Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
EP 24: Transforming Poultry Health – VOC Fingerprints, Early Alerts & On-Farm Monitoring
2 Playsin 12 hours
In this Poultry Network Podcast, hosts Tom Woolman and Tom Willings speak with Angela Curtis, CEO of RoboScientific – whose presentation at the BFREPA conference created a real buzz – about how “electronic nose” technology could change flock health management.
RoboScientific has spent around 10 years developing a system that samples the air in poultry houses and analyses volatile organic compounds (VOCs) – the chemicals that make up smells – to detect disease.
Each shed is monitored using a compact box (about 18 inches cubed) containing 24 sensors.
Different combinations of those sensors respond to different conditions, producing a distinctive “digital fingerprint” for specific diseases or challenges.
Rather than continuous streaming, the system draws shed air for around an hour, often between 5–6am, pulling roughly 3,000 litres across an adsorbent pad before analysis.
This concentrates VOCs from across the whole building, giving a much more representative picture of flock health than boot swabs or individual bird checks.
Once processed, results are sent to the cloud – or via Bluetooth where connectivity is limited – and presented to the farmer through a simple dashboard or smartphone alert.
A major advantage is that VOC changes are not dependent on visible symptoms.
RoboScientific’s broiler trials suggest the system can identify disease one to two days before clinical signs, and in the case of viruses such as Newcastle disease or avian influenza, potentially two to four days earlier.
Alerts are designed to be practical and actionable for farmers, for example “likely infectious bronchitis – check this shed”, or a general prompt when something unusual is detected.
A deeper veterinary dashboard is also in development, enabling flock-history analysis and tracking of disease progression, such as the transition from coccidiosis to necrotic enteritis.
The technology has shown strong performance in broilers and is now moving into layers.
Early work is focused on diseases such as erysipelas and Mycoplasma, and RoboScientific is exploring whether VOC patterns could even detect red mite infestation levels. APHA trials are planned to confirm early detection windows for key notifiable diseases.
Commercial rollout for broilers is expected by the middle of next year.
Work in layers, hatcheries and rearing units will follow as disease signatures are mapped, with layer deployment anticipated around 2027 once validation and on-farm trials are complete.
Units are expected to cost £7,000–£8,000, with one or two typically required per broiler shed depending on size and airflow.
Angela encourages producers and vets interested in the technology to get in touch or visit RoboScientific’s website or LinkedIn page.
The episode highlights how rapidly developing sensing technologies could reshape flock management, enabling earlier intervention, better decision-making and a more complete understanding of flock health than traditional sampling methods can provide.
Recommended
Transcript
Introduction to the Podcast and Guest
00:00:15
Speaker
Hello and welcome back to the Poultry Network podcast. I'm Tom Woolman. And my name is Tom Willings. Hello, Tom. Good morning. Good afternoon. Hello, everybody.
00:00:26
Speaker
Thank you for tuning in again. Whatever time it is you're tuning in, wherever you are, um whatever you're doing. um Yeah, we're really excited this week to be joined by the CEO of Robo Scientific, Angela Kursis. Hi, Angela.
00:00:41
Speaker
Good morning. Well, afternoon even Yes. I'd say yes. It loses all meaning, doesn't it?
00:00:50
Speaker
I'd like to think as well we're going out across all sorts of different time zones. You know, there are people in Siberia listening to us um and and and people in in Mexico sat on the beach on their holiday. um So, yeah, wherever you are in the world.
Angela Kursis at Bifrepa Conference
00:01:05
Speaker
So. so we We're hot off the Bifrepa conference um you yeah speaker speaker list, Angela. And and i'm going to I'm going to say first off, from where I was stood um stock still all day, there were numerous people that that came up to me to say, did you see the robo-scientific presentation, Tom?
00:01:29
Speaker
Did you manage to get in there? And the answer, no, no, no, I didn't. But no one came for the rest of the day and talked about any other presentation that had been delivered. and And as we've already covered in a previous episode, I would say that the lineup was was a very, very strong one. Some really good um people that I would have loved to have listened to. But from everyone I spoke to, your presentation was the one that that that that got the most positive remarks, got the only positive remarks. But um yeah, we thought it would be fantastic to get you onto the podcast, just in case anybody that wasn't at the event um didn't get a chance to listen to you. And and we can you know give this as much airtime as possible.
