve got DFIs, we’ve got institutions. They all have nuanced investment philosophies. One might be purely for profit, others might be impact investing, and they’ll have different kinds of metrics and everything else.
I feel like it all comes down to measurement and risk. At the end of the day, that feels to me like what it really comes down to. In your eyes, how does this technology make that whole process better, from the perspective of the clients and what they’re looking for?
Rishi Juta: Thanks Ghost. I think that’s exactly the right way to frame it, because once you strip away the jargon, the real issue is risk, and the ability to make good decisions early enough for that decision to still matter.
Different institutions absolutely do have different philosophies. A purely commercial investor may be focused more narrowly on return, downside protection and exit value. A DFI may also be thinking about development, impact, mandate alignment, policy relevance, governance and long-term sustainability.
But in all cases, you still need a way of converting a messy information environment into a usable view of risk. What this kind of technology does, is give you a much stronger risk measurement and decision support layer.
It allows you to bring together structural financial information, covenant positions, operational indicators, management commentary, reporting behaviour, and external signals into one environment.
That matters because real-world risk is rarely one dimensional. It is not just one ratio moving. It is often a combination of deterioration in the numbers, changes in management tone, missed reporting deadlines, governance drift, sector pressure, and weakening narrative consistency.
So from a risk perspective, the benefit is that you can start seeing patterns much earlier. You can move from static, retrospective reporting into a more dynamic view. You can assess not only where an investor is today, but how it is trending, how close it is to stress, what the pressure points are and whether the warning signals are financial, operational, behavioural, or external.
And that becomes especially useful in environments where early intervention matters. If you wait until the business is already in serious distress, the room for recovery is much smaller.
But if you can identify signs of covenant pressure, cash strain, operational slippage or management inconsistency early, you have a much better chance of engaging management, challenging assumptions and pushing a remedial program where there is still an opportunity.
It also helps because risk is not just about the business. It’s about leadership quality, execution, as well as discipline. If the qualitative material over time is showing strategic inconsistency, repeated excuses, declining clarity or weak accountability, that is relevant. Good risk management should be able to consider that, not just the income statement.
The real value here is that it gives institutions a more complete and more timely picture of risk. It supports better prioritisation, faster escalation, and more confident decision-making.
Shane Cooper: What I would add to that, Ghost, is that the covenant intelligence provides very useful information, because once the data lands, you instantly have your covenant position, your key ratios, and then of course your distance to breach, which is useful information.
The second is the early warning triggers. Those go well beyond pure covenant maths. They have to include things like reporting delays, management churn, sector stress, and adverse news, especially in today’s challenging economic environment. Those are all the signals that help you stop the rot early, rather than only documenting the distress after the fact.
And the reason why we feel strongly about this is that very few companies come back from deep formal distress gracefully. For us, the real leverage here is this earlier engagement, the earlier challenge and early remediation.
The Finance Ghost: So the way I like to think about it is, I listen to all of this, and I know that when you have smart people working on your portfolio and on your due diligence, if they have a relatively small number of data points to look at, chances are very good that they will pick this stuff up. I think we accept that. And even then, there’s a risk that they won’t, because human error is part of it.
Shane, in that podcast we did last year, you pointed out some really cool use cases internationally, where some of the best engineers in the world are using AI because of the human error factor, and because the AI rules-based approach catches a lot of that stuff. So even on a relatively small data set, human error is a risk.
But when you get to a huge portfolio (and here we’re talking institutional level applications, really) and you’ve got multiple portfolio companies, lots of different board packs all the time, you’re getting inundated with this stuff, you’re looking at new investments all the time where you have to make really good decisions – it’s basically a guarantee you’re going to miss something, right? When you’re managing a portfolio like this, it’s just a reality.
So this is where this AI model steps in and says, “Listen, we can give you the early warning triggers”. And like you say, it still needs a human in the loop and someone who can apply that judgement to it.
Rishi, when you’re working on that due diligence, you’re applying judgement to what’s coming out of it. But if there’s a tool that gives you a better way to catch everything, or at least to get closer to catching absolutely everything, that can only be a benefit.
That really does feel very sensible to me. And again, I’ve had the benefit of actually seeing the product.
So I would encourage people who are listening to this, if you think this is interesting, speak to the team. Because actually seeing the stuff is pretty cool. It really does help you visualise. “Oh, this is interesting. This is very colorful”. This dashboard shows you, across all these different names, where it might be going wrong, where it’s not, where it’s already in breach, where it has a high probability of breach, and all those kinds of things. Which I do enjoy.
And Shane, I’m going to come back to you here, because I can see these use cases. If someone is listening to this and going, “Oh, I’m not sure what this is for”, then I would be so bold as to say it’s probably not for you because you haven’t lived these problems. These problems exist.
So if you’re listening to this and going, “Oh my goodness, that sounds like my life”, then this is for you. The point is that it just takes away a lot more of that menial work. Both of you have already alluded to that. And it gives you more time to actually do something with the outputs, right?
I just want to throw it back to you, Shane, for some commentary on that. Is that the theme that keeps coming through in how these AI solutions are really adding value in companies? Because I’m certainly seeing that a lot, when I read company narratives about AI projects and what they’re doing.
Shane Cooper: That’s exactly the point, but I’d phrase it slightly more carefully – the value of AI is not just that it saves time, the real value is that it reallocates human effort from low-value mechanical work to high-value judgement work.
And that shift is really important. The additional benefit is that it can do for you, what, as a human being, you were never capable of doing before.
As human beings, we’re not put together to understand large data sets and to make connections and correlations that may not appear to you at first glance. So applying AI to huge data sets is extremely powerful.
