Episode #13
Peter Bracher is a highly experienced food auditor and compliance expert having started out in life in the UK as a government environmental health officer, before moving into retail for Tesco then Asda.
Having left Asda, Peter helped the richest man in Asia Mukesh Ambani set up a retail empire in India, before moving to NSF and setting up its Asian Pacific operations.
He is now based in Thailand and is helping Supply Chain In-Sites (SCI) marry the world of compliance with the use of big data and algorithms.
Peter sees the on site audit by an auditor as the last resort. There is so much data available now that sending someone on a plane is hugely inefficient, and just encourages gaming of the system.
Instead he cites a prawn farm where satellite imagery, combined with verifiable data such as purchase receipts for animal feed, and a simple calculation reveal the maximum amount of shrimp that the farm can produce. If something doesn’t add up, then further inspection or spot checks can be carried out.
He foresees a future where everyone knows exactly what they are eating and drinking and can be confident it is safe.
He thinks we’re gonna get to a point in an interconnected world, with wise use of data, where we use algorithms to replace the routine bits, but you still have oversight of a person. AI and RI – real intelligence and artificial intelligence – working together so that you can be sure that what you buy is going to be safe.
He can’t think of anything much more valuable than that.
SPEAKERS
Dom Burch:
Welcome back to the ubloquity podcast with me Dom Burch. This is a podcast where we get to speak to thought leaders from around the world. I’m absolutely delighted this week to welcome Peter brasier Not least because we used to kick around Asda House in Leeds together, but now he’s far away in the exotic world of Thailand. Peter, welcome to the ubloquity podcast.
Peter Bracher:
Thanks a lot. Dom. Yeah, it was like three countries ago for me. It was a good time that was a high energy company in those days.
Dom Burch:
It certainly was what have you been up to since then? Where’s that journey taking you because you’ve been in the world of compliance and regulation. And so what you’ve been up to?
Peter Bracher:
Yeah, well, I went from from Leeds to Bombay. And I help Mukesh Ambani. He’s, he’s the richest guy in India, I helped him set up a retail company, which was fun, but seven days a week working 16 hour days, Indian style. I’d had enough after three years. So then I helped NSF, the American company set up their operations throughout Asia Pacific. So that’s like a food safety certification auditing business, based in Thailand, but working throughout the whole Asia Pacific region. And that was good fun. But now I’m working with SCI and I’m helping them to develop alternatives to traditional auditing.
Dom Burch:
And what does that look like now, because I guess the world of technology has moved so much in the last 10-15 years, you know, we’ve been looking at big data, we’ve been talking about algorithms, and we’ve seen, you know, technology companies like Facebook just completely revolutionise how digital advertising works. In the world that you come from, you know, regulation compliance, and how are seeing that world develop, are we sort of crossing point now is is it really getting to that point where new ways of working have really beginning to accelerate?
Peter Bracher:
They are Dom, it’s just starting, I think regulators being government tend to be a bit behind the curve. But Dubai is one shining example where they’re, they’re really at the forefront. Dubai have linked all their food safety enforcement systems into one single online system. So every food business in Dubai must be registered and must be online. And they have to have like a person in charge of food safety trained, and their training records are online, has to be on site and like login, you know, I’m here, I’m now the person in charge for the following few hours. So they know in every food business, who’s the boss at any point in time, they also know what they’re producing. And they can just at the click of a button, they can do a recall. And they can do food poisoning investigations. They can target their inspections. And they’re also starting to use data from from the training records, which are online, the complaints records, the food poisoning allegations. It’s all online. So the Dubai enforcement officers target their inspections. It’s no longer, well, in my day, I was a an environmental health officer on the beat years ago. I mean, we’d just think, well, where are we going to go today? You know, we haven’t been been to that Chinese restaurant for a while, let’s pop in. It’s now targeted. I think they’re leading and I think other enforcement agencies will
Dom Burch:
And as you see those sorts of things beginning to follow. occur, and it must be refreshing for you, you know, I’ll get back to the retail days. And we’d have sort of by exception reports wouldn’t we? We’d have a store estate of maybe four or 500 shops, and you’d be looking at various measures wouldn’t you? And going, the last time we went and inspected them six months ago, they performed badly, here’s a little black mark. And that would, to some extent, allow you to sort of be a bit more surgical, a bit more forensic about targeting your limited resources, and trying to fix where the real issues were, rather than just sort of, as you say, randomly popping up to the next shop and hoping that you were going to find something, or conversely hoping you didn’t find something. Data and technology and algorithms, I guess, are really beginning to do some of that hard work for us.
