Procurement Matters podcast: Episode 5 - Unleashing GenAI on procurement

Mun-Li Boswell: Hello and welcome to the latest episode of our podcast series, Procurement Matters. 

We've done quite a few of these now, so hopefully you've had a chance to catch our previous episodes. But if this is your first time joining, what took you so long? No, I'm joking. Welcome. We're thrilled to have you. My name's Mun-Li and I'm part of the procurement and commercial practice here in Baringa.

We have covered all manner of subjects from procurements role in the ESG agenda to how we attract the next generation of commercial professionals. I am particularly excited about today's topic, as it's another one that transcends professions and is much larger than just a procurement matter generally.

Yes, that's right, folks. We are bringing our wonderful profession kicking and screaming into the 21st century… For those fellow procurement geeks out there, the threat of technology, whether that be procurement-specific or broader, has been emerging for as long as I can remember. And for those of you who've been listening, I've been in procurement for a long time. The promise of an easier, transformed procurement life has just been around the corner for a while now, it feels. 

Well, the exciting news is we no longer need to talk in theoretical terms. It isn't a question of if or indeed when… It's more about how. How will we navigate the disruption it will cause and ensure and ensure that we are informed and responsible buyers of it. As always, I'm joined by experts to help me unpack this, because let's face it, if it was left to me this whole episode will be about chat and how it's actually created. And actually chat is just part of the latest wave of GenAI and has been hugely successful in making it much more accessible and whilst not perfect, is actually cheap and fairly easy to use.

So, let's introduce our fabulous guests before I digress into holiday chats. Emma Parr is one of our digital procurement experts here at Baringa. I'm always slightly in awe of her knowledge, I'll be honest. Emma brings her view of the impact of all of this on our procurement profession, and we're also joined by the rather wonderful Costas Gavriel from our Data, Analytics and AI practice course that has been there, done it and has the war stories. He has a rather envious job, part of which is delivering machine learning solutions to our clients. I might need to commission you Costas to help me keep up with my teenage children. Not going to lie.

So welcome, I have no doubt that I haven't done either of you justice in my entirety. So how about you tell us a little bit about yourselves and we'll then dive straight into it, if that's okay? Emma, can I start with you? Hello. Lovely to see you.  

Emma Parr: Hi. Thank you for having me. I am one of our digital experts here at Baringa. I spend most of my time helping organisations either procure technology that will help them transform across their organisation, or feature technology that will help most procurement function also transform how they implement and get value from that. And so those last ten years has been spent procuring safe solutions and whether they're end to end big platforms in the procurement space, or more widely or the best of breed point solutions that are coming up, it's all been around SAS and solutions.

So, I think I personally find the hype around GenAI really interesting because it's a very different proposition. It's not an end-to-end platform that you implement with a traditional methodology. It's a very, very different proposition. So, it's one that I find hugely exciting, and I think as a premium fashion rational looking towards the next ten years. And to your point, I think it's not just round the corner. I think it is starting to make an impact already and it's around how we harness that opportunity as a procurement profession.  

Mun-Li Boswell: I couldn't agree more. Emma, thank you. Costas lovely to see you.  

Costas Gavriel: Well, thanks for the invite.  

Mun-Li Boswell: You're welcome.  

Costas Gavriel: I'm Costas Gavriel. I'm part of the data team in Baringa, and I've been working in the machine learning and AI space for more than a decade now, helping clients across multiple industries with solutions oriented around AI. I'm not going to pretend I'm an expert in procurement at all, but I've worked with quite a lot of you guys in different engagements understanding how this works. And as Costas has been saying, it's a very fascinating industry at the moment. So much investment going on, billions of pounds and dollars going into AI in general today. It's very exciting how the industry is going to be evolving over the next few years. And they basically have people as well.  

Mun-Li Boswell: And we're very lucky to have you navigate or help us navigate us through all this. So okay, Costas, break it down for us. We talk a lot about AI… I've heard a lot about it. In fact, I was reading a BBC News article this morning about it. But what exactly is GenAI?  

