PetRescue launches dog obsessed AI via DDB Sydney to help pets get adopted across Australia
Online, pets get a lot of love. Especially on Instagram – some dogs have more than five million followers and get a million likes every day, while rescue dogs struggle to get attention and, most importantly, find a new owner.
Knowing that people who follow these dogs online are most likely to adopt if they find a pet of their favourite breed available for adoption, DDB Sydney together with PetRescue have combined data and machine learning to create a solution. We’re calling it PetMe.
PetMe is an AI capable of identifying dog breeds, temperament, size and coat, and matching those qualities to over 9000 available rescue dogs across Australia.
This is how it works – post any dog image using #petme. The AI will recognize the breed and auto-comment back to everyone offering a dog of exact or similar qualities up for adoption.
Many influencers have already joined the cause, and we’re inviting everyone to help out by posting as well.
Agency: DDB Sydney
MD: Priya Patel
CCO: Ben Welsh
ECD: Tara Ford
Creative Art Director: Alex Newman
Creative Copywriter: Guilherme Machado
Creative Developer/Technologist Patrick Schroen
Creative Strategist: Augusto Correia
Design: Dandreti Design
44 Comments
This is clever. Well done everyone
I don’t get it.
How is it ML?
And #petme is an existing hashtag with over 145k posts.
Simple, clever and for a great cause. Love it. Not bad, DDB.
Why hasn’t this existed yet? What an awesome idea.
Really great!!!
A really positive step towards responsible pet ownership 🙂 Awesome!
Bloody awesome!
Build the cage. Lions are coming
Best thing I’ve seen here in a while. Can see this being a really effective way to get rescues dog a new home. Deserves some metals.
Allow me to break it down for you, nice and slow, so you get it.
The AI responds to every NEW use of #PetMe.
Cheers.
Tried it, came back with a small dog that wasn’t even similar to what I posted. But the main problem is that it is finding me an animal to adopt in another state 100 of 1000’s KM away.
So unfortunately the AI smarts aren’t very good and it would be better to develop in a geo location tool that can match you to an animal that you can easily adopt in your current city.
Probably pick up a couple of bronzes but can’t see this solving the adoption of animals in shelters, which is whats its all about after all.
Gui and Alex are two legends. Well done guys!
http://www.campaignbrief.com/2011/04/colenso-bbdo-creates-facial-re.html
This is awesome guys. Much better than Westpac 😉
Troll 1) aka ‘Done so much better’
The old doppelganger dog-finder idea that finds a dog that looks like you, is not the same as AI that finds a dog that closely resembles the one you want, in looks and a whole range of other factors. The only thing done here is your cerebrum (Google it).
Troll 2) aka ‘Close but so far away’
Just using geo will not get you a dog you’re likely to want – which is the insight. Your thought lacks any insight, and is not even an idea: it’s just a technology use on its own.
So you’re not even close, you’re on the other side of the intellectual universe.
Now go iron our your finalist certificates and leave DDB to polish their, ahem, statues.
100 of 1000s of km away?? Where do you live, the moon? https://spaceplace.nasa.gov/moon-distance/en/
I don’t understand who this is for…
The poster already has a dog so has no use for this tool
The post’s audience is not involved and therefore unaware unless they read through the comments so will likely miss it
… is someone supposed to post a picture on instagram of a dog they’d like to use this tool?
If you’re going to build a complex piece of tech like this- surely you’re going to get more mileage by connecting it more explicitly to its target audience…
You obviously work at DDB. If you live in NSW your never going to adopt a pet from Perth, Victoria, Queensland etc etc etc
I get it you’re trying to win awards but this could be so much better, even slightly useful for the general public. That’s why it’s not very good and will be hard to convince a jury that it is.
Have you even visited the site?
It says the AI “Finds a match up for adoption in a local shelter.”
LOCAL.
– The End –
This is awesome. Definitely going to win some metal. Haven’t see anything like this
Great job guys.
Hey there
Good idea.
But use a spell check on the site.
I presume “characterisc’s” is supposed to be “characteristics”?
I think the point is that – the creative doesn’t live up too / or doesn’t answer the insight (and that Colenso’s version did)
So in summary, nice insight poor execution = no awards
This is soooooooo badly put together. It could have been OK.
