Intelligent by design: the business case for AI
Why are venture capitalists so taken by AI?
New technology drives all industrial revolutions. Steam engines powered the first in the 1700s and electricity the second in 1870. In the 1960s, electronic transistors gave us the computer age and the third industrial revolution. Now, the fourth industrial revolution is being fuelled by artificial intelligence (AI).
As with previous revolutions, one of the outcomes will be a step-change in productivity due to increased levels of automation. The socio-economic implications of this have quite rightly been given a lot of air-time already, so there is no need to repeat them here. Instead, I will focus on why venture capital is focusing on AI and what types of business this technology will change.
The rise of intelligent automation
In previous industrial revolutions, automation was driven by the defining technology of the age. While that’s easy to visualise with steam power, electricity and electronics, people have been less aware of software’s power to automate. Yet, even before AI, existing software has already created lasting automated change:
- Automated writing in all sectors with word processors
- The replacement of physically delivered postal communications with email
- Automated and personalised marketing through email marketing tools
- Automated manufacturing process management with SCADA (Supervisory Control and Data Acquisition) software
AI is taking software automation to a whole new level
Historically, automation has been ‘dumb’. The mechanical replication of step-by-step recipes produced by humans. However, some processes were too complex to write a recipe for, either because the number of steps was too large or the ingredients too intangible. Within the last ten years, AI advances have allowed programs to create their own recipes.
For example, while it’s always been easy to explain all the rules and strategies for winning a game of noughts and crosses, it was impossible to write a recipe capable of guaranteeing success at more complex and strategic games. The ancient Chinese board game ‘Go’ has been the exemplar of this.
However, towards the end of last year, Google announced a significant AI breakthrough, when its ‘AlphaGo Zero’ program taught itself to master the game, without human influence. AI’s victory over humans at Go marks a turning point, where an increasing number of processes can now be automated. These advances are the basis of venture capitalists’ feeding frenzy around AI. It is a technology with the potential to touch all markets and industries.
“It’s more powerful than previous approaches because by not using human data, or human expertise in any fashion, we’ve removed the constraints of human knowledge and it is able to create knowledge itself.”
David Silver, AlphaGo lead researcher, Google Deep Mind
The investment case for automation and AI start-ups
As investors, one of the things Octopus Ventures looks for in a business is its ability to scale rapidly and disrupt a sector. Start-ups that automate tasks fall into this category, because they use machines and software to augment or replace human tasks.
Speed of implementation and higher-margin business mean that companies selling technology based on automation can grow quickly and cost effectively. Additionally, the speed, safety, reliability, repeatability and accuracy of automation – combined with reduced costs – creates a powerful economic case for disruption. Taken together, these factors make AI-based start-up businesses extremely attractive to venture capitalists.
Most of the start-ups coming to us are using intelligent automation in some form. But, because it is still relatively early days, few of them have core products, services or business models that rely entirely on AI. Since AI technology can be applied to many problems, we haven’t seen AI start-ups cluster around a particular sector. However, AI-based businesses require a couple of features that other start-ups don’t. For example:
- Access to data: the more data you have, the better your algorithms learn
- Access to AI talent: it’s currently in relatively short supply
Octopus and AI business
Our part in the story so far has been as the backer of three of the most influential Artificial Intelligence companies that Europe has ever produced:
- EV is a natural language search engine developed in university labs. It uses voice recognition software to answer any questions. If this sounds familiar, it’s because we sold Evi to Amazon in 2012 and it now forms the foundation of Amazon Echo. Evi became the basis of Amazon’s first R&D centre in the UK, and now has more than 400 people building on the technology we first backed ten years ago.
- The founders of SwiftKey created an app for faster, easier typing on mobile phones and tablets. It’s now being used on more than 300 million devices worldwide and was sold to Microsoft in 2016.
- Magic Pony Technology uses machine learning technology based on algorithms. These algorithms can understand the features of imagery, allowing it to look at a source image or video and then improve it. With our support, Rob Bishop (who is still only 28) sold Magic Pony to Twitter in 2016 – just 15 months of developing the technology. Rob now runs Twitter’s video team and reports directly to Jack Dorsey (CEO and co-founder of Twitter).
What’s next for AI?
By its very nature, it is hard to predict where disruption will come from. That said, it seems clear that AI-powered businesses and products will push through a lot of disruption in the near future. That’s why we believe that now is a fantastic time to be an entrepreneur or early stage investor.
One of the most exciting things about the application of AI isn’t that it provides the answers we want, but that it can ask the questions we didn’t even think about. It seems likely that the most disruptive AI start-ups and companies will be the ones that create the products and services we didn’t even know we needed.