00:02:11
Speaker
Well, that's really exciting to
Detecting Diseases with VOCs
00:02:13
Speaker
hear, I must admit. I think the thing that we give is we give hope. And it's always good to be the person giving hope, not reporting on the latest avian flu figures or whatever it may be, as important as those things can be. so yes but let's Let's jump into it. alllan just and Sorry, Tom, I apologise. I was going to say, should we jump into it and just start from the the the very, very top of, you know, what what are we talking about? What is it that robo-scientific do?
00:02:40
Speaker
Well, we're very excited because finally we are at the point where we can say that we have got working prototypes about to go on farm. And I say finally because it's taken us 10 years to get here.
00:02:56
Speaker
and What we do is we take volatile organic compounds, VOCs, and we sniff the air in a chicken shed to see what VOCs are present and to see if we can find a relationship between the digital fingerprint of the VOCs and a specific disease or condition. So this technology is used in quite a lot of other sectors already, isn't it, Angela? Oh, it's as old as the hills. if if you are If you were as old as I am, you would have watched George III, that film about when he had that awful urine issue.
00:03:33
Speaker
disease and and all that the doctors were doing was sniffing his pee and effectively I hate to say it that's what we do which we just like ah an automatic dog we go around sniffing the air to see what's there and I didn didn't catch George III but I've been into plenty of chicken shit and used my nose um and can tell the difference between things going well and things not going quite so well but yeah we're breaking new ground with the George III reference
00:04:04
Speaker
Exactly. So, yes. So we're volatile organic compounds. It's the chemical components that are in the smells that are created by the host organism. So that can be a human, an animal, bird, a plant.
00:04:22
Speaker
um And the patterns that are being emitted are the things that we've been looking at to see when we can identify the specific pattern of the VOCs that are coming out that relate to a specific disease. It's a bit like a digital fingerprint.
00:04:40
Speaker
And then how many of these sort of sensors are we talking about in ah in a house or in you know take a typical broiler shed? um you know how How big a kind of investment in equipment are we talking about?
00:04:54
Speaker
Well, I get a bit... It's very difficult to describe because actually we're not talking about sensors per se. It's not like, um I don't know, little CO2 sensor in the in the the roof or anything, but it is a box. It's a box that's about 18 inches cube.
00:05:16
Speaker
I did tell you i'm I'm old enough to be not metric. And in that box, we have a sensor array of 24 different VOCs And ah they can they can maybe be 24 different VOCs or they can be a variety of say, eight, one VOC and two of another. It just depends what we're looking for so that when the when the air is passed over the sensors, all of the sensors react. we actually say, okay, we're looking for coccidiosis, say.
00:05:51
Speaker
So what we're going to look at is we're going to look out of those 24 sensors, we're going to look at, say, sensor 3, sensor 6, and sensor 18. And we're going to look at what the responses of those three sensors are.
00:06:03
Speaker
Or maybe we're going to be looking at, um I don't know, campylobacter, which is not something anyone worries too much about these days. But if we're looking at campylobacter, we could be looking at sensee sensor sensor 11 and sensor 20. It just depends. So that's how we get the differentiation. And it's how we're able to look for a variety of
Impact of COVID-19 on VOC Research
00:06:23
Speaker
diseases. But each of the 24 sensors is an independent sensor providing to the whole answer.
00:06:31
Speaker
And so and you've been trialing this technology, obviously, fairly extensively for a good number of years in both layers and broilers, then, is that correct? Not in layers. We're just entering the the wonderful world of layers. um And I must admit, I love it. I go to a laying farm and I come home with a tray of eggs. It's just wonderful. It's not worth an awful lot of money at the moment. Yeah.
00:06:56
Speaker
yeah yeah they don't last long in my house I have to say. but you've You've mentioned Coxie you've mentioned Campylobacter and obviously layers those two infections were probably not the top of the list of concerns for um for for farmers what what could you detect in a laying shed that would be you know most relevant to a laying farm?