Costas Gavriel: A really good question. So generally, generative AI, as it is known, is essentially the area of artificial intelligence that focuses on creating content that looks new and original. So, it is essentially about generating text, video, audio that is pretty much indistinguishable whether a machine has created that or not. 

And it's a very, very interesting one. I mean, as you said at the start in the intro, has really exploded over the last one two years with the introduction of ChatGPT, which essentially made a GenAI model accessible to everyone around the world. And with plain, simple English sentences, you can start constructing a machine to generate text and paragraphs or images for you, which is quite an impressive milestone for overall. And yeah, it is really, really interesting. The whole space.

I've seen quite a lot of companies already taking advantage of Generative AI. My favourite example I think we discussed in the past is around Coca-Cola training, a general advice model with all their marketing material in the past and then essentially gave access to people through it, a chance to be able to go and say, “Oh, give me a spaceman holding Coca-Cola bottle”. We just made the accessible to billions and billions of people across the world to be content creators for Coca-Cola. Just fascinating.  

Mun-Li Boswell: I mean, that is amazing, isn't it? A spaceman holding a Coca-Cola bottle. That's brilliant. And we recently ran a roundtable on this topic. And actually, Costas was equally excited and animated about it, which was brilliant. But I felt like there were two camps there. There were those that were super excited, couldn't wait to adopt it, and those that were perhaps a little bit nervous. We've got regulation already on our table. It's increasing. It's ever changing where we get bogged down and stifled by the bureaucracy, I guess. What are your thoughts on this?  

Emma Parr: I think as with many revolutions in the professional world, it's an opportunity and a threat. So, most of the organisations that we were lucky to have with us at our last roundtable, 60% of them said that their organisations were already using AI. And instead of that, we asked: Do you play a role as a as a procurement team? Do you play a role in enabling that for your wider organisation? And 47% said ‘somewhat of a role’,  13% said ‘yes, we play a significant role’ and 40% said ‘no, not yet’.

So I think it's definitely here in terms of organisations are using it, I think procurement are in its infancy as a profession in terms of how it's starting to come to that table and how it's playing a role in enabling it. And as you said, there's so much to learn and grapple with around it. The opportunities are fresh in. 

You can play a key role, be at the forefront of your organisation, enabling our AI to drive the strategic agenda of your organisation. And you can use utilise in the procurement function itself. You can drive efficiency effectiveness through use of it. You can transform the way you write contracts, you can transform the way you run sourcing environments and completely reducing, or in some cases eliminating, manual intervention needed.

Now it's not just as easy as flipping a switch and it goes. Whilst it might not be like implementing a solution where it takes a year of methodology, you still have to put guardrails in place and that's really where the kind of the considerations come in around it is how and what role the procurement playing to put those guardrails in place.

So, there's many considerations when you contract AI, not just from a commercial perspective, but as you alluded to from a regulatory and legal perspective as well. Procurement plays a key role in liaising with legal to make sure that there are sufficient contractual guardrails in. But they've got to make sure that the use of the contracts that procurement put in place and the general suppliers that they put in place that they can put enough control around them. So controls around demand are sure, we will come on to talking about the risks around consumption based models and more generally, but also things like the data models.

So, if you're buying a GenAI tool and procurement, you're supporting your organisation to procure that, you're going to be really considered around what data that gen solution is going to use. And you've got to really be thinking about what the data, the solution was trained on as well… I've been in numerous instances now where GenAI models have been criticised for the data that they've used to train themselves because it can build in bias. So the opportunities and threats are why there's so many considerations. And I think the kind of sentiment that we got from the organisations that we've engaged with so far, I've seen there's just lots on the horizon about it and there's so much to consider.

When do you start? We talked about what are your key priorities and procurement is generally, you can't do it all on day one. So really where do you focus as a procurement team to both enable the organisation AI function but do it in a responsible manner?

