A machine chooses the best startup at PAPIs Connect


Artificial Intelligence is a fact, a machine is able to predict the success of a startup through analyzing data and asking questions. This is what happened in Valencia, where PreSeries chose Novelti as the startup most likely to succeed. This is possible due to the alliance between Telefónica Open Future_ and BigML.

The event coincided with the world-famous Valencian Fallas. The noise, gunpowder and fireworks were around the city, but the sparks were also present in the Paraninfo conference room at Universitat Politècnica de València (UPV) during the first Artificial Intelligence startup battle, celebrated at PAPIs Connect. The startup battle was the perfect context to present PreSeries, a predictive platform able to predict the success of a startup in its very early stages, based on data analysis.

The two-day conference was filled with interesting talks and panel discussions presented by Artificial Intelligence experts such as Nuria Oliver, Telefónica’s Scientific Director, and Ramón López de Mántaras, Director of the Artificial Intelligence Research Institute of the CSIC, among others. The most anticipated moment of the PAPIs Connect conference arrived with the startup battle, featuring five contestants: Datatrics, Emotion Research Lab, Intranetum, Novelti and Restb. The originality of this event was that, for the first time in history, the winner was chosen by an algorithm, without human intervention.

The run-up to the battle was presented by MªJesús Alonso and Poul Petersen, from PreSeries. They introduced the joint venture between BigML and Telefónica Open Future_, represented by Francisco J. Martín and Ana Segurado respectively. “In the last years, we have seen that around 80% of the startups we meet fail in the first five years of development. With the implementation of this innovative and must-have technology we can mitigate the risk involved in investing in new businesses and help the startups implement a better business model to secure their future” explained Ana Segurado.

20160316_144256 (1)

The audience was expectant and the contestants were nervous as the moment to present their projects neared. Each had to present their company in three and a half minutes.

Five startups with five ambitious projects:

  • Datatrics. Provides a predictive marketing platform very affordable and easy to use for marketing teams.
  • Emotion Research Lab. Analyzes human behaviour through facial recognition.
  • Intranetum. Uses Artificial Intelligence to classify the knowledge and know-how from a given organization.
  • Novelti. Offers analysis in real time to detect and prevent outliers from the cloud in all devices connected to IoT.
  • Restb. Monitors, with Deep Learning technology, the content generated by users that upload their photos to online spaces.

After the presentation of each project, PreSeries platform assessed them by asking questions such as the number of funding rounds, whether the founder has lead other projects previously, or how many team members the startup has. In this context, and together with other parameters the platform already knew, PreSeries gave a score between 0 and 100 to crown the winner, which was Novelti with a clear difference compared with the second place finisher, Intranetum.

It was a great result for Novelti, who won with a score of 86/100. Marco Laucelli, its CEO and Co-founder, together with Álvaro Agea, Technology Director, received the prize from Ana Segurado: the entrance in one of the Wayra academies, Telefónica’s startup accelerator. This means to be invested by Telefónica Open Future_ and accelerated at a Wayra academy.

“Telefónica inviting us to be part of their Wayra initiative and its ecosystem is very important. On top of that, having the opportunity to present Novelti in Boston is also a great prize for us”, highlighted Marco Laucelli, CEO and Co-founder of Novelti.

Finally, as closing fireworks for this celebration where Fallas and Artificial Intelligence got together, Novelti will participate in the AI startup battle world finals in Boston, which will take place at the end of 2016.

20160315_174958 (1)

An important milestone for this project that thanks to Machine Learning has been chosen without any kind of human manipulation.

This post has been translated from the original post published in blogthinkbig.com in Spanish by Pablo Requejo Rodriguez. The author of the video is Melissa Hortet.

Re-thinking Early Stage Technology Venture Investing

Venture Lifecycle

Warren Buffet once famously said “Only when the tide goes out do you discover who’s been swimming naked.”, which is distinctly relevant as of early 2016 thanks to a very volatile start to the year in public stocks — including even the eternal favorite tech blue chips. Unicorns, those privately financed tech companies with valuations well over $1 billion, have not been able to weather the market sentiment storm either, thus being forced into the unsavory choice of down round financings to be able to keep executing on their world domination plans.