00:07:18
Speaker
Almost anything, I would say, in time. At the moment, we're starting to look, um and the thing I haven't got in front of, we're looking at erysepolis because that's clearly something that's got a high mortality and it's um very important. Mycoplasma, we think we're going to be looking at that. I'm particularly excited because I have a meeting. This is my life. I have a meeting next week with an academic who spends his life looking at red mites.
00:07:48
Speaker
And we're going to discuss how we might be able to um sense red mites. We know we'll be able to see them because they're a parasite. Therefore, they're an organism. Therefore, they will be emitting VOCs. However, what we don't know is can we just see red mites or can we say if there are a few red mites hiding in the crevice or are there a shed full that needs some attention? Because I gather that is a fairly significant challenge when you get to about April or May.
00:08:15
Speaker
Yeah, i I would have put infectious bronchitis on the list. That was the other one, yes. Yeah, and and and as you say, red might but red might not necessarily being a disease, but you've already answered my question, really. And yeah, well, fascinating to hear how the yeah how the conversation with the academic goes.
00:08:32
Speaker
Yeah, well, the the the thing um that we have established and COVID remarkably was a huge help to us in this. One of the nice things that came out of the pandemic and who thought we'd be saying that was um the amount of work that went into um viruses and VOCs because as you were saying, Tom, there's an awful lot of work has gone into VOCs over the years. So it's it's not a brand new technology. um what we do with it is. um So viruses, although they're inorganic, they do change the way that we metabolize things. So the the way we metabolize the chemicals within our body changes the volatile signatures that we're sending out. And therefore, when we are looking for a virus in a shed, um we we can see that using the VOC signatures
00:09:29
Speaker
of the host. So that's how we do that. so And how do you think this this would work in practice then with this technology when it's up and running on farm? How would it work with the farmer and their management of the of the flock?
00:09:43
Speaker
Well, interestingly, we did a lot of work with pigs where we were working with a company using cameras for behavioural um indications. And what we discovered was if you're doing a camera um recording of of what's happening on a farm, you need an enormous amount of data.
00:10:03
Speaker
If we're looking at our VOC information, it's tiny packets of data. um because ultimately what the farmer needs, as I understand it, and as the ones that I've spoken to have said to me, is they need to know at the start of every day what's happening with their they' birds. If anything like avian flu is a real concern, then they probably need to know, say, every six hours. But they do not need to be told every 10 minutes that, yeah, we checked the birds, they're fine. um And it won't work anyway, because the key thing with VOC stuff, sensing is the sample quality. It's like any any test, it's only as good as the samples. So what we do is we sample the air, we gather in the air that's in a chicken shed for about an hour typically, and that is about 3000 litres of air in a broiler shed.
00:10:57
Speaker
And once we've got all the air gathered in, we've collected it onto an adsorbent pad, we're concentrating down all the VOCs that have passed through in that hour, We then um remove them from thee the adsorbent pad, pass them over the sensors to see what's there. That's how we get a much higher level of concentration of the sensors. It also incidentally means we get a much more universal sample. So rather than relying on somebody to wander through the shed with boot swabs or you know picking up maybe an occasional bird for culling, it and and it's a random choice,
00:11:35
Speaker
we're actually looking at the whole picture of what's happening in the shed. We're analysing it to see what patterns we get from that. And so so that's how it works. And so then once that's happened, all of the data is passed up onto a cloud, ideally, although I live near Norfolk and connectivity isn't the best.
Farm Implementation and Challenges
00:11:54
Speaker
So we're probably going to be working with Bluetooth as well. um And ah if we've got that level of connectivity, then the farmer will get it on his smartphone and or he'll get it on a dashboard. um So he will be notified. So, for example, in Broilersheds, which is where, as I said before, we've done most of our work, and what we're what we're seeing is we tend to sample at five o'clock in the morning. We finish sampling at six o'clock in the morning by five past six in the morning. If anything's um
00:12:23
Speaker
happening in the shed that needs to be notified, the farmer knows about it so that when the stockman goes in at seven o'clock, he knows what the status of that shed is. And is that recording always at the same time each day? Because I could imagine that the atmosphere is changing with different levels of activity or or birds doing you know particular functions at particular times that would again change change the atmosphere.