Peter Bracher:
They are yes, and as well as targeting limited resources better, like you say, it’s also more positive, it drives the right behaviours. I mean, I’ve I’ve made a career out of enforcement, and then and then traditional audits. But I’ve always had a suspicion that it really wasn’t the best way. It creates gaming. When you turn up to a supplier or a manufacturer. There’ll be a QA person who who knows exactly what you’re going to look for, knows the standard. One of them even said to me, he has what he called the yellow brick road. And he leads the auditors on this yellow brick road and he’s carefully placed a few a few minor nonconformances you know, wash basin with no sanitizer. He’s also placed his best trained people who can give the right answers. They’ve checked all the records that they know the auditor is going to look at. It’s just a game. And with restaurant auditing in the operation I set up in Asia Pacific, we dominated restaurant auditing. So we’re doing 65-70,000 audits a year for like all the big QSR businesses and the retailers. But they can predict when we were coming, there was there’s one example where the area manager even employed a private eye to follow the auditors, and then they look at, oh, this is the path, we’re gonna go to this restaurant next, let’s get everything ready. This is just wasted effort. So what I was looking for is something that can not dilute that responsibility of the management and the owners. Stop it becoming a battle between auditors or government inspectors and the local management, and actually make it something that will benefit customers. So that so that was the aim of what I’ve been doing with SCI.
Dom Burch:
And talk me through an example then let’s use a prawn farm in Thailand, you know, the start point might be desk research, it might be good old fashioned spreadsheets and data analysis and asking questions of a supplier. But then over time moving that into something that does create an algorithm. And then potentially you’re going to get what machine to machine data feeds, APIs coming into a system. And your ability then to be far more proactive rather than reactive.
Peter Bracher:
The way I see it is the on site audit by an auditor is the last stage it’s like a last resort. There is so much data available now that sending someone sometimes on a plane or go one country to another is hugely inefficient, and just encourages this gaming. Instead, we can use the big data that’s available. So yeah, a prawn farm is a good example. What you want to know is, first thing is are the prawns actually coming from that farm. And what happens is some countries, they run into problems with antibiotics or disease and their exports get stopped, or the big buyers stop buying from that particular country. So all they do is they ship it through another. So you don’t know this prawn farm in Thailand could be selling Indian shrimp for all you know. The first thing you could do is satellite images, you can look at the size of the pond, how many square metres surface area of pond can you see. And then there’s just a direct mathematical relationship, how much shrimp of the particular size you could possibly produce from that farm is directly related to the size of the ponds. So that’s the first thing you can look at without even leaving your desk, you need to obtain verifiable data. So you get a feed from the farm. And you can do this through a simple free to download mobile phone app. All farmers now have mobile phones, smartphones, India, Thailand, they’re universal as are improving cellular connections. So what you do is you ask him rather than do a self audit, where he ticks everything to say everything’s wonderful, you ask him verifiable questions, you say, how much feed did you buy last month and upload your receipts, so that’s verifiable. And you can see, okay, cross reference, size of pond, average size shrimp they’re selling, how much feed have they bought, and then have a look at how much you’re buying. And the way the market works now, is food is actually short supply worldwide. So a buyer will typically buy everything that small farm in Thailand produces, so you know how much shrimp they’re buying. And you just cross reference that. And as you say, this is the work of algorithms, you don’t even need people to do that. So you get a dashboard that will just alarm out and say This farm is supplying more shrimp than is possible for that size of pond from the satellite records and for the amount of feed they’re buying. So you’re checking the data, then you can ask other questions about numbers of staff and ask to see payroll, which also helps check social compliance, you can check for things like that, because there should be pay records that show how much you’re paying that should be above minimum wage. And you know, another algorithm how many people are needed to produce that much shrimp. So again, it’s cross referenced. So use data first, and you use a site visit, just to check that your data is right, someone’s not being really clever and trying to fool you. So it’s not an audit, you send someone for spot checks, just they have the data in their hand and say Okay, does it all look right on the farm?
Dom Burch:
And I suppose once you start getting that verifiable data and you know, it sounds really similar to the world we’re operating in where we’re trying to say, what data feeds exist, which are digital, which are held centrally, which are verifiable. Can we use multifactor verification so can we use two ways to kind of look and pincer movement if you like that data, just to say, you know, let’s use a satellite image, but as less also do a spot check. And as you start to build up confidence and trust, I guess in those data feeds, then it’s about where do you put that data and do you see a future then where your world starts interlocking with the world of the blockchain, where you start to then say, let’s store that data in an immutable format. And only give access to those who need to have access to it could be government could be customs, could be auditors, or it could even be the customer couldn’t it? Could be the retailer, who says, Actually, I want a finger on the button right now to say, are all of my prawns coming from a sustainable source?
