First of all, who are you and what do you do?
EnergyWay is a data science and artificial intelligence company that offers innovative mathematical models to improve the efficiency and the sustainability of companies. We like to say that our models transform numbers into knowledge and with that, we help our clients to optimize available resources for the benefit of the people, the organization and the planet.
I’m Andrea, responsible for the business development, strategy and marketing activity of Energy Way.
How long has EnergyWay been operating? How big are you?
Exactly today we are turning 6! It is really exciting to see how much we evolved in this period! Today we are 47 scientists: mathematicians, physicists, engineers, neuroscientists, economists, designers and philosophers. And we expect to hire other 20-30 talents in 2020. The annual turnover around 5€ million and this is also growing very quickly.
Today we have offices in Italy and we are opening a subsidiary in Israel, in Haifa.
We have a customer portfolio of over 150 customers, from large corporates to SMEs in very heterogeneous sectors, from power and gas utilities, telco, mass market retailers, food, banks and manufactures.
To them we deliver custom artificial intelligence deployments, such as predictive controllers for water treatment plants, demand forecasting and churn prediction for telco and aerodynamic defect prediction for Formula1 racing team.
What are your plans for growth?
For the near future we have really ambitious targets. In 2020 we will kick off a new industrial plan that we are completing these days with the collaboration of our outstanding advisory board. We are really happy to have the opportunity to work with Mr. David Bevilacqua, former managing director of Cisco in Europe, Francesco Caio, which is the president of SAIPEM and former CEO of giant corporates and advisor of UK government for digital agenda. And all the other members of the advisory board. They are providing us the guidelines and the best challenge we could desire before going live with the plan.
The plan has 3 pillars:
The first one is about the business model.
From one side, we will keep on doing what we successfully do today, that is super customized Artificial Intelligence projects with a sort of advisory model, but in addition to that we are setting up new business units to replicate successful projects with a completely different business model in order to create recurring revenues.
The second Pillar is about the recruitment
We will double the number of headcounts shortly. This is the reason why last week we launched our Academy in partnership with several Italian universities. This program consists on 1,5 months of advanced training (replicated every 3 months) with 20-30 post-grad and post- PhD students for a deep dive on Artificial Intelligence. Our aim is to keep in Energy Way the best performers.
The third pillar is about our Education Projects
Since 2016 we are fostering new educational approaches to secondary schools and high schools. At the beginning many didn’t understand why a startup like us was investing so much energy in this activity rather than focusing exclusively on business growth. But we really believe we must play a role to contribute to the evolution of the community we are working in and the way we can do that is by promoting in schools how mathematics and data can be fundamental to understand reality.
For instance, we created a 6 week program for secondary schools which is about coding, energy management and CO2 management. It’s name is “allora Spengo” literally “then I turn it off”. We teach students how to create programming open source hardware tools to monitor the energy consumption of the school and how to manage the building. It is a nice and fun way to bring coding and data science basics together with environmental, energetic and civic concepts.
We are also receiving great feedbacks and media coverage every time a public administration give us the opportunity with new schools so it’s really a win-win situation
Can you comment on the difficulty that companies have in accessing talent in artificial intelligence, data science, etc?
I understand the difficulty and can guess the reason why. One reason is the heritage of the policy of the past years: we have very few students that are getting degrees in STEM subjects. It’s something we have to face – and it’s the same for startups and large companies.
We try to stimulate in the young generations the beauty and the creativity that these subjects can stimulate, trying to lower the distances between everyday experiences and mathematics or informatics.
I don’t really know what the recipe is – but when you have the opportunity to get in contact with students that completed a university program in mathematics or informatics, you need to give them the environment that allows them to express the creativity of the subject.
I remember the opening day speech of the Dean of informatics engineering at my university in Lugano: he said the biggest mistakes was to combine the word “engineer” with the word “informatics” as at the end of the day there is nothing more creative than informatics or mathematics. But teenagers usually misread the etymology of the word “engineer”.
In September 2019 we launched the manifesto for “sensible rationality” because we want to highlight how all this world of AI / machine learning / deep learning (use all the buzzwords that you want) is nothing more than mathematics – something human beings have used since ages.
Today we have the opportunity to use mathematics with cutting edge informatic tools – we understand we have a complex environment to interpret – and without these tools, we’re not able to see the correlations of elements that influence the processes we want to understand..
Young professionals want the opportunity to express themselves through the tools they learned. In a flat organisation this is for sure much easier than in a vertical organisation such as a large corporate.