The problems the industry is facing are highlighted in this award hunting work.
If you can forget the terrible UX, this is pabulum. A cross between a low rent creative’s version of ‘melanoma likes me’ and ‘dog-a-like’.
Am sure the case study video will be unbelievable and it will pick up awards because of it but very few people will actually understand or use this so it will not help dog adoption thus being a giant waste of effort and resources.
Such a shame to see DDB outsourcing their design work (and clearly not even to a particularly talented agency). I remember a few years back when the internal design department was rather formidable…
Here’s the launch video guys https://www.youtube.com/watch?v=37tlF-qb2fI
Has anyone actually successfully used it?
Will reserve judgement for when I have experienced it but hours have passed so far and no reply.
Such a mess of a campaign from idea to execution.
The great thing about campaigns like ‘melanoma likes me’ and ‘dog-a-like’ is the fact that they helped the cause then they won awards because they are insightful, simple ideas which here beautifully executed.
Shame on you DDB.
Are you jealous? Yes, you are lol
your obviously on the credit list.
But the last laugh is at the bottom of the website you created.
The ‘watch it in action section’ is pairing a man in Myrtle Beach, South Carolina, USA with a Small Female Shih Tzu Mix Dog in North Arm Cove, NSW, Australia.
Ahahahahaha. Another dog saved from the pound.
This is rubbish.
Which is ‘Not available for interstate adoption’
https://www.petrescue.com.au/listings/644637
Terrible. Hahahahahaha
@ “re: Does It Work”
Not jealous. I think it’s a cute idea. But surely it is not too much to expect that it actually does do what it says in the press release and website. The entire idea is dependent on it actually functioning.
Still no reply btw. Better get the AI working harder.
I’ve tested it and it works. Got a Husky pretty similar to the one I posted… Some bugs need fixing, sure! But it’s a machine, it will get there eventually.
If the bugs need fixing, why not wait until they are to start talking about it. worse still, launching it?
“It will get there eventually” is not a great endorsement, machine or not.
No interest in actually helping the cause. And probably the worst scam campaign (both idea and execution) that has ever been posted on this site.
I hope this gets thrown out of any award show by the jury. The peeps on the credit list should be made to go out and actually adopt a shelter dog or at least visit a shelter and understand what the problem looks like. About 250,000 animals are euthanized in Australia each year (both cats and dogs) and to promote this half arsed, poorly executed idea with the only purpose to try and win awards is absolutely disgusting.
I’d also be interested to know if any media has been put behind this (7 views of the video last I looked) this idea is way to clunky and poor for this to take off on its own. Campaigns like ‘Melanoma likes me’ and ‘Dog a Like’ are clever and fresh enough to gain their own momentum so deserve the metal that they received.
All involved should be ashamed.
Another blatant and poor attempt to bolt on a piece of tech to a creative idea, if any in this case. Learn from the masters at Colenso for their work on Pedigree before launching this half-baked idea.
Shame on you DDB! This is evident it’s for your own award winning cause!!!
99% sure this isn’t “AI/Machine Learning” or automated and is manually performed by a person working at the ad company.
Instagram’s API as of July 2016 doesn’t allow you to automate comments. Google “instagram comment endpoint” and you will see no such endpoint exists as they removed it in 2016.
It appears to work “too” well. Machine learning is based on making constant mistakes and making adjustments. In all of the tags I’ve seen, it correctly identifies the dog and breed with too much accuracy.
If this was legitimate, the tool could be used to help people who’s pets have run away and been handed over to the shelter however as this isn’t legitimate, I suspect this will never be the push.
I’d love to hear these points discredited and some evidence to prove me wrong, I want to believe, but pretty sure they know it’s not possible.
‘Bile Release Duct’ totally engaged. I hope everyone got that out of their system
Yes please back your claim that this is indeed ML and AI at work.
If it uses a set of data to simply identify a dog breed, it is neither AI or ML.
Does it become more accurate over time without human involvement?
The mic is yours.
It won’t work because no one will know about it beyond trade media. No editorial media pick-up because a news journo will smell the BS on this a mile off.
i was once an AI.
Good try but it honestly looks like digital by people that don’t understand it. A lack of understanding of digital behaviour, not tech. Our industry needs less gimmicks and more utility/value.