Four years ago this month we founded Peak with the vision of building a new kind of tech company. A company that truly enabled its customers to do great things with data. That vision has taken us on an amazing journey and, today, we’ve reached the latest, greatest and most exciting point in our short history. We're making Peak’s AI System publicly available to anyone wanting to build and deploy AI solutions in their own enterprise through our beta programme.
Up to this point, Peak has provided our customers with both the technology and the services they need to succeed with AI – all for one fixed subscription. This has led us to build a product which supports enterprise AI deployments at scale. One that enables data scientists to concentrate on their core job of creating machine learning solutions. A product that enables IT teams to rapidly and securely manage the services needed to support enterprise AI.
In creating this product, we’ve realised something; it is the future of enterprise AI and every company needs the capability it offers. We call it AI Platform-as-a-Service, but let's just call it Peak to keep it simple!
Why every enterprise needs AI
Let’s pause for a minute to discuss the impact AI can have on the enterprise. It’s just hype, surely? It’s just another fad – the next “big data” – and won’t really take off, right?
The power of AI, specifically machine learning and deep learning, is profound. What’s more, it is entirely horizontal in its application. It won’t just shape one industry or particular vertical, it has application everywhere.
For example, our research has shown us that AI-powered retailers are growing 30% faster and have 50% higher profit margins than those not adopting the technology.
Moving on to our platform, Peak is a new category of product – one that combines the three critical components necessary for enterprise AI deployments:
1. Infrastructure: Data handling, storage, compute, scaling, extensibility, security and robustness
2. Workflow: The workflow of taking raw data, interpreting it, joining it, transforming it, feature extraction, model training, orchestration and deployment
3. Solutions: Ready-to-use machine learning solutions and business solutions
In short, the entire Peak platform is built from the ground up for AI.
Complex AI problems, solved
Our early experiences as a company, as well as much shared research, show us that data scientists spend more than 80% of their time on tasks other than training machine learning models. These tasks range from cleaning data, collecting data, data munging, feature extraction and building training sets. Proof of concepts often fail to make it to production due to the complexity of deploying within the enterprise and, for those solutions that do get productionised, a data scientist is often supported by two, three or four engineers. This is all caused by complexity in the data science workflow and infrastructure, and these are the complex problems that Peak solves.
Data scientists are now able to spend their time focusing on machine learning. Armies of engineers are no longer needed to support machine learning in production, and time to deployment has been rapidly shortened. All of this means our customers are extracting more value from their data, and more rapidly than they would otherwise be able to.
Our public beta programme gives every data scientist, data science team and engineering team access to the power of Peak. We’ll help support everyone who wants to be part of the programme and are looking forward to the feedback we’ll get to make our platform even better over time. You can register your interest here and we’ll get back to you right away.
Getting to this point has been a huge team effort, so some shout outs are definitely in order. First, the amazing Peak team. Specific call outs to the whole engineering and product team (check out Praneet, our head of product’s upcoming piece on the making of Peak and the AI System). Peak’s data science team (take a look at our head of data science Tom’s blog on how our platform enables his team at scale). And let’s not forget our customers – the trailblazing early adopters! – and our hugely supportive investors and board.