One mistake we see in corporates is that they often tend to acquire startups/scaleups that can quickly improve their business but then when they try to integrate these group of professionals in their rigid/vertical organization, they experience a “leakage” of talents.
Would you say this is because of culture, hierarchical needs, or?
A combination of these factors. There is probably a common cultural element that comes from the hierarchies and compartmentation of corporates. Hierarchies can create a slow environment that does not fit with the ambition and potential of these experts
You don’t keep these people with money and benefits – you keep them with corporate culture and with an environment that allows them to do what they are able to do.
And most of all, they need to be contaminated and to contaminate others in the organisation they are working on
This is something we really feel – there’s a sort of misunderstanding on the figure of the data scientist. A data scientist doesn’t mean anything. If you want to complete an AI project, you need multiple expertise that rarely the same person has:
You need someone to translate business domain into a first mathematical framework (this is a real rarity);
Then you need a pure mathematician to optimize this mathematical framework;
Then you need an informatic/numeric expert to translate this into code;
Then you need other informatics competencies to translate this code into something that can be used by the business.
And all these people need to contaminate each other and get contamination from the business. If this team gets placed in an environment which does not allow them to share between each other or don’t allow them to communicate with other business units – then it doesn’t get results, people get depressed and then leave.
Is it true that data scientists are stuck in the dichotomy between very flat startups and very highly paid positions at larger corporates?
I must admit my conflict of interest in answering this question! I’m promoting the idea that a digital transformation project works better if data science units are outside!
My point of view is that usually one of the important ingredients is that data scientists need to work on different topics. Both for their attitude but also for the seek of the project.
Usually, the insight that you need to complete a project can come from very distant use cases.
For example, once we were working in a project for a supplier of Lamborghini: it was a computer vision project for detection and classification of defects on car-painting… that time we created super performing algorithms.
But when we were asked by another automotive company to help with a similar use-case, as this was not in the luxury segment and had a completely different processes, we understood that what we experienced before was almost useless for the second one: in the first case we had a corporate painting 3 cars/day and we could use 3d images. When you want to classify the defects you need the altimetry of the defects to distinguish between scratches, holes, dust, etc..
In the mass market car plant, we were facing a process where they were doing 36 cars/day – so not enough time to encode the 3d scanning images from the cars. So we needed to build a completely different tool for a case that was similar in theory.
When we faced another project in the medical sector for the detection of tumors – thyroid tumors out of echography – a good portion of know how from the Lamborghini case was perfectly fitting. When we were asked for a corporate in agro to detect insects in vineyard plantations, what we developed in the mass market car manufacturer was a good beginning for that project.
So you can see how distant are these use cases and how they empower the team to find the right solution.
What about an organisation working on a very vertical topic with a team of data scientists to work efficiently on the same topic – we find this very unlikely and typically demotivating for the people.
Even if salaries and conditions are good.. Sometimes they find a scaleup in the province of Italy can be much more motivating to work.
Let’s take a step forward and talk about matchmaking: what’s your story with us? How did you get involved, what activities have you been involved and what’s been your experience?
We got in touch late 2018 for the Budapest event, very few days before the event but we gave it a try. We came back with a very hot prospect and several leads. Some weeks later we signed a deal with GEICO Taikisha – a super challenging computer vision project that we fulfilled and now we’re doing already a 3rd project with them.
Then, we participated in almost all the other events, Graz, Brussels, London and Milan keeping a relations with almost every attending corporates.
We’re super happy with the results of these Scaleup Summits – as we got the opportunity to get involved in discussions with companies such as ENI, ENEL, ACEA, Acciona, SAIPEM and many others – last week we managed to sign the first contract with ENI, and we are honestly very close to sign deals with some other corporates met at the summits!
One of the reasons we invite you to multiple events with us is that we always get positive feedback from the corporates and that continuing a conversation over a longer time period helps to continue the process. Do you have a strategy for your success?
Ahah, I don’t know if I can call it a strategy – we have always been very honest from the beginning. One of our assets that we discovered to be rather unique is that we don’t approach the corporates with a ready product / solution. At the scale-up summits like in many other contexts this is not what is expected by corporates- many startups have an off the shelf approach.
That’s fine, but what I think is interesting when we have talks with heads of innovation in big corporates and midcaps – is that we tell them about an approach. A successful approach in several cases and industries. And we understand how this approach can fit with their main issues.
We mastered a format for explaining what is AI in our perspective – using concrete examples – we always show the correlation map that we patented to explain our findings in terms of correlations of variables affecting a process. In this map we show linear and non-linear correlations (weak correlations). And with this map we demonstrate how we can make explicable the mathematics behind a complex project.