With all this activity, 2016 makes for a challenging investment climate to say the least. As pundits have heated discussions on financial news networks, investors are getting ready for another potential investment winter against the backdrop of a 7 year bull run losing steam by the day. This points out to a climate, where the easy money of years past funding startups with the right buzzwords in their pitch decks may not be a guaranteed thing in the near future. In other words, we are likely facing a conjecture requiring an elevated level of discipline in assessing a tech venture’s fitness as early in the cycle as possible.

Chances are you have read about the controversial topic of Series A crunch that have been experienced in tech investment strongholds like Silicon Valley. This phenomenon has been a direct result of the supply side expansion of early stage investment sources. As a result, the average startup has found it much easier to secure enough funding to get started. At the same time, costs associated with building a minimum viable product and testing for product-market fit have come down thanks to more flexible product development methodologies (e.g., Lean Startup), scalable and affordable cloud infrastructure services, new open source software frameworks, as well as social media engagement and content marketing powered word of mouth that took the place of paid marketing campaigns (at least in early phases of evolution). To boot, crowdfunding services have sprung up and rapidly scaled such that they have become the go to initial funding platforms for many first-time entrepreneurs in highly fragmented market segments such as e-commerce, consumer products and even consumer hardware e.g. Pebble. They also offer the added benefit of helping recruit paying beta users without having to shell out any equity. They are no panacea though, and many startups find it difficult to cross from early adopters to the mainstream side of the famous “chasm”.

You may be thinking more early stage startups are not necessarily a bad thing, albeit the truth remains that (with intense competition for later stage and growth funding a reality) building a notable company of regional or global scale is still as tough as ever. Every year thousands of metaphorical eggs are spawned only for a few to make it to adulthood after natural selection has its way with the upstream swimming bunch that have to clear one hurdle after another. This proliferation of incomers have also added to the challenge of identifying winner-take-all type entrants early enough for serious investors to be able to recoup the costs of their entire venture portfolios instead of catching those at the tail end, when the valuations are already driven sky high by less sophisticated angels and/or private equity. Without a doubt, Tier 1 VCs still have the inner track given their strong networks and the loyalties of serial entrepreneurs from their past investments. Nevertheless, the task of spotting the next blockbuster has not necessarily gotten any easier for them either.

All of this makes inquiring minds ponder “There has to be a better way!”. At Preseries we are of the opinion that it is high time for early stage investing to catch up with the times and evolve what remains a very manual, clique-driven and subjective practice. At the end of the day, it is proof enough that the current approach leaves a lot to be desired since Uber had to receive seven rejections in one week while trying to raise $150k in exchange for 10% of the company in its early days. We live in an increasingly data driven society awash with information. It is unreasonable to expect any one expert to be able to cope with this incredible amount of intelligence arriving by the minute. We need machines to step up and help us pick up the relevant signals 24/7 in order to be able to stay up to date and people, projects, companies, competition, macro-forces and the consumers at large. As hard as it maybe to imagine for some, we believe venture investing is not necessarily immune to omission from the list of skills that were formerly deemed to be only performable by human experts.

So who needs services like Preseries today? The answer is pretty straightforward: Anyone with a vested interested in picking the wheat from the chaff when it comes to early venture investing. This means private investment offices and angels, who are looking to write a check to the right team of founders. Traditional VCs that aim to zero in on worthy companies earlier than they otherwise would with their current processes. Incubators and accelerators who are looking to systematically accept applicants with the most upside to their programs. Pension funds, endowments and strategic investment fund managers who are looking to evaluate the investing chops of those private investors that they are about the trust with their contributions. Even founding teams that would like to compare and contrast how they fit in within their ecosystem on multiple dimensions.

At Preseries we don’t yet claim to have the magic formula for early stage investing 100% nailed down. Anyone claiming to be in possession of the perpetual energy machine equivalent of the technology investment management will rightfully meet rolling eyes and scrutiny. As such, that claim in itself may be crazy. But what’s event crazier to us is denouncing the value of methodically piecing together relevant traces of information hidden out there in the vast and growing knowledge graphs the web has to offer so as to better assess the outlook for any given technology startup. First augmenting, then perfecting and ultimately replacing the semi-informed yet fallible human decision making is the difficult yet very worthy journey that Preseries has officially set on today.


And not without a bang either. PreSeries is launching at PAPIs Connect 2016, Europe’s 1st Artificial Intelligence conference for decision makers, managers, and app developers. Today’s program includes a first in history startup battle segment, where the winner will be chosen solely by the PreSeries algorithm without any human intervention. So wish the participants luck and stay tuned for more!