Finally got around to watching this and I agree entirely with what @Avishai said.
I’d maybe try showing both Android and iOS at the same time where possible, especially if you’re trying to show feature parity between both platforms. Separate videos are a bit of a pain to link people to unless it’s highly platform specific and involves a 3rd party brand (I’m thinking Apple Pay and Android Pay in this case). Possibly take some inspiration from how videos such as this one by MKBHD are set up. If you do some of the setup up front with an aim to reuse it across videos, it’s very little extra production work to show both.
I’d definitely cut down some of the script’s specific wording and use more visual elements, tap markers or a very careful overhead camera setup will help guide and show the user what you’re talking about, rather than telling them to “tap the large blue button at the bottom of the screen”.
As for titles, I’d consider dropping the “How to Monzo Ep2:” bit of the title entirely, these aren’t particularly episodic videos that follow an order and by dropping the numbers, it allows you to remove or remake videos in the future without any obvious break in numbering. Dropping the beginning of the title also means that it’s less likely to get badly truncated by space restrictions, I get “How to Monzo Ep2: Send Money, Request Money and…” on my subscriptions page and the TV YouTube clients barely make it to the end of “Send Money”.
A good way to get around this is to use some kind of graphics on the thumbnail to make it clear what series it’s from, then consider using the “Set as official series for this playlist” option of the playlist settings and YouTube will add some extra metadata around the video on almost all platforms to make it clear that this video is part of a series. For bonus points, set a specified order that isn’t manual and use auto add with a tag to automate the task for new videos.
This is such a minor point but when capturing video for the background, try to select a portion that doesn’t have people making large movements around the room or walking too close to the camera, large areas of motion distract from the video’s key content. This is why the screensaver videos from the 4th generation Apple TV are drastically slowed down from their original speed, though that specific method wouldn’t work for capturing video of an office.
In unrelated comments, it would be nice if you could work within the Monzo testing environment to populate the feed with some realistic data and highlight some cool merchants in there. Same with the contacts, there’s some potential to fill those with some community members (with permission), staff, or create a series of clearly fake but fun users such as Monzonaut Red, Monzonaut Green, Monzonaut, Blue, and Monzonaut Yellow (clearly stealing from the colours in the M ), with Hot Chip being the demo user. It’s a bit of work up front but done right, captures from that basic setup is endlessly and safely reusable across all marketing and presentations, including live streams and even in print. This would also help to keep the continuity of data between takes and devices from the same video. The continuity can help show the feature parity between the two apps and means you can pull tricks such as showing the same notification/feed item arrive on both apps (maybe a small time warp in editing to show them arriving at the same time!).