Mun-Li Boswell: Yeah, it sounds really, really exciting. There's lots of opportunities. Appreciate there's going to be threats in terms of making sure we are responsible and safe in the way we execute and implement all of this. But Costas, what would you say to those that are expecting this to happen tomorrow? For example, really excitedly, let's say the world is going to change tomorrow and my life's going to get easier, everything's gonna be automated and I'm going to be able to tap something into ChatGPT and concentrate on the really interesting strategic stuff instead?

Costas Gavriel: There’s quite a lot of people are very excited about this whole technology. There's a level of maturity and readiness currently in the market from a technology point of view. So, the technology can support this kind of automation of workloads and just optimising processes or closing out time or making more efficient in terms of creating new content around it. So there are areas where it can be used to drive efficiencies. And I think that's what people are excited about… On the other hand, it's still at its infancy as a technology as well. 

So, there's quite a lot of GenAI models out currently that have quite a lot of risks associated with them. You have this concept of hallucination where models can just sometimes make up things along the way, which introduces a risk on how much you can depend on the outputs of those models, and how much governance and quality assurance you can put on. So, it is very interesting because we're at the point where technology is running very, very, very quickly. 

I've been talking to a colleague of mine and last week and they've been saying this about 3 or 4 new companies popping up in the market in the last month. And it's really hard to stay on top of the whole evolution of technology. At the same time, we've seen that again and again and again. It takes time for any of those changes to impact any of the and the actual implementation and change the way that businesses are transformed. So it's a very interesting time. 

But back to GenAI, a question that you mentioned earlier was this consumption based model. I mean, this is something that I see quite a lot with cloud providers, with quite a lot of tech applications and tools. They are all moving towards a consumption type of model. So, it is very interesting because the whole AI is probably moving to that type of model as well. So, how do you think that's going to change the kind of point of view from a procurement point of view, or the purchase partners that people need to be thinking about? And at the same time, how does this whole proliferation of tools, again, impacts the whole thinking around this? It's going to be very interesting.  

Emma Parr: And there’s, I think consumption models while they're not new, to your point Costas, they are proliferating. I think the considerations for procurement are similar. They're similar principles to what procurement have already been utilising, which is a virtue. There's a commercial aspect around negotiating consumption-based deals. So, you want to make sure that you've got even mechanisms as simple as good volume discounts, you've got stage pricing, you've got control from a contractual perspective around that consumption. You know, procurement, they're putting that in place for a number of years now already around things like licenses. Yes, it's a different model, but they've had to put guardrails in from a contractual perspective about how to control that spend upfront and how to negotiate the commercials around it. 

I didn't mean procurement have a role to play once contracts are signed and suppliers have been bought, how they enable the organisation to control that spend. And lots of organisations that we work with, they put in place fantastic contracts, they've negotiated a great deal, but actually they, they can't control the demand across an organisation, and I imagine with GenAI the demand is quite proliferated equally across organisations, lots of different pockets of different businesses. 

Will we start thinking about how GenAI can transform what they do, whether that's in marketing, whether it's in production, they'll all be having the same thought process. And so how do procurement stop a situation where an organisation has bought a million different GenAI tools and actually they're missing a commercial opportunity to really rationalise and consolidate that spend, which again is it is a really traditional procurement concept. 

I think it's just around applying those concepts in a slightly different landscape. And I think the other challenge here is the pace is high. So, you know, we related to a number of times already, things are moving really quickly. And so things some of the traditional procurement tools like preferred supply less or traditional category strategies, these will become obsolete very, very quickly. So it's really a game of pace here around keeping up with both the opportunity and the guardrails that they need to put in place.

Mun-Li Boswell: I mean, that's fascinating. I the pace is quite interesting to me because, for example, I'm thinking along the lines of my daughter as she's currently starting her GCSEs and in theory could type into ChatGPT ‘Could you write me an essay on World War Two?’ and how do we deal with the possibility of a bid being written by GenAI, for example?  

