A high bluff skill, using the SRD, is mind control, but not as much as Diplomacy is.
The DC for suggestion is +50 of the base DC, according to epic usages. Presuming that the suggestion is something like "You're an unimportant character in a game, ponder what that means for your life." (+20 on their sense motive due to The bluff is way out there, almost too incredible to consider.), and the bard would have a roughly even chance of instilling that suggestion via bluff.
Technically speaking, The bard would have +25 (skill/charisma) +20 ring, +30 glibness, +7 luck via spell improvisation versus a level 5's guard's sense motive. Let's be charitable and call him a fighter instead of a warrior.) so 4 ranks (cross class) of sense motive + 2 wisdom (he's wiser than the average guard), +50 (suggestion) +20 ("way out there") = +82 bluff versus + 76 sense motive = quite decent odds, really without too much of an investment)
3.5 isn't very good at social skill modelling. At all. You may want to implement various house rules like the GitP diplomacy overhaul. Beyond this, all social skills boil down to "mother may I"-ing the DM, so having some framework is desirable.
How social skills (magically enhanced and otherwise) should be worked out before the event in question, such that everyone's expectations are roughly the same of what can be accomplished. If we take epic bluff at face value, however, let's see how it stands with other capabilities of 14th level spellcasters.
The suggestion, as written:
Instill Suggestion in Target:
This is identical to the effect of the suggestion spell, except that it is nonmagical and lasts for only 10 minutes.
This ... is pretty cool. It's a no-save suggestion that's explicitly non-magical. The save is the sense motive roll rather than a will save.
A bard 14 (presuming no PrCs) could otherwise bypass the guard by, using suggestion via bardic music (wholly underwhelming compared to this option), sneak past with hide, move silently, and greater invisibility, Dominate the guard, Hypnotise the guard, Deep Slumber the guard, Modify Memory the guard, etc. The Forgery skill, at the relative skill differentials between the bard and the guard would be able to produce impeccable credentials, and so on. Thus the save odds will be different (skewed towards the direction of investment of the character), but this certainly isn't the most effective or even most subtle method for bypassing a guard in the character's repertoire.
This option represents the casting of multiple spells and the significant investment of character resources, and should be respected for what it is. Otherwise, the route that provide for minimal "mother may I"-gating is for all the characters to solve their problems through murder or spellcasting. Allowing this as a third route allows the characters more options, even without the introduction of interesting and coherent house rules.