Indeed, I've done a Samaritans course and actually suicide can be seen as the tip of an iceberg (or similar analogies); it can just be one particular event (some bad news or even something more trivial) can tip someone over to actually act. Even small barriers to their progression such as platform gates, restricting access to known hotspots or Samaritans signs or a short conversation can deescalate people and increases thinking time which is vital.
For a subject which receives an enormous amount of coverage the public understanding of suicide is extremely poor. Many people fear the topic of suicide because it is not rational, and prefer to rationalise the event of suicide as being like the end of a long downward journey, especially with regard to mental health (another topic which has mushroomed in remit, making it unmanageable to discuss). The idea of "mental health > downward spiral > warning signs > suicide attempt" is attractive because it helps people imagine that it is easy to intervene or even recognise the symptoms themselves if they ever became susceptible.
In reality, a lot of people who commit suicide may not have formed a logical, firm ideation which they carry for weeks or months, as you say, but may be completely illogical or transient. Many people also have no mental illness at the time of suicide.
Hotspots are an issue because susceptible people form ideals about their preferred method of suicide, and the success of one person may validate it in the mind of another.
Hence, a fencing intervention, stopping someone killing themselves off the fast lines at Harrow and Wealdstone, will not just prevent a delay to trains but will in the long run actually save lives because many people who find their chosen suicide method to be unavailable actually don't go on to commit suicide at all.