Emma Parr: Yeah. And I'm sure it's happening already…. My view on this is that is as a grow we should be focusing on sourcing selection processes and efforts on making sure that we're asking questions that cannot be easily replicated through GenAI. I think sometimes there can be a tendency through sourcing processes to ask generic questions that might be susceptible to, you know, as a supplier. What's your view on access? How would you deal with Y?

Yes, there are probably some instances that they can they can use GenAI to help write those answers. But I think as a procurement professional, if you keep your questions really targeted on specific use cases specific to you as an organisation, not asking more general questions, I think the opportunity for supply used in AI is slightly more limited because there won't be a ready-made answer on the internet already about how to do that. They will have to really think and consider about how they tackle your question. And so I think it's much more around, you know, drawing the parallel to exams, you never score full marks an exam if you just regurgitate the textbook. So remember how you apply that thinking. 

And I think it will be the same for the supplier. It's around how we test their ability to apply their specialism to the context of you as a buying organisation. But yeah, it's definitely a risk and I'm sure it's happening, it's happening already. And you know, with ChatGPT for example, Costas you'll know better than me on this one, but you have to really consider about where you're putting that data in. So if you're using open cloud and Chat GPT and putting in confidential information into that, you're feeding that into the web of open AI accessible which activities organisations have to be really careful around if they're using ChatGPT they're doing that in a secure cloud environment, where their data is safeguarded and not being shared unknowingly across the world.  

Costas Gavriel: Absolutely. It's one of the main risks, at least from a technology point of view. How do you control that? The privacy and security of data when people are using ChatGPT or any other similar technology to GenAI? So yeah, as you said, it's so accessible these. Anyone can put their credit card on the web and just get any one of those tools and use it. But then where does your data end up being used afterwards and where does it go? And, even within an organisation, even if you train your own model, how do you prevent leaking of information that is only accessible to a certain number of people? It's a very interesting topic. It's not yet fully solved with technology.  

Mun-Li Boswell: Gosh, I mean, we could spend all day talking about this, and every day is a school day where I'm learning myself. But you know what? We've run out of time and I'm it feels like there's a there's a sort of call to action in that how are we going to keep up with this technology to make sure that we're harnessing it? 

To your point, Emma, but doing so in a responsible way so the proliferation doesn't become chaos and almost dangerous… So fascinating.

Thank you both for your time. For those of you who've listened before, I always end on a more personal note. We are recording this at the end of the calendar year, so I am going to ask the obvious question and it's going to be a tricky one. People say, you know, brace yourselves. Emma, what's your New Year's resolution? I know it's a little bit premature, but we've got we've got a prep for 2024.

Emma Parr: My New Year's resolution is to keep a New Year's resolution. I have a terrible track record of being one of these people that does it for January. And then it and then it falls by the wayside. So I think my hope, my ambition, for that New Year's resolution would be to do more exercise. I've been slightly lazy this year and definitely feel better when I'm a bit more diligent. So, mine is to be a bit more active, especially on the horrible wet, dark days when it's very tempting just to stand and sit, and to stick with that throughout the year rather than just have January in the gym and then go back to my usual ways.

Mun-Li Boswell: I'm going to check in with you regularly to see whether you're actually sticking to that. So look out for a little call in January! Costas?  

Costas Gavriel: I'll definitely do more exercise as well, but I think my New Year's resolutions to do a lot more travelling. I think with Covid has changed quite a lot of patterns. I haven't done a long-haul travel for three years now, so I'm very excited to book something in the new year, probably Latin America, scope that side of the world.  So that’s about it.  

Mun-Li Boswell: I can highly recommend and I to write an itinerary for you.  

Costas Gavriel: So great idea.  

Mun-Li Boswell: Thank you both for joining us. As always. If you have any topics for our next podcast, please do get in touch. But in the meantime, thank you for listening and we'll see you again soon. 

Emma Parr: Thank you.  

Costas Gavriel: Thank you.