00:12:49
Speaker
ah It doesn't change the atmosphere because um the the wonderful thing about VOCs is that they're either there or they're not. If they're there, then the reason that they're there is there too.
00:13:02
Speaker
If they're not there, then the reason has gone away. So it's a very real time status report on what's happening in the shed. and And we tend to do it um at five in the morning because, um well, obviously for layers, it's ideal because if you have free range layers, they're in overnight. and So we've got time to get a nice healthy collection of VOCs in the atmosphere and broiler sheds again. um atmosphere tends to close down a bit overnight. everything's There's no human activity to corrupt the VOC signatures or anything. And so everything that's in the shed overnight tends to be coming from the chickens. And and so we can see what it is.
00:13:46
Speaker
And how how clued up does the the the farmer, the producer, the grower need to be in interpreting results? Do they see um you know an outfeed of pure data and that they need to be able to identify where things have changed? Or or is it being presented to them in an already interpreted view that they can see you know something has changed and and this is the consequence of what's changed? This is the interpreted kind of impact of that change.
00:14:14
Speaker
Yeah, we're still developing the app and the dashboard, but the intention is pretty much like the other dashboards and the apps that you see. It's very much geared to giving the farmer the information he needs. So it will, ah for example, it will say if we see um ah bronchitis, then it would say, you know, we believe there's an indication from the VOCs that this is in the shed, you should go look.
00:14:44
Speaker
Or it could be that there's something going on in the shed that the senses identify as being different, but they don't know what it is. So then it will just give a general alert that says, we think you should go and check your birds. Yeah. Ever get a false reading or ah or a false alarm? You know, you set people off with, my goodness, we've got avian influenza here. Oh, no, don't panic.
00:15:07
Speaker
ah We can never say never, can you? But... um Our belief is it's as once we've once we've got sufficient data, then every every packet of data that we collect will add to the information that we're we're delivering. So at the moment, the indications are we're very accurate. I mean, typically, if we get beneath sort of 80%, 80, 85% accuracy, sensitivity and specificity is what we're talking about with sensors, um
00:15:40
Speaker
then we would be concerned. but But typically we would expect to be higher than that. But again, you can never say never because there's always going to be something going on that hasn't occurred before or you haven't come across. So there there is a kind of a health warning with all this stuff. What we're also looking at is in terms of the more complicated data is creating a dashboard that would work for vets so that we can give vets a kind of an advanced training so that they can do deep dives in because the ultimate, we believe, with our system is it will be predictive. So for example, with, um again, if I go to broilers, because that's where we've got the work and I'm im most familiar with it, so forgive me. and
00:16:23
Speaker
With necrotic enteritis, you have coccidiosis, then you get clostridium perfringens, and then you get necrotic enteritis. and So what we can do is we can say, oh yeah, and and we did a um ah study of this actually in America, which proved it. um Yep, Coxie was introduced to the birds on day 17. Yep, we saw it.
00:16:44
Speaker
Then um I think about 13 days later, Clostridium perfringens was introduced. Yep, we saw that. And we were able to say at that point, look out because it's coming.
00:16:55
Speaker
And that how much notice are you giving? What is the the, you know, the whole point of this is early warning, isn't it? Real-time analysis for early warning. So how many days ahead of, of um you know, the clinical signs, the physical signs are you with VOC?
00:17:09
Speaker
Right, well, the first good thing with VOCs is VOCs don't rely on symptoms. So they will see asymptomatic cases as clearly as they'll see symptomatic cases. The other thing is we we typically expect to see um disease outbreaks one to two days
Predictive Power of VOCs in Poultry
00:17:28
Speaker
earlier. We're actually waiting. We're doing some trials with APHA. They've got some issues with their containment facility at the moment. But when it's free, we're going to be looking at Newcastle disease and obviously avian influenza. And the hope is that we should be two to four days ahead of when symptoms make the condition of the birds obvious.
