Currently, mid infrared (MIR) spectroscopy of milk samples is used to predicted milk fat, protein, and lactose. While the technology has been utilized for milk recording for numerous years, the application of mid-infrared to measure other parameters in milk is a recent development. More recent studies have shown that MIR can predict different milk quality attributes (e.g., Soyeurt et al., 2011, 2012) as well as other performance traits such as energy balance (McParland et al., 2011, 2014), methane emissions (Dehareng et al., 2012) and health and fertility traits. Here we extend the predictions of novel traits from milk MIR to implementation in the industry as part of routine milk recording. We will discuss the system established to routinely capture and used milk spectra in an operational milk recording business. The system established has allowed us to apply prediction equations for a range of traits to national spectra on almost 117,000 animals (over 1 million test-dates) from 355 farms across the UK (from Feb 2013 to Jan 2015). Systems for routinely predicting novel traits have been established and routine extraction of data for the estimation of genetic parameters and breeding values developed. For example, heritability estimates for cow energy balance, body energy content and energy intake were 0.12, 0.23, and 0.15 respectively. Mid-infrared-based traits predictions from routinely collected data at a national level have potential use in genetic improvement of livestock for previously difficult and expensive to record traits.