Thomson won’t give financials for the privately-held firm, but its progress has been sufficient to support that half-billion dollar valuation and trade investment rather than more speculative parties.
The co-founder says there’s no immediate need for another raise.
“We’ve got good capital and we’re actually progressing nicely towards profitability over the next 12 months as well, which gives us options.”
Over the past decade, Auror’s meat and drink has been its crime-reporting templates, which make it quick and easy for a retailer to report shoplifting or violence to police, and its automated number plate recognition (ANPR), which helps service stations prevent drive-aways and police locate suspects.
Late last year, Auror, with the path cleared by the Privacy Commissioner laying down new rules following trials by various retailers, it also added facial recognition – or Subject Recognition as the company calls it.
It works like this: as a person walks into a store, their image and a “biometric template” is captured by a camera.
Each store in the “Auror community” keeps a list of “POIs” (persons of interest) – “high harm” individuals who have carried a weapon into a shop or acted violently or are a prolific shoplifter.
If the Subject Recognition system says the shopper who has just walked in is not a POI, their image and biometric template are immediately discarded.
If you have a POI, the store is alerted. It’s then up to a store’s staff how they act. Thomson says the key element is that Auror has given them a real-time warning.
Safeguards include that a store can share information about a POI among its own group – say, a Briscoes store with another Briscoes outlet, but not another retail chain.
Nor can police access a retailer’s POI watchlist.
Signs outside stores in the Briscoes group – one of the first adopters of Auror’s Subject Recognition system – warn shoppers that the real-time system is in operation. (Auror has no brag-list of early Subject Recognition adopters, but Briscoes and Woolworths have self-identified.)
The Herald snapped the notice below at a Rebel Sport outlet. It included: “If you engage in harmful behaviour in our stores, we save your image for a period of up to two years for future comparison.”
Thomson adds that there is a “human in the loop” at every phase of the Subject Recognition system.
All the safeguards notwithstanding, does Thomson sometimes lie awake at 3am, worrying that his company is part of an accumulated system of surveillance and relentless tracking that has the cumulative effect of destroying our privacy? Is there a danger of things getting too Big Brother?
“When I wake up at three in the morning, my concern is more around the people who are being threatened in those stores, day in, day out,” Thomson says.
“And I’m really proud of what we’ve done here, starting from New Zealand, taking to the world, around bringing technology that can actually help stop that problem.”
He adds: “We’ve seen a big shift in the past 10 years. When we started, it was more covert offending. People would try and hide things in a backpack or within their jacket.
“Whereas now people are blatantly just walking out and not caring if they’re being recorded or watched, or even stopped by someone.
“Post-Covid, we’re seeing this rise in violence being associated with that as well.”
Thomson’s go-to stat is that 10% of offenders cause 60% of retail crime – and that the serial offenders are four times as likely to be violent or aggressive.
“We’re keeping retail workers safer with real-time alerts.”
In early 2024, Māori woman Te Ani Solomon entered a New World in Rotorua where she had shopped without incident for 12 years, where a facial recognition system was being trialled.
She said two male staff approached her in the meat section and one got “literally in [her] face” and loudly told her: “You have been trespassed and you need to go.”
Solomon said she felt “racially discriminated” against. The supermarket’s owner, Foodstuffs North Island, said it was a “genuine case of human error” that saw Solomon mistaken for a previous offender and that it would apologise.
The Privacy Commissioner said the Foodstuffs trial complied with the Privacy Act. Lessons from it helped shape the Biometrics Code finalised late last year.
Various studies have found racial bias in facial recognition technology, including a 2019 US National Institute of Standards and Technology study that found “Many of these algorithms were found to be between 10 and 100 times more likely to misidentify a black or East Asian face than a white face”.
“Mismatching in the past has been a big issue. It’s one of the reasons why we’ve been following the technology but not actually engaging with it up until this point,” Thomson says.
“But it’s got to a point where it is very accurate across a wide range of demographics.”
He adds that there are additional safeguards. No one is automatically confirmed as a POI. There’s always a “human in the loop” at every stage of the decision-making process.
All retailer-entered information in Auror is restricted to prohibit the collection of sensitive information such as ethnicity, race, religion, political affiliation or sexual orientation.
Retailers can’t arbitrarily or manually add a person to a POI watchlist.
The POI list cannot be used for profiling, categorisation, prediction, tracking, monitoring behaviour or targeted marketing purposes.
And law enforcement and other retailers can’t access the list.
Auror says “built-in transparency and auditability tools are available,” but there is not 100% transparency. The firm has not named the third-party facial recognition technologies it integrates for Subject Recognition. (A spokesman said all had passed “rigorous” testing. He added the key element was that Auror’s Subject Recognition applied a “governance layer”.)
90% of sales offshore
Auror has more or less a lock on Kiwi retailers, but the small size of our market and rapid growth offshore mean New Zealand only accounts for around 10% of its sales.
As well as being its home turf, New Zealand has been a more straightforward market because we have a single police force.
“Australia’s got eight. The UK has 43 and then the US is 18,000 – every country, state and city, then the federal overlay as well,” Thomson says.
Interconnection between the various local and ultra-local law enforcement authorities has been historically poor.
“The complexity means it’s actually been helpful for anyone that’s been offending because they know they can just jump across that imaginary border and offend again. If they get caught, they can pretend it’s their first time.”
In the US, Thomson says new anchor investor Axon has been key opening doors with big retailers. The taser maker also knows how to navigate the patchwork of police departments.
In the UK, Auror got a big boost last year when then Home Secretary Yvette Cooper urged more retailers to adopt the Kiwi company’s software.
Earlier this month, London Mayor Sadiq Khan jumped the Auror bandwagon as well.
“We’ve brought on a number of new large retailers like Tesco’s and Morrisons [in the UK], but we’re also getting a lot of support from the government, the Met Police and others because the retail crime problem that they’re facing there,” Thomson says.
Where next?
Auror’s largest shareholders are Australian venture capital (VC) funds Reinventure (22%) and Folklore Ventures (16%), Wellington-based VC firm Movac (14%), Axon (10%) and Thomson (also 10%).
Could its next move be a public listing?
“Our friends, family and members of the public often ask, ‘Can we invest in your company?’. That [listing on the stock exchange] is an option for us. But right now, the focus is purely on expansion.”
Will that include a push beyond retail?
“We’re doing more retail adjacent spaces, like shopping centres and business associations,” Thomson says.
“We’re also doing a fair bit in the US around supply chain crime. Offenders and groups are moving earlier in the cycle so rather than targeting goods sitting, it’s the trucks on the way to the stores. So what you see in Fast and Furious movies is now happening in real life.”
Thomson is also eyeing online fraud where small-scale rip-offs are rife but people often feel they have no mechanism to report them.
“It sort of goes nowhere. No one reports it.”
Chris Keall is an Auckland-based member of the Herald’s business team. He joined the Herald in 2018 and is the technology editor and a senior business writer.