Artificial intelligence offers smart opportunities for PE
An increasing number of larger-ticket investors are being attracted to European startups in the artificial intelligence space, with the UK in particular likely to be home to a boom in the years ahead. Christopher Papadopoullos reports
The terms artificial intelligence (AI) and machine learning often conjure up images of driverless cars, grey jumpsuits and mannequin-like robots performing household chores. But beyond the current hype a quiet revolution is underway, with more businesses making use of the technology in innovative ways, and more opportunities for investors beyond the venture stage as the sector matures.
"AI has the power to transform industries, but it won't be as soon as everyone thinks," says Simon King, a principal at Octopus Ventures. "AI is now in the peak of its hype cycle, but there's a disconnect between what it can deliver and what people are expecting it to deliver. Autonomous vehicles are a classic example in my mind – it's going to be decades before we have self-driving cars that require zero intervention from a human driver."
On the so-called Gartner Hype Cycle, AI is approaching the trough of disillusionment, King says, where technologies cannot match initial expectations, and only after this realisation do they start coming back out and delivering. King helped build the investment case for Octopus's investment in video compression startup Magic Pony Technologies, which was later sold to Twitter for $150m. Octopus has also sold language processing company Evi to Amazon and predictive keyboard Swiftkey to Microsoft, which both made heavy use of AI.
Many of the companies successfully funded by VCs have been mopped up by tech giants keen to acquire rare talent. "The strategic acquirers recognise there's an opportunity here," says King. "They want to get in early, and it's really core to their proposition – increasingly so for Google, Amazon and Facebook. They recognise they need talent, and that talent is scarce."
AI is now in the peak of its hype cycle, but there's a disconnect between what it can deliver and what people are expecting it to deliver" – Simon King, Octopus Ventures
David Kelnar, investment director and head of research at MMC Ventures, says this was part of the first wave of AI activity – firms developing and acquiring foundational capabilities. Now the second wave is underway, and this is attracting the attention of private equity. "The second wave involves firms applying AI to specific business functions, such as marketing, advertising or security, to make them more effective or efficient. Or, AI companies are sector-specific, focusing, for example, on finance, health or manufacturing," Kelnar says.
The second wave involves carrying investment opportunities from the venture stages to the mid-market. One GP active in this second wave is Kennet Partners, which invests £8-25m in minority growth equity and growth buyout transactions. The GP has made two big investments in financial automation: Rimilia, which targets larger firms, uses machine learning to automate accounts receivable, cash allocation and cash forecasting processes, significantly reducing the amount of manual work required in finance departments; and Receipt Bank provides an automation platform for bookkeeping firms and small businesses. It is not just potential growth that makes these types of firms attractive to investors.
"The software-as-a-service model provides a highly visible revenue stream, as opposed to the traditional licence model," says Kennet managing director Hillel Zidel. "These companies are also dealing with huge amounts of data, which has a lot of value." Zidel is wary of buzzwords such as AI, big data and internet of things. "AI now is what big data was 18 months ago," he says.
Instead, Kennet is more careful, focusing on businesses that are more established and past the early stages. Kennet takes the technology very seriously, so much so that it even uses machine learning technology to trawl business databases and news sources to seek out investment opportunities. Zidel says financial automation is "massively underexploited", with plenty of roles in finance departments ripe for automation, which provides disruptive growth businesses with opportunities to scale-up.
Many of these developments in the B2B sector are going unnoticed. MMC's Kelnar says: "There's a quiet revolution happening whereby AI is dramatically improving firms' abilities to gain insight from data and their ability to automate previously manual processes." MMC has made five investments in the past year in companies using AI.
The opportunities for private equity investors are likely to remain in the B2B sector. AI is being used by a range of B2B businesses, whereas only a select number of consumer businesses are using AI. The reason for this is data. AI-based businesses need a lot of data from their clients, and it is easier to get this data on businesses, but more difficult when it is consumers.
"You usually don't start a consumer business with a lot of data about your users," says Kelnar. "New B2B suppliers can plug into businesses' extensive corporate data sets to train their algorithms. But it's difficult for B2C companies to overcome a 'cold start' challenge regarding data. They usually have to develop their AI capabilities over time, as they gain data about their customers."
"AI has profound potential in sectors that are considered less exciting by some – such as manufacturing, professional services and retail. In these sectors, between 30-50% of peoples' time can be spent collecting and processing data," Kelnar says.
Kelnar expects significant developments in predictive maintenance in the manufacturing sector, where production downtime can be costly, and in healthcare, where AI can increase efficiency and reduce costs. In healthcare, applications of AI include medical imaging (in which AI is used to identify tumours), improved drug discovery through the processing of vast amounts of published research, and automated diagnosis. He is also optimistic about the future of autonomous vehicles and believes the transformational impact they could have on other areas – insurance, for example – is being underappreciated.
In Europe, almost all of the activity in AI has been in the UK. This may be for several reasons, of which the number of top universities is certainly one. "The field is old and the developments that once again kick-started it as a field happened in 2006, so there are few people outside of academia with expertise," says Octopus's King. Many of these, he adds, are based in UK universities, notably in London, Edinburgh, Oxford and Cambridge.
Entrepreneur First, a startup incubator based in London, has a network of academic computer scientists on hand to help aspiring entrepreneurs. It then helps them to grow, pitch to investors, and takes a stake in them. Its most successful startup is Magic Pony Technologies.
But it is not just in London where such businesses are springing up. Kennet portfolio company Rimilia is based in Birmingham, while MMC's latest investment, a data analytics company called Peak, is headquartered in Manchester.
The UK also has a traditionally strong environment for startups, with venture capital trusts and enterprise investment schemes helping to direct funding to early-stage companies. This blend of research, early-stage investment and momentum in the proliferation of AI-focused entrepreneurs, means the country is well positioned to provide a wave of investment opportunities for private equity backers in the years ahead.