FAQ

What’s the core of Transformation, the real deal?

From our experience, it is both control & compliance by design. Control relies on protocols that allow the company perform in its most efficient version (Business-as-a-Factory). Else, inefficiencies and overflow of projects. Compliance is a zero or one problem. It can be a killer. On one hand, if you transgress any obligation you can be fined largely. On the other, if you are late at getting ready for compliance, if you wait and do not include it by design along your transformation, it can become an unbearable cost down the road. Obviously, those that can comply within the most sensitive areas will be able to have the largest impact, protected by large barriers to entry. But beware, if control & compliance are not by design they are prone to high operational risk.

How can I start transforming myself? What do you recommend me?

Nobody knows ex-ante - not you, not us. It is always the solution to an ad-hoc problem. It requires expert analysis to discover your optimal roadmap. At the very least, you should onboard tactically your previous projects on an algorithmic platform and start your new ones natively upon it. Also, resist the hype: focus on the platform, not on the data.

I don’t know where to start with the algorithms...

It would be surprising if you did. First, we would advise you to seek to minimize science to reach a certain impact. Nevertheless, there is model risk. And yet, if that minimum requires a shift of the frontier of knowledge, so be it - we do it often. Second, we always advice to start from your expert heuristics. Upon them start giving room to machine learning wherever you wouldn’t know what actions to take. The projects will evolve naturally thereon weighing human insights vs machine ones very organically. Note at this point that, as a result, algorithms can really be anything - starting from the own governance of the company (on-platform organisations) and ending by very deep analysis of challenging data.

What’s an algorithm, anyhow?

We define it as a sequence of smart actions (across internal or external apps, directly or upon APIs) following a series of smart signals (machine learning and expert heuristics) that lives in a specific architecture (ideally, federated). Every piece can change the performance of the algorithm hence, they are all crucial in its definition. Note that is not a mere model of machine learning as often confused.

I want to protect my intellectual property and my data. How can I still collaborate with 3rd parties?

As far as we know, federated technology is your best option. We have been pioneering it (what we call “the machine in machine learning”) since the mid 2010s - Data MAPs technology. Beware, it is heavily underpinned by applied science which in turn is a very difficult challenge where, we believe, very few players in the world can deliver properly - it mixes domain, tech, machine learning, game theory, synergies exploitation and regulation. This is, your most respected providers may not be suited enough.

I don’t use almost any sensitive data and it is anonymized. Control is fine. I don’t need compliance.

First, you either comply or you don’t. It is a zero or one game. Second, AI regulation is being built globally as we speak. We have privileged insights by being members of ISO’s AI, ISO’s Metaverse and GPAI/OECD - where we are further co-chairs of the AI Regulation working group. We expect it to have a large impact on many algorithms you may need to rely on - note at this point that it means it will go well beyond the models, reaching the robustness of the architecture and its cybersecurity. Also, anonymization is a process that can often be reversed by combining the anonymized data with other sources of information. It is quite a subtle source of risk. A lot of focus has to be put into it. Where possible, federation helps improve things as you avoid bringing data that you don’t necessarily need to have.

What would you do if you were us?

We are probably one of the most ambitious companies in the world when it comes to algorithmics, efficiency and compliance. Our own internal tech is federated by design and it allows for a very flexible approach to algorithmics - from governance protocols to complex strategies and impact metrics. We further account for a synergies recommender system as well as a compliance bot directly pointing at production libraries (i.e. control & compliance by design). Thereon, not only can we best estimate the cost and impact of a new project but we can also take control of those being developed up to combinations of products and services never unlocked before. You can start that ambitiously or less so but, at the very least, you should pay attention to the legacy that you are setting out in your company (both in terms of tech and culture). The right one, flexible, is a booster, a competitive advantage. The wrong one, a burden, it can strangle your evolution with respect to your peers. There is very few people out there who can judge insightfully what you should do optimally.

Yes but we are different. Our culture is slow and very traditional. We are massive/very small…

Echoing that you are slow and traditional or talking about your size or your past or this or that gives us goosebumps. We typically avoid clients without the right attitude and these are very red flags. In these cases we advice you not to do anything. At least you won’t be creating the wrong legacy.

How can we start working together?

Typically, we would advice you to engage via a SciTheWorld’s strategy consultancy (e.g. as-is, to-be). SciTheWorld is our Centre of Excellence and the holder of the intellectual property and its licenses. Then, we would propose you an ad-hoc tech stack to build upon our platform apps you should be buying as well as proposing project managers and engineers from 41OPS to deliver the strategy roadmap - i.e. we cover from the PowerPoint to the production code.

Note at this point that we are used to scale with consultants from other companies around the world, your preferred developers providers, which we supervise through our internal tech of control & compliance (C-Squared). As a result, the way we transform you from centralization or decentralization into federation is gradual and frictionless. You won't notice the difference - just in costs and time-to-value.

· FAQ ·