Episode 013 | April 18, 2022 | 34:27
Guest: Malthe (Inniti, co-founder)
Published: April 18, 2022
Duration: 34:27
Description
This week we are again guested by Malthe Muff, co-founder of lab-automation software company Inniti. We get deeper into the sustainability potential of automating labs, why Malthe believes the biggest innovation potential lies in software, the biggest blocker for entering the lab 4.0 era, why autonomous labs are not about replacing workers but empowering them, and where digitalization is heading.
Full transcript
About this transcript
This transcript was automatically generated and may contain inaccuracies, typos, or mistranslations. Episodes recorded before 2024 were transcribed by an on-site model and may have a higher error rate. The content reflects the original conversation to the best of our ability. For the authoritative version, please listen to the audio episode.
Everybody talks about how AI and machine learning will revolutionize the space. I completely agree. To really utilize those technologies we need to focus on the quality of our data. The main challenge holding labs back from Industry 4.0 is the lack of interconnectivity. A lot of lab equipment is functionally exactly as it should be — the only limitation is that it is hard to connect to the internet, cloud, or any software platform.
Hi, this is Joachim from Green Innovation Group. Today we are back with Malthe from Inniti to talk about digitalizing labs — practical barriers, how to start, and what labs will look like in ten years.
What does Inniti do? We connect the whole laboratory to our software platform. There is a lot of laboratory equipment out there that functions exactly as it should — the only limitation is connectivity. Data is scattered across many different places. Our little box is an IoT enabler we attach to lab equipment. Bidirectional communication: we gather data AND control equipment. End user builds experiments with equipment from many different vendors on one platform.
During COVID a lot of labs shut down. Our customers could keep working — sometimes on-site to fill tanks, but otherwise control experiments from home and monitor them.
The state of the lab industry is dominated by hardware manufacturers: Thermo Fisher, Mettler Toledo. They have built high-quality hardware. Future innovation will be more incremental on hardware; the new game-changer will be software. The main challenge holding labs back from Industry 4.0 is the lack of interconnectivity — each manufacturer has its own way of communicating, and even within manufacturers it varies. Some speak Chinese, some French, some Spanish. We write drivers for lab equipment so they can communicate with our platform.
Without offending anyone — equipment manufacturers may not have an interest in standardizing. If you spend time integrating a scale of a specific brand into a process, you need to buy the same one next time — lock-in effect. Some have launched closed-ecosystem platforms — only their own equipment. That just creates yet another generation of data silos.
Most labs today have equipment from many manufacturers. Some are connected to a LIMS (Laboratory Information Management System — like SAP but for labs). Some have hard-coded LIMS connectivity, but it does not allow researchers flexibility. From day one we wanted to put researchers in control — drag-and-drop programming to build complex experiments. Digitalization needs to be easy to do. It should not turn researchers into slaves of the system — it should free up time so they can focus on the fun stuff: analyzing data, coming up with new ideas (and yes, emails).
Sustainability benefits:
Data integrity — take data directly from the equipment instead of writing it on paper then typing it into Excel. Big quality gain.
Empowering people to do more interesting work. Better data over time becomes a database that powers AI and machine learning.
Monitoring equipment usage — many organizations discover utilization is low and can centralize equipment, reducing space and machines. When you automate experiments, quality goes up — fewer failed experiments, fewer consumables, less raw material, less rework, less energy.
Partnership with a design-of-experiments software company using statistical methods — instead of 100 experiments to find correlations between factors, just 10. Same insight, less waste.
Inniti enables existing lab equipment to become smart — expand the lifetime of equipment without replacement. Equipment manufacturers may push out new equipment that has Wi-Fi or Bluetooth as a USP — but is that enough value to replace a whole HPLC?
Analogy with ovens — connecting our fridge to Wi-Fi just gives us convenience, but doing the same in the laboratory has much greater impact.
Barriers to adoption:
Increased awareness — people often have digital transformation strategy slides. Need to educate what digitalization means in a lab context.
Fear of “laboratory automation” — sounds like huge pipetting robots. We can also automate simple setups.
Cloud anxiety in large organizations. We can deploy on their cloud too. Communicate with the right stakeholders.
Compliance (GLP — Good Laboratory Practice) makes things take more time in pharma/healthcare contexts. That’s there for a reason.
Saying: there is no impact in slides. Actual changes drive the CO2 down.
Mission is autonomous labs — not replacing workers, but enabling them. Western world has a STEM talent shortage. To stay competitive we need to get more out of each engineer or scientist.
Possibility: brilliant scientist in India operating equipment in Denmark. Connecting global R&D and QC labs — “follow what my colleague in Singapore did while I was sleeping.” Easier knowledge sharing. When people change jobs, complete documentation of their work survives. People sometimes redo experiments because they cannot find what was done before.
10 years from now: a lot will be the same — same tasks — but more efficient, interconnected, way more tools to analyze data, much higher data quality. Everything will be connected. Two competing standards (Spectaris and CELA2) based on OPC UA are being developed — risk of two standards instead of a golden one. Big organizations have the buying power to demand standard interfaces. Don’t accept 40 different ones.
Interplay with virtual labs / in-silico experiments — startups using AI to design molecules — saves tremendous time. Especially in pharma you eventually need to test in real life, but virtual experimentation pushes the real-world test later. Helps upscaling: predict what 2.5 litres looks like at 500 litres.
Thank you Malthe. Digitalization is your friend. Thanks for listening — please check greeninnovationgroup.com.