We, in the research industry, have been encouraged to think of automation as something new, something that we can “sell” to research buyers; however, I have concluded that is not the case. And here’s why…
As soon as the research industry started to collect quantitative data as a commercial endeavour, we have looked for ways to make it less labour intensive (and thus more cost effective). We moved from manual data entry to punching cards to CATI to CAWI, making the most of automated (and ultimately computerised) processes.
For now, automation in research relates most strongly to data collection, data processing and visualisation. We should be using automated/ programmed processes to remove manual labour in all three of these facets. We’re already doing much of this through our use of computer programmes, and we should be asking ourselves how we do more – we’re starting to hear more and more about the potential of machine learning and AI in research.
One way in which we can do more, is to standardise (or harmonise) the research inputs. Automating custom research projects still requires a designer and sophisticated tools; however, if our programmes can anticipate the need (in the first instance because we can tell it what to expect) then automation can truly support the research process. This is the principle behind Research Now’s automated research solution where we harmonise a client’s research effort on multiple, unique, custom projects into a single, templated research solution – that still possess the ability to accommodate customization – allowing them the ability to move quickly, build norms, and remove manual tasks.
By using a templated input, data collection can be turned around as fast as we can ask people to participate. The data can be processed seamlessly without interference and the visualisation is pre-set. And this is where I’ve struggled up to now. That description doesn’t sound like “automation” to me, it doesn’t sound like artificial intelligence, it sounds like a logical sequence of pre-set tasks/ events replicable only because they’re always the same. And that’s just it: it’s the very basic level of automation – it’s not sexy or that trendy, it’s mechanical and restrictive and crucially it doesn’t involve any interference from the researcher.
The industry should continue to look to the future to harness automation principles in the effort to provide improved, error-free research services; however, automation has been a key part of our industry for many years and we continue to use it to enhance what we do. We just need to continue to take use it as we evolve our industry.
If we understand it, if we plan for it, if we apply it well, automation…can remove dullness from the work of man and provide him with more than man has ever had before.US President Lyndon B. Johnson
Recognising that there is a level of fear around automation in research regarding the future of research jobs, I will end this post by paraphrasing a 1964 quote from US President Lyndon B. Johnson: “If we understand it, if we plan for it, if we apply it well, automation…can remove dullness from the work of man and provide him with more than man has ever had before.” Let’s embrace the streamlining of our industry processes, redeploy our efforts to be more consultative and drive intelligent business decisions.
