According to you, which factors contributed the most to help create a fertile ground for the early-stage companies and investors in London?

I don’t think it’s a single factor, but the catalyst was the SEIS scheme which offered very generous tax breaks to investors influencing their risk appetite for earlier stage deals. There was then a multitude of other significant factors that played a role – smart professionals branching out from all of the world-class companies based in London, who believe they can make a difference with their ideas, a mature asset management industry with the expertise and capital to allocate and a boom in the number of coworking spaces and accelerator programs. Each of these factors served to reinforce one another and create a fertile startup ecosystem.

What was the most interesting trend that you noticed in 2015?

We saw an explosion in the adoption of microservices in 2015 and specifically containerised application deployment with Docker leading the way here. Think of it as having everything you need to run an application, the source code, system tools etc. all within a single container hosted on your server. This ensures that it will always run consistently irrespective of the environment it is running in.

We also started to see startups being built almost entirely on APIs. At the end of Q4, WordPress released their official API plugin to add a Rest API to a WordPress site within about 5 minutes flat. You can then easily connect other services to your own product through API integrations and we started to see startups being assembled very quickly together using this methodology.

Is there a market crunch recently for Series A funding?

I don’t think there is a series A market crunch….currently; but I think it will be harder for startups to raise Series A funding in 2016 if they don’t have attractive unit economics – or at least a clear path to getting there that is not overly capital intensive. Investors are more hesitant about the valuations and many are expecting the climate to get harder.  I expect later stage valuations to continue falling as they were already falling in Q4’15 in many of the pre-IPO stage companies.

It’s unlikely that this will trickle down to materially impact Series A valuations though as the cost of scaling companies hasn’t really changed. The best companies will continue to be able to successfully raise Series A rounds. If they are the first company in the world to launch a particular product/service, then funding will be easier, but if the idea is not globally unique, the case for the vast majority of London Start-ups where there’s probably 10 or 20 more other startups around the world doing the same thing, it’s going to be a tougher fight.

What is your view of the current AI based start-ups in the industry?

Let’s break that down first in terms of how we define an AI based startup, as AI, machine learning etc. really became buzzwords in 2015. I look at so many decks of companies who feel they have to be doing (or at least have to be seen doing) AI or Machine Learning. To me, that doesn’t necessarily mean they are an AI company. It just probably means that they are using some open source learning machine library and classifying themselves as AI Company. But the true AI cos like Deepmind or Vocal IQ, are actually still rare, they are more akin to R&D projects, focusing on understanding the full environment and making predictions based on pattern recognition and optimal data routing.

If you look at the number of companies with an ‘AI’ related keyword in their description, in the UK we have seen approximately funding into about 50 cos. This is still minuscule compared to the US where about 603 companies have received funding. Playfair was the first seed VC in the UK to really focus on AI and we have since seen a bunch of other investors start to take an interest in the sector –  which is great as these AI companies often usually need to be supported over a long period and require substantial funding to really achieve their objectives. I think we will continue to see an acceleration in the number of AI companies as well their capabilities and commercial applications in 2016.

What were the marquee deals since DeepMind and what impact did they have?

One of the biggest London deals is Jukedeck, who recently raised £2m in seed funding last November. By combining artificial intelligence, music composition and audio production, they let you create your own music at the touch of a button. Try it! (even if just for fun).

The special thing about this project is that a lot of people have woken up to the fact that AI is not just for doing boring mundane tasks such as extracting information from images or natural language processing, but actually, it can be used for more creative tasks as well.

Another company that made news was Vocal IQ, a Cambridge-based company acquired by Apple for an ‘undisclosed’ amount. VocalIQ was a deep learning startup that used voice recognition to make voice conversations with your smartphone more natural and a 2-way dialogue. Apple snapped up this technology to integrate it into Siri and fuel the arms race to develop the world’s best AI-based personal assistant.

What are the challenges that AI startups may face before they taste success?

Long production cycles: As mentioned earlier, most true AI companies are more akin to R&D projects. It takes time for these to come to fruition as well as to generate a sufficient amount of high-quality data on which to train the data models.

Open source competition: A lot of AI technology becomes open source very quickly, so maintaining a competitive advantage over a longer term can be challenging.  

Lack of technical understanding: I’d say there is a FOMO investor sentiment right now with regards to AI. They know it’s going to be huge and they need to get on the bandwagon. But understanding AI software is highly technical with a long steep learning curve. Many investors are therefore hesitant to invest in areas where they don’t fully understand the inner workings under the hood and the potential wider commercial applications of the product.

If 2015 was a year of …….. 2016 will be the year of a ………………

If 2015 was a year of Slack, 2016 will be the year of a whole suite of products built around it. In the new year, I expect to see increased automation of tasks within businesses and an acceleration in the adoption and availability of tools and bots built around task management and communication platforms such as Jira, Asana, Slack or FB messenger. The cost of building startups will continue to fall, competition will become more intense and users will start to feel overwhelmed by the extensive choice of products to use.