In a lot of institutions today – (you would know this, Rishi knows this, I know this) – very smart people are spending way too much time on report preparation. Pulling numbers together, checking templates. There are lots of memes about how PowerPoint and Excel essentially run the world. Reconciling versions, writing commentary from scratch, chasing updates. Seriously heavy administrative burdens that we take on in our professional capacities on a daily and weekly basis.
It’s useful in that it has to get done. But that’s not where the biggest value lies. The real value sits in asking what the numbers mean, why are they changing? What risks are forming? Where is the portfolio the most exposed? Which companies need intervention? What should management be focusing on? What scenario should be modeled and what should the next decisions be?
Very often in today’s corporate world, you don’t get enough time to handle those conversations. Now you do. So AI and automation should absolutely take away a large portion of the grind.
Intelligent ingestion certainly helps, automated calculation certainly helps, and even the report generation (certainly, at first pass, that breaking of the white space) really does help.
The biggest story is that once you have a governed base of structured information, AI can also help with anomaly detection, pattern recognition, peer comparisons, and correlation studies with data that you’d never have potentially brought into your data set. And, of course, propensity analysis, which is for me extremely useful based on the unstructured data that you’ve now brought in.
The platform is not just helping produce a report faster, it’s helping you interrogate the report better. It can surface unusual movements, non-obvious patterns, combinations of signals that may indicate emerging stress, and then areas where human attention should absolutely go first.
Now, where it gets super exciting for me, is exactly in this point – because the end state is not just a faster reporting machine. The end state is a better decision environment. And I think that’s particularly important in the institutional world because the cost of late insight can be very high.
If a team spends most of the cycle preparing the pack, and only a small slice of the cycle time thinking about it, then the organisation is structurally late. Now if you flip that, the report comes together much faster. The signals are therefore surfaced much earlier and the anomalies are highlighted automatically.
And then your conversation changes completely. You get more time to challenge management, more time for scenario thinking, and more time for proper engagement. That’s where it allows for intervention before things go properly sideways.
So yes, less time reporting, more time doing something more useful with the outputs. And that’s exactly the ambition.
I would add one more thing. It’s not just more thinking, it’s better-timed thinking. In a portfolio environment where you are managing risk, timing is often everything.
The Finance Ghost: Yeah, I really like that. Judgement work versus mechanical work. I think that’s exactly the point. It’s not just the time saving, it’s what you can do with that time and where you can spend your time. So, very cool.
Rishi, I’m going to bring it back to you. Maybe last question in your direction. Something else that I understand about this product. Again, this has been built by Forvis Mazars, which I will remind everyone is a proper professional services firm.
A lot of the work you do is technical. So when you guys are putting together an AI-type product, it is with that application in mind. Stuff like scenario planning, but also IFRS, generally recognised accounting practice standards, all that kind of work.
There are applications here of this technology, right? Would it be fair to say that it’s been built for regulated environments?
Rishi Juta: Thanks, Ghost. Yes, that’s the essence of it. The point is not to build something flashy for loosely governed environments. The point is to build something that actually stands up inside serious institutional settings where control, traceability, governance and explainability matter.
Once you’re operating in those kinds of environments, whether it’s a financial institution, a DFI, or another regulated entity, you quickly realise that speed on its own is not enough.
You need speed with discipline. You need to know where the data came from. How it was transformed, what was validated automatically, what needed intervention, what assumptions were applied, what changed over time, and how you would evidence that to internal audit, external audit, or regulators if required.
That is where the more technical applications become important. Scenario planning is a good example. If you have a governed data layer and the ability to integrate external data, you can do much more meaningful stress-testing and sensitivity analysis. You’re not just looking at static history, you are looking at how the portfolio behaves under different scenarios, sector shocks, liquidity conditions or broader market stress.
Then you bring in things like IFRS 9 and GRAP 104. The point is not that the platform replaces professional judgement or accounting standards, but it gives you a much stronger information base for those processes.
If you already have governed data, historical patterns, covenant status, commentary and external signals in one environment, then your credit assessment, impairment thinking, expected credit loss and related reporting become much more robust and much faster. This has clearly been conceived for environments where governance is not optional; it’s part of the operating DNA.
The Finance Ghost: Shane, maybe if we then bring it home with you. Who do you want to be speaking to this product about, at this stage in its development journey?
I understand you guys have made a lot of progress already. I’ve seen it myself. I would imagine it’s at the period now where you are open to conversations with particular clients who are maybe looking for solutions like these.
So perhaps just to bring us home, if someone could phone you two minutes after they listen to this podcast or perhaps send you an email, who would you want that person to be? What are the conversations you want to be having?
Shane Cooper: I think it would be any executive who is sitting in an asset management or a DFI environment, who’s experiencing corporate pain at the moment. It’s the visibility of your portfolio. Are you uncomfortable with the degree to which you manage risk in your portfolio? We’ve got a solution for you.
The Finance Ghost: Very nice. I like that.
Gentlemen, I’m going to leave it there. I think this is pretty exciting, and again, I’ll remind listeners that this is a software solution being built by Forvis Bazaars. So congratulations to the team on that side. I think this is quite innovative and probably somewhat unique within these professional services environments.
I really like what I saw in the demo that you showed me. Obviously, it’s still in development, but there’s a lot of cool thinking there, some great visualisations, and I can see the use case.
So I will include in the show notes ways for people to contact both of you.
From my side, good luck with the ongoing development of this thing. AI is such a fast-moving space that I suspect we’ll be checking in on this technology in a few months to come, perhaps. Hopefully, by then it’s already being used in some really exciting applications.
So Shane, Rishi, thank you very much for your time today and good luck with the ongoing development of this thing.
Shane Cooper: Thank you Ghost, great to be with you again.
Rishi Juta: Thanks Ghost.