When we explain this approach to the interlocutors, then it’s easier for them to tell us the “stomachache” that their company is facing.
Then our goal is to approach this “stomach ache” with the experiences gathered in totally different use-cases.
I think you’re describing very well how to make your counterpart see the value proposition. What we find the best fit for our events is when companies are coming to co-develop value with the corporates. I assume that typically you’re not really an investment or acquisition target?
Actually, we have had several acquisition or share offers that we’ve consistently refused.
Let me also add that this picture of co-development does not always work like that. There are many innovation departments that are specifically looking for ready-solutions. One question they often ask is: what’s your track record for that specific deployment. We’re not able to answer those kinds of questions.
If the interlocutor is on this fixed pattern, it’s really difficult to change the value proposition that they might see in our work.
Let’s see if I understand – if the innovation departments are looking at very specific evaluation metrics, because of your nature, you might not fit this?
Yes, exactly, for example including reports for track records and so on, excludes us immediately.
Our mantra is that when you build a data driven project, it’s really rare that you can replicate a project from one corporate to the other even though they are working in the same area.
This is true in industrial deployments, probably less true in digital marketing or fintech.
Luckily when we have the opportunity to talk with the C-level of these organisations and this is understood, then we can make progress.
When we have to face a super structured department of innovation, this can sometimes be not the best way to present yourself, as they need to fit your value proposition in their formats.
If their format is not elastic enough, we waste a lot of time. Time is money for both. But for a scaleup like us, it has a much bigger impact. If I need to spend 10 meetings to make my value proposition fit to the pattern of my interlocutor, to be sincere, today, we close the discussion early and we move to other playground and another lead.
In our experience, this speaks to the importance of innovation at the C-level. We typically say that if you work in innovation, you have to work differently. If it was just another project with the same metrics as before, then it’s not the level of innovation that we’re aiming for. It’s something that we find very stressful between startups and corporates, these different languages.
So on a different topic, is there a difference in working between larger companies and smaller companies?
The answer is quite obvious and I’ll try to be original in my answer..the answer is definitely yes.
With midcaps and small corporates the main issue that is the time to get on board is much much shorter. On the other side, it’s not often possible to tackle a really ambitious project from moment one, as the investment capabilities are lower. So there is a trade-off to be considered.
The main reason why we put so much effort to work with larger corporates in our early stage was for the referrals. We are working with many large corporates: Telecom, Poste, Hera, Barilla, British Telecom, A2A, etc- not that it’s such a challenge to work with them once they are on board – and we work with many SMEs too, we come from Modena, which is a land of SMEs. And we have positive feedback from both segments. But getting on board with large corporates it’s really painful yes. Usually we take 3-5 months to sign a contract from the moment when we agree on the topic, methodology and budget. While for a midcap, the deal is done when you agree on these 3 things!
But with large corporations then you have to work with legal and procurement departments, and as we are a “strange beast”, it’s not easy for them to classify us in terms of supplier – are we a software vendor? Are we an advisor? What about intellectual property? When you engage in these types of discussion, time tends to get longer and longer.
After that, working with large corporates you get to work on tasks that are stimulating, it’s easy to scale up but you need to be ready to fight a lot at the beginning and keep your mindset steady along the collaboration – there is always the temptation to acquire you – and you need to be sure about your targets and your objectives for the future.
One last request: for our corporate listeners, large and small, what would be your advice?
To give us a try! Not joking, give us a try quickly – should be the mantra of any innovation department – it’s better to try quick and fail quick, rather than take ages on onboarding and then discover something is not working.
Don’t be too obsessed by including a scaleup or startup in a conventional process if the value proposition is good sounding to you then why not finding the right way to let the startup you selected to work on the topics you want to ameliorate.
I think that’s the crucial point.
For entrepreneurs, what would be the advice?
Be super professional, show a lot of enthusiasm and find a way to impress the interlocutor.
Forget the classic recipe. We don’t have a “product” and somehow we’re making it!
Don’t be too blocked by the format the corporate would like to receive. Be consistent with what is the value you can bring to them and find your recipe.
Manuals on innovation are not really effective. Even corporates understand that is an experience on fields and also there are no fixed rules on how to establish relationships with external bodies that are contaminating their organizations.
Thank you Andrea!
If you’d like to know more about Mind the Bridge’s matchmaking activities, check out our Scaleup Summits or our reports on Open Innovation.
By Ricardo Silva
Head of EU Projects – Mind the Bridge