Amid continuing global excitement about AI breakthroughs and governments talking big on AI futures, at Intellegens we’ve been getting enthused about a slightly different question. What can we do to make the AI technology we already have as useful as possible? That’s led us to some deep thinking about the machine learning tools people actually need for R&D – and to a big new product announcement.
It started, as it usually does for us, with conversations with our customers about exactly what they are doing with machine learning. We’d had much positive feedback on how our Alchemite™ user interface enables more scientists to apply machine learning. But, each time we added new functionality, the user experience got slightly more complex, particularly for those with a focused task, like a creating Design of Experiments (DOE) projects. Alchemite™ was easier-to-use than alternatives, but could we optimize it further by recognising ‘one size does not fit all’? And could we do this while improving the ability of different user types to collaborate?
1. The right app for the right task…
Our Product Manager, Rachael Clarke, led an in-depth study that identified a number of key recurring tasks among our users. Fast DOE was one. Another was quickly generating an ML model to fill data gaps and test out ideas – for example, when designing a new formulation. Then, there were users whose focus was delving deeper into the full suite of analytics that an ML model opens up. Finally, ML is also applied to automate and inform other lab and production systems. From this analysis sprang Alchemite™ Suite – a set of apps, each focused on a different one of these tasks. Alchemite™ Designer, for example, provides only the features needed to quickly set up DOE runs in a few button clicks. This approach also enabled us to add a host of new features and analytics for users of Alchemite™ Innovator, the app that offers the complete ML toolkit, without complicating the experience for other users.
2. …but, collaboration matters too…
R&D organisations can now match the right app to the right team member. But our customers also want to retain and improve another key benefit of a shared approach to ML across their teams – collaboration. A strong collaborative focus sometimes pushes companies towards monolithic, integrated software systems that involve too many usability compromises. We rejected that approach with our targeted apps. But these apps are based on a common architecture and user experience that makes them interoperable, with project management tools and the ability to share, re-use, and collaborate around ML models and results. We also added Alchemite™ Viewer, a lightweight app for managers and other team members who simply want to explore project results and apply them in decision making.
3. …and so does integration
Finally, there is still a bigger picture in which ML needs to integrate with other systems and workflows to, for example, automate decision-making or feed-in insights from data – we discussed a few examples in a recent blog. Alchemite™ Architect provides full API access to our ML methodology, enabling Alchemite™ to play its part in enterprise digital transformation strategies.
It ain’t (just) what you do…
The excitement around AI and ML is all about what you might be able to do with it. But, in making the hype a reality, it’s critical to recognise the wisdom of the old song – ‘It ain’t (just) what you do, it’s the way that you do it’. That’s what we’ve been concentrating on at Intellegens in recent months, and we’re delighted to be able to share the product of all of this work with Alchemite™ Suite. We’re confident our customers will appreciate this new way to get things done. Because that’s what gets results.
