Although the focus is on the task, it means sense to devise tasks so that significant language is recycled as much as possible, since we know repetition helps build memory.
Here are three examples of easy to implement tasks which produce a great deal of meaningful recycling of language.
Opinion poll surveys are an obvious way of bringing in an element of task-based learning. One of my favourites with near-beginners was to do a classroom survey on the popularity of various school subjects. Once you've taught the subjects and practised pronouncing them accurately you can get your class to go around asking as many of their classmates as possible within a time limit what their favourite subjects are. You could manage this in various ways, but my own was for each student to respond with three subjects (not just one, since this might lead to very skewed results, e.g. in favour of PE). Students make a note of the answers on a grid (or simply on their own piece of paper) then you bring the class together to analyse the results. Students could enter their results in Excel if you think it's useful for them to practise their ICT skills.
In addition, rather than giving them a ready-made grid of subject names you could dictate them to offer some focused listening (transcription) practice.
Make sure you insist that the class keeps working in the target language and that questioners and respondents give full sentence answers, e.g. Quelles matières préfères-tu? Je préfère histoire, sciences et éducation physique (EPS).
Just think how many repetitions of the target vocabulary you will get in a lesson all of which will build long term memory.
Apart from school subjects you could work on food and drink or, say, means of transport.
This example is a reading and doing homework task which students enjoy. Provide a simple TL recipe (or maybe a choice of two or three) which students must translate then make at home. Collect the translations and have a recipe tasting in class during a subsequent lesson. Biscuits, cup cakes or a basic tart would work well. If you can choose a culturally significant recipe even better. Make sure the recipe does not require unusual ingredients which parents would have to go out and buy specially. Make sure also that all students are able to take part, of course. Perhaps the school's food technology department could play a role in that regard.
If you are worried about the Google Translate issue, then provide a gap-fill version of the recipe with options provided. Students must fill gaps before doing the translation.
Although this task does not naturally produce lots of repetition of target vocabulary, it is a good source of imperative verbs and is both motivational and memorable for students. This in itself makes it worth doing, in my view.
A favourite task of mine and one which was enjoyed by students at advanced level was to produce a TV news bulletin. To get the activity going explain that they will be in small groups as an editorial team producing a news show. Give them a list of invented news items and tell them to work out a running order based on the importance of each story. This should generate a good deal of discussion, which you can encourage by joining in as and when it seems useful.
Once the running order is established the group writes a script for the news broadcast, produces a teleprompter (this can simply be lots of pieces of paper, or a teleprompter app for use with a tablet), then records the broadcast. The class could use on the sport reporters and interviews to make the task more fun, perhaps recording outside or around the school. Maybe they could even find a TL-speaking teacher colleague to do a spontaneous interview.
For the initial discussion part of the task and the broadcast itself you could provide lists of useful TL phrases such as "We could..." "We should..." "This would be best first..." "Here are today's top stories..." "Over to our reporter at the scene..."
At the end of the sequence of lessons (about three hours' work?) show the video broadcasts.
For another view of this topic try:
Note that that article mentions information-gap activities as tasks. Bill VanPatten, in his book While We're On the Topic (2017) makes the point that information gap activities may or may not be effective tasks. Many are far from "real life" activities, but this does not necessarily invalidate them in my opinion. Just "practising" language does have a place, I would suggest!