Peter Bracher:
that I think is really exciting area, because if you put the data in the blockchain, and you have it linked to the product, I’ve seen, I’ve been around shops now where customers can scan a barcode with their mobile phone, and it’s supposed to tell you where the clothes come from, how much carbon or water was produced, but you don’t know if it’s right. But if it’s verifiable data that’s been checked in the blockchain, then you can trust what you see. I think it also is an advantage to businesses if you’re running a restaurant that your customers expect ethical standards or a hotel. And if you want to say, right, this, the shrimp has come from this farm where they pay a living wage to all their people. And our checks have shown that, you know, they’re not using antibiotics again, you can check growth rates, that gives a lot of faith to the customer. Now the customers can trust the data and they can see the whole chain themselves all the way back to the farm. If they want.
Dom Burch:
I guess over time, then that also begins to open up more insight, doesn’t it? So once you know that, that data is verifiable, it’s had multifactor verification, you can trust it. Over time, I guess that builds up, you know, a history for that farm for that product for that country for that supplier. And gradually you know, fast forward five years, 10 years time, the richness then of that data and been able to overlay other things, environmental factors, weather factors, economic factors, even because we’re in a time now, aren’t we, Peter, where, you know, economies are really, really struggling cost pressures, huge. I mean, we’re seeing potentially up to 20% inflation in food, which is, you know, for a generation or two, just unheard of, we’ve never lived through a time like that most of us on the planet, that’s going to put huge pressure, isn’t it onto the system into the supply base? And I guess, you know, from your experience, when cost pressures are really, really extreme. Does that lead to more corner cutting, you know, less compliance, you know, how do we overcome those things?
Peter Bracher:
I’m old enough to have been an enforcement officer at a time when we had bad inflation. And the UK had a recession. What it meant was, we were looking for food fraud in that time, the early food safety legislation was about adulteration and fraud, because people were simply trying to cut costs. And watering down the beer was the first cases. But that’s coming back. I mean, in China, just last year, there was a company actually making artificial eggs. And you think the the effort required in making an artificial egg, surely it would be easier to let the hen do it, but actually, it’s cheaper. So that sort of behaviour we’re going to see more of and also cutting on on staff, you can see post COVID, hotels and restaurants are really struggling to get their act together. Because they lost a lot of the best staff. Now they’re trying to recruit at the lowest possible cost. So one of the other things we’ve been doing is also simple available data is you want to know if people are properly trained in food safety in restaurants, for example. So what we do is we ask people to upload their training records, and that’s online. So we get a feed from that. But we also look at the HR records. So you can see how many new starters has a restaurant had. And you can cross check that to see if they’re properly been trained in food safety. And then you were talking about data before Dom, it’s big data, we will be able to see exactly how many people are needed to run a particular sort of restaurant in a particular region. So, inner city one’s near business premises have certain peak times, they’ll tend to need 50% more staff, than one that has steady trade in a shopping mall, for example. So you can see have they got enough staff? And have they been properly trained, including the new starters. And that’s a vital element in food safety. And by using data you don’t need to send someone in to see if someone’s cost cutting, because that’s cost additive. But an on site audit is around 250 pounds for a restaurant and about 1200 pounds for a supplier. Now if you could cut that down so you only need to maybe do one a year to check that there’s nothing really odd going on just to verify that your data flows are right. That’s a huge saving for companies. One of the biggest QSR companies spends more than $50 million a year just on auditing. So in an inflationary times, this is a great opportunity to not just save money, but replace it with a better system.
Dom Burch:
What are your hopes for the future Peter, because, you know, let’s be optimistic for a moment and allow ourselves to dream of this kind of Nirvana that’s just out of reach at the moment. But you know, as you see the world shifting so much, and, you know, I talked somebody else about this recently, and it always feels like every generation always feel like they’re on the cusp of this amazing, you know, revolution in the world. But it really does feel like that at the moment. What are your hopes for the future? What What can you see coming around the corner that, you know, us mere mortals haven’t yet spotted?
Peter Bracher:
I think it’s absolute confidence in what we’re eating and drinking. At the moment, we all suspect you know, there’s probably pesticides on the fruit and vegetables, but we don’t know. You know, we hope that our water is safe. And there’s nothing more important, nothing more important than what we eat and drink is going to keep us healthy. So I think we’re gonna get to a point in an interconnected world, with wise use of data, it can’t be completely automated. The systems that we’ve been developing with SCI we start off using experts, people do desktop reviews, then we study what they do. And then we just use algorithms to replace the routine bits, but you still have some oversight of a person. But this cross reference verifiable data in a blockchain is the oversight of a real person. So you’ve got AI and like RI, you know, real intelligence and artificial intelligence working together so that you can be sure that what you buy is going to be safe. I can’t think of anything much more valuable than that.
Dom Burch:
Absolutely. Well, listen, Peter, we’ll have to leave it there. But thank you so much for for joining us on the podcast this morning. And this afternoon, no doubt where you are. It’s been an absolute pleasure catching up with you. Let’s not leave it quite as long to catch up next time around.
Peter Bracher:
Yes, I’ll let you know when I’m next in the UK Dom. Great to catch up again.