Skip to main content

Enterprise Grade Lean AI platform

Periodic Reporting for period 1 - BrainMatter (Enterprise Grade Lean AI platform)

Reporting period: 2019-05-01 to 2019-08-31

AI models are trained via supervised learning using huge amounts of categorized data. Preparing the data is very time-consuming, because currently the data needs to be manually labelled. As a result, productivity is low, and costs are high. Secondly, there is a general lack of large high-quality data for the models to be accurately trained. Thirdly, AI platforms are narrowly focused on just a few data types.
Overall aim of BM is to develop an intelligent and highly customizable platform, that facilitates the full AI lifecycle from the creation of high-quality data sets independent from the data source to the training and deployment of AI models.
The importance for the society of the BM project lies in paving the way to industry-wide AI adoption which is a huge market estimated to create $3.7T value by 2027.
Technical feasibility: we defined a technical roadmap, including work plan for the coming two years with technical specifications, and further a path to reach TRL9. We have analysed the technical risks and defined a contingency plan. Our planning also includes task-by-task planning of budget and time.
Commercial feasibility: we carried out a market analysis of the Big Data and AI for enterprise applications markets (Western EU and global scope): software market to reach $46B value by 2027 with 16% CAGR growth. We narrowed down our target verticals to manufacturing and defined the market drivers, barriers, as well as conducted a SWOT analysis; we also benchmarked against competitors to re-define USPs. We identified stakeholders, standards and regulations. We concluded an FTO and defined an IP protection strategy. We updated the business model canvas with findings of the study, including revenue model and pricing.
Financial viability: according to the findings of the technical and commercial studies we updated our financial projections with the re-defined revenue and pricing model and deemed BM to be financially viable.
We have signed a collaboration project, BrainPower, with IBM that will allow us to test our solution in their machines. The revenue from the collaboration with help us to finance further development activities.
The wider impact of BM is allowing the adoption of AI industry- wide and drastically speeding up automating of error-prone, time-consuming and boring data labelling tasks which will allow data scientists to centre themselves to tasks of higher-value creation and therefore contribute to unleash the potential of $3.7T AI value.