May 2021 Summaries

Speech recognition technology has improved significantly, with accurate output for many spoken languages, but still faces challenges such as background noise, punctuation placement, capitalization, correct formatting, timing of words, and domain-specific terminology, which can affect accuracy on a case-by-case basis. Additionally, data security and privacy concerns have increased, with 42% of respondents citing mistrust following media portrayals of tech giants, and language coverage remains a barrier to adoption for global businesses. To overcome these challenges, providers must focus on deploying and integrating voice technology in a simple and secure manner, addressing language coverage gaps, and prioritizing data security measures such as on-premises deployment, dark site environments, and private cloud deployments.

This month I spoke to the Daily Telegraph’s James Cook and Morgan Meaker for an article they were writing about Cambridge – ‘Why Cambridge’s Silicon Fen is fast becoming Europe’s answer to California’. Whilst we are home to fewer celebrities and considerably less sunshine, it is true that Cambridge, England, home to the Speechmatics HQ, is a hotbed of the world’s best technology talent. The Fens provide the perfect environment for agriculture and my interview with The Telegraph made me reflect on what it is about Cambridge - an ingrained culture here creates a fertile environment for big thinkers leading on world-changing innovations. Anyone who lives or visits here marvels at the surroundings, with natural beauty and nods to great city greats like Stephen Hawking and Charles Darwin dotted across its footprint. Getting out of the same room sparks creativity, particularly after stillness from the last year. Big thinkers think big, promoting a culture that accepts failure as a side-effect and setting 'BHAG' goals drives excitement and energy into businesses. People come to Cambridge to study and often end up staying, taking years to research and perfect their ideas, and it's essential to structure days to allow time for thinking and reflection, discussion and collaboration. At Speechmatics, we encourage teams to take time to think in a way that best suits them, with initiatives like 'Hackamatics' - a 'sprint-like' event where teams collaborate over unique projects with no distractions.

The adoption of voice technology is increasing as companies recognize its value in improving customer experiences, streamlining workflows, and increasing revenues. The region expected to see the largest growth in speech recognition technology post-COVID-19 is Asia Pacific, driven by the growing economy and population. North, Central, and South America, as well as Europe, are also predicted to adopt voice technology rapidly due to economic, social, and technological factors. Voice technology will cover additional languages such as Spanish, with a focus on any-accent language packs, in the next three years. The perception of accuracy for voice technology is specific to use cases and businesses, with improvements expected in word error rate, noise handling, and accent-independent language packs.