Collaboration and our customers
‍

We value the deep, long-term partnerships we’ve built with customers – founded on a shared desire to solve impactful engineering problems. Our customers choose PhysicsX because they want a software platform they can use to power the whole engineering lifecycle from initial simulation or design phase through to operational workflows, where models aren’t stuck in a lab but are used.

Our most successful engagements are the most collaborative, with direct involvement from all three of business leadership, engineering leadership, and our engineering users. Working alongside them, PhysicsX deploys multi-disciplinary project teams with skills across data science, simulation, machine learning engineering and software development. They embed with customer engineering teams to deliver tools and capabilities, rather than insights or trained models alone.

Our goal is to empower our customers. For this reason our teams focus on enabling engineers to use PhysicsX tools as an integrated part of their existing toolset, and to expand the product to meet people where they are. At the same time, during our collaborations our engineers embed deeply and many of our customers consider them part of the team for those phases – and value that!

What is a pilot?
‍

PhysicsX partnerships with customers begin with a pilot, jointly scoped. The objective of a pilot is to use the PhysicsX platform and tools to solve a core engineering problem – albeit one discrete enough for some solution to exist and be in active use in a short time frame.

Prior to starting a pilot, PhysicsX engineers work with customers to identify a real engineering problem and identify pre-requisites. Most pilots will last for 1-3 months depending on problem complexity, during which time the PhysicsX team and the customer engineering team work side-by-side.

During the pilot, PhysicsX engineers work closely alongside our customers’ engineers – one team tackling the pilot use case together. This collaboration can be in person, remote or hybrid (though at PhysicsX, we’re big believers in the value of being in the same room at least some of the time), and takes many forms depending on what’s most effective.

What makes a good pilot?
‍

• A good pilot is hard
  ‍A worthwhile pilot must move the needle on some critical outcome. It has to be hard enough for you to care about it. It’s not a proof of concept, or a proof of technology, it’s something that matters.
• It’s immediately valuable and not a means to an end
  ‍We don’t work on toy problems, we work on something that’s useful to you from (maybe even before!) the day it works. It might be a subset of a bigger problem, or it might be the first in a series of things that become increasingly useful the more elements exist but it is, standalone, useful.

• Relatedly, the time scale of a good pilot is short
  ‍Just because the problem is hard, doesn’t mean that it should take a long time to tackle. We want you to be able to see impact quickly, in a matter of weeks, not several months. A good pilot typically lasts between 1 and 3 months.
• The success of the pilot is not a factor of chance
  ‍“Win Le Mans” is not an example of a good pilot. We might help you optimise that car, or an element of that car, but “winning Le Mans” is a function of too many other variables. It’s also not falsifiable. The pilot outcome will always depend on some factors outside the project team’s control, but it must be possible to verify or falsify the claim that impact came from using what you built with PhysicsX.
• And you have some clear concept of the “objective function”
  ‍A clear target to achieve is necessary for any good pilot, but that target has to be measured in terms of both impact and engineering outcomes. Whose life is going to be different after this? Why? What will they be able to do that they couldn’t do before? And why does that matter? “Better” isn’t good enough.
• You have users and data (or data can be generated)
  ‍Pre-existing simulation or experimental data is not a pre-requisite – as we can generally work with customers to model the physics and generate this data as a first task – but it is an advantage. Engineering users who will work with the model or its outputs are essential, as our team and product work as a force multiplier for our customers.
• Does that mean I can’t pick a Moonshot as a pilot?
  ‍No! Typically, the critical business outcome is something that can be improved upon with a fast and accurate model. PhysicsX models run in near-real time, but provide the same or similar accuracy as numerical physics models that could take hours or days to run. If the problem fits into that paradigm, and has the above features – great! – but rules are made to be broken, and if it’s ambitious and valuable it might still make a great pilot.

‍

Timeline
‍

Choosing an engineering challenge that fulfils these criteria is important. This is what will give you conviction in the PhysicsX platform and in the collaboration between our teams. We believe it’s worth spending energy in advance of each pilot to identify the pilot scope, and to identify key customer counterparts.

Our ultimate goal is to enable you, our customers, to generate value from the PhysicsX platform independently and at scale. Along this path, you can rely on us to be your partners in overcoming your toughest engineering challenges, wherever it's needed in your organisation.

At the outset, PhysicsX engineers help our partners get started with PhysicsX by configuring the platform, performing initial data exploration and sometimes generation, training models and configuring model optimisation. We also build product. The effort required is the effort to make the outcome happen, which isn’t constant, and isn’t a linear decrease but always corresponds with increased empowerment and involvement of our customers.

As the scale of data and model networks grow and customer engineers become more comfortable with the platform, they own more of the administration of platform and leverage their domain expertise to develop entirely new applications and use cases, often independently of PhysicsX teams. While PhysicsX engineers will continue to build and develop high-priority use cases, many of our engineering customers like to drive the design, testing, and implementation.

Whichever way works best for your organisation, we will support you all the way to tackle your hardest and most impactful engineering challenges. If you have ideas you would like to explore –whether they are fully formed concepts or just initial thoughts – please do not hesitate to reach out.