Kelsey Hargreaves, Technical Manager at BICSc, reports.
Before my time in the cleaning industry, I spent a lot of time at university studying the practical use of artificial intelligence and understanding its place (UIP): the positives it can bring to a situation, the negatives, and the ‘embedding actions’ we may need to take to allow peak performance of the AI. What do I mean by this? Let’s think about this in a day-to-day example of the industry:
We have a battery-powered vacuum cleaner:
Positive: It’s easy to move with, and it allows safer practice without the trip hazard of a cable.
Negative: The batteries do not last as long as a direct inlet power connection.
Embedding action: The designated person (preferably the user), will need to charge the batteries routinely and we should invest in multiple batteries per user where required.
In this example, the embedding action we need to take to ensure that the battery-powered vacuum works without issues in relation to the ‘negative’ would be to ensure that we have multiple batteries available and that the batteries are charged.
Now, this seems elementary, however, the same ‘embedding action’ thought is not applied when we think about cobotics in the cleaning industry. New cobotics continue to be introduced to the industry every year, with powerful and insightful updates formed from the response of the industry and the ‘real-time’ experiences from cobotic implementation. Even with these developments, the same 3 challenges to the introduction of autonomous cleaning remain, namely: Cost; Standard of clean; and Usability.
Now, I am not using this article to sell a new robot I have created, although I do like the sound of the iKelsey 3000. I want to establish that outside of initial preconceptions, we need to think of what ‘embedding actions’ we could take to negate problems with autonomous cleaning. Surprise, surprise, I’m going to talk about training. Yes, it’s that key tenet of service delivery. We need to think about training within automated cleaning as essential and not ‘nice to have’.
Let me show you what I mean: let’s take our earlier UIP and think about cobotics.
Introduction of a cobotic vacuum cleaner:
Positive: Allows cleaning operatives to have extra time for detailed work.
Negative: Cost, standard of clean, useability.
Embedding action: Training of staff to the full implementation of the cobotic.
Let’s look at the impact the embedding action has on the negatives, the same as what we did with the implementation of a battery-powered vacuum.
Cost: The highest recurring cost we have within a cleaning contract is the cost manpower. The accumulation of the increasing hours on detailed work gives us an ever-increasing cost. When we embed new machinery that alleviates these staff responsibilities, we can save money on extra hours needed for things like ‘detail work’. With correct training for operatives, we can offset the costs of new cobotics against the money saved on extra time.
Standard of clean: Training an operative and training (mapping) a cobotic on the correct ‘co-working’ they are undertaking is instrumental in providing a good standard of clean. If an operative is not trained in their new responsibilities or time allocation to new tasks, we will not see a change in results.
Usability: The average person knows how to ask Alexa to play Bohemian Rhapsody, but they don’t know how to ask a cleaning robot to focus on the corners and edges, how can we change that? Hint: it starts with a T.
We wouldn’t expect an operative to use a ride-on machine without training, so why do we neglect to think about the impact that training can have on the introduction of autonomous cleaning? We need to realign our thought process; innovation only occurs when we embed understanding alongside it.
http://www.bics.org.uk
About the contributor
Kelsey Hargreaves
Technical Manager
BICSc