[RPG] How to make finding an NPC interesting?

If ramping up tactics/strategy is what you're after I've got to suggest you take a look at "The Angry Guide to ... Combats", parts 1, 2, and 3. Some takeaways:

• Two of one monster and two of another is much more interesting--for all involved--than four of the first. Even if the CRs are the same any variation in abilities, ranges, &c. lead to strategic and tactical thinking, rather than toe-to-toe slugfest. On both sides of the fight.
• Make sure you have the objectives of the monsters/NPCs well-established in your own head before starting combat. When total annihilation isn't the sole objective on each side, things get more interesting. Remember, combat != encounter: combat is only part of an encounter.
• Terrain, terrain, terrain.

Secondly, there's an initiative variant in the DMG called Speed Factor (DMG5e p.270). This variant separates declaration and resolution of actions--don't worry, with practice it goes just as fast as individual initiative. One upshot of speed factor initiative is that the situation in-game feels like it's evolving much faster. Faster evolution=more chances to change tactics in interesting ways.

But combat doesn't just have to be about tactics vs. going toe-to-toe. I challenge myself to include more role-play in combat every session. (It's actually a written note-to-self on the backside of my table-tent.) Some of the actions this brings out:

• Intelligent monsters and NPCs call out directions and warnings to each other and address each other by name. They had objectives/motivations going into the fight, and every round the cost-benefit analysis changes.
• Throwing out a plot call-back or a plot hook can often get the PCs to put on the brakes and re-think their approach. Suddenly it's half the party saying "hold on!--these mooks might come in handy" while the other half's all barbarian-rage and seeking the Dark One's Blessing. What's more interesting than PCs in conflict during combat?

The first rule of RPG plots is that no plot survives contact with player characters. You cannot rely on players to do what you expect. If a puzzle has two solutions, the players will invent a third, and you will need to be responsive to that if you want to avoid the feeling of railroading.

Your main risks are as follows:

1. The plot requires the PCs to be captured, but given the chance, they may evade capture
2. The plot requires the captured PCs to attempt escape, but they might not even try
3. The plot requires the PCs to succeed in their escape attempt, but they may fail
4. You need the PCs to feel responsible for their own successful escape, but they not come up with a plan that justifies the success mandated by the plot

Have the campaign begin with the players already captured. Tell the players in advance that this is the starting premise of your game. It conveniently allows the PCs to keep their starting weapons and equipment (they're gladiators, so those are the tools of their job). This also explains how they know each other and prevents the very real danger that the PCs will win or escape from a fight they're supposed to lose.

Next, the PCs must be motivated to try to escape. If the players do not decide upon this spontaneously, you must create that motivation. Have an NPC gladiator give them a warning that the arena is planning to set the PCs in an unwinnable fight where they'll be slaughtered (against a golem, manticore or the like, as a grand spectacle), and must make a run to escape.

Let the PCs come up with their own escape plan, but if they're entirely clueless, have the NPC help (he knows how to escape, but needs the PCs because they're going to have to fight their way out). Be responsive to player ideas, even if you had something else planned.

If the players escape based on their own plan, they will certainly feel that they did it by their own wit. However, the next best solution is if they fight their way out, so that they feel responsible the success of their own escape.