Counterpoint to the above: No is not a bad word.
Depending on the circumstances some things just won't be possible or will simply annoy the other players at the table. If a player starts in on something that has no business in the setting quickly apply the word "no" and move on with things.
I'll build on that a bit. When you say 'no', give a short explanation why. Then, follow that up with 'we can talk about it after the game if you disagree, and if I'm wrong, we'll do it differently next time'. That goes over a lot better than a flat 'no' and avoids arguments happening then and there.
--
As far as the discussion afterwards, IMO stick to two things: First, you're trying to be fair and consistent. Secondly, you're always solving for making the game more fun.
For example, one player thought he could make his verbal component for the Suggestion spell just the specific words of the suggestion, which he'd indicate to the DM with 'finger air quotes', so if the target made their save they wouldn't even know what happened. I said no, and explained the problem with that. "Casting with undetectable verbal components is a) a specific Sorcerer-only ability, b) costs spell points so it's an expensive, limited resource, and c) bypasses the ability to be Counterspelled so it's insanely powerful. And the PHB specifically says Verbal components 'require the chanting of mystic words.' So it's too much to be able to pull off right in front of someone's face."
The player saw the point we made, but still disagreed somewhat, so we talked later.
We discussed it from several perspectives - game balance, class abilities, enabling player cleverness, etcetera - and I made the case that the problem is with it being a 'win' button for casters that trumps anything any other class can do, with no risk. Only Sorcerers should be able to pull that off, with a valuable expense of resources. He saw that, and agreed, but he also made the case that since Suggestion is innately a social spell, being always automatically detected or understood nerfs the spell to the point of uselessness. So, he wanted the potential to get away with it in some situations, because that would be a lot more fun too.
The whole group talked through it over Slack for a few days. We came to an understanding and an interpretation that totally works within RAW. Basically, since verbal components are chanted clearly at a low speaking voice, depending on the range, ambient noise, and other distractions, people other than the target might not hear them. So, if it's a bustling city square, it's low risk. In a quiet chamber, with other potential witnesses, that's a high potential to be found out, and that's where the fun comes in.
Basically, we figured out it's a Help action to influence the DC of the other observers to not notice the spell. It involves the whole party in on-the-spot improv or planned shenanigans to generate more background noise, create distractions, shield the spellcaster, or something else. Those actions are executed using one skill or another against the observer's Passive Perceptions. Depending on the circumstances - i.e. trying to pull it off against a city official when guards are waiting right outside - it requires more creativity and becomes more exciting. So, instead of the spell being an automatic 'win' for the caster alone, it now incentivizes teamwork and everyone gets involved in making sure it goes off.
Judging from how the group has taken to it - the amount of planning, and the creativity of their schemes - we made the right call. We all came out of it with a better understanding of the game too.