00:17:49
Speaker
So again, um that's something that will be very useful. The most important thing that I haven't mentioned that should be mentioned with VOCs is the fact that all samples have to be collected in an enclosed space. So for example, although we expect to be able to see avian influenza really clearly in an enclosed space,
00:18:08
Speaker
We can't promise to see it when ah wild bird lands on the farm because it would the VOCs would be just too diluted. And also the bird probably wouldn't hang around for the hour that we would need it to, to be able to gather in the VOCs. No, and i and but I think everyone needs, from a biosecurity point of view, needs to assume that avian influenza is there and and keep that mindset really, isn't it? So so in terms of the...
00:18:32
Speaker
the the the timeline of this technology you were saying you've got you've got prototypes yeah out in the field now so how close is this technology um to being something that ah that a farmer could could buy off the shelf and and and use on their own farm right well for broilers it would be um we're hoping by the middle of next year uh with egg laying flocks i'm at the moment i'm talking to some farmers who will be doing some air sampling so that we can get just a basic feel for what VOC signatures of laying um sheds look like, um then we need to look at the diseases and ensure we can identify those. Then we need to trial it on farms. So that will be about 2027 before it's actually working. The really good news that is that the the the box is ready because the box was just a nightmare because obviously you've got all these moving parts and all these mechanical engineering bits and pieces that you have to make working. The sensors actually were a doddle compared to just getting the mechanics right. So, um so yeah, so we reckon early 2027 for air glares. And then obviously we'll be looking at hatcheries um and the rearing farms as well.
00:19:45
Speaker
Can I ask the dreaded cost question, Angela, for a broiler shed? um what what's what's the What's the setup cost? um we We need to confirm how many units we'd need in ah in ah ah a laying flock. In a broiler flock, we think it's one or two, maybe two, it depends how big. um But again, the density of the birds tends to be higher. So that would probably be one box, maybe two. The boxes themselves, the... um the sensing boxes come in at about £7,000 to £8,000 a unit.
00:20:18
Speaker
But of course, once they're in then they just sit there and they do all of the sampling, all of the testing, all of the processing, all of the reporting every day on their own without any additional costs. We are going to have a subscription model, obviously, that will go along with that to ensure that we're managing the data appropriately for people. And we will also, in addition to that, we'll be collecting everybody's data and anonymizing it to ensure that um the whole system just improves its performance um as it as it goes as it develops.
Contact Information and Future of VOC Technology
00:20:54
Speaker
but um If people want to learn more listening to ah to to you on on on the podcast, we've obviously put some details in um in the chat, as it were, underneath this. But yeah what would you encourage people to do? um Well, get in touch with me. I mean, they can email me. They can call me. I'm very happy to talk to people. and Or they can go to our website where we've got some very basic information on the moment. I'm afraid I am our webmaster. And as you've already learned, my technology isn't that great. So it's there. But yes.
00:21:27
Speaker
Fantastic. Yeah. And the the other thing is obviously we'll be um ah we're on LinkedIn. That's the other one. So, yeah. Great. Well, thank you ever so much for talking to us today, Angela. um It sounds really exciting and very interesting to see where this goes. I think it's just amazing how we really seem to be at a point where really new, exciting technologies are coming really coming along at a pace of knots and things that we thought were...
00:21:54
Speaker
were pie in the sky, would have thought a few years ago, now are really ah really starting to bear fruit. So yeah, excited to see how things go once they actually start getting out in chicken sheds and and people start using them.
00:22:05
Speaker
I think the integration of those sensors in the sheds with the support services into farms, you know vets or or processing companies, et cetera, and taking a kind of holistic view of how the environment that the birds are being kept in is changing and then being that much faster on the response times.
00:22:24
Speaker
you know points to a um ah ah yeah serious development in in progress in in the industry as a whole. We'd love to think in only a handful of years down the track, maybe this will become um kind of commonplace in ah in commercial poultry.
Podcast Conclusion
00:22:40
Speaker
Thank you. All right. We'll wrap it up there. Just a quick reminder to anyone um who is listening or watching and hasn't subscribed, please subscribe to the podcast. um That way you won't you won't miss one when it when it comes out every week.
00:22:53
Speaker
um But otherwise, we'll see you next week. Goodbye.