[Quality] Unify AI Assistant docs across solutions

[benironside]

Description

There's overlap and inconsistency among our AI docs. We are repeating ourselves with minor variations across different docsets, and there's useful information in some docs that isn't present in others, none of which is ideal. We should come up with and implement a unified approach to documenting these features.

Background & resources

We have AI Assistant-related docs in at least the following locations:

https://www.elastic.co/docs/solutions/security/ai/ai-assistant
https://www.elastic.co/docs/explore-analyze/ai-assistant
https://www.elastic.co/docs/solutions/observability/observability-ai-assistant
https://www.elastic.co/docs/reference/kibana/connectors-kibana/openai-action-type
https://www.elastic.co/docs/reference/kibana/connectors-kibana/bedrock-action-type
https://www.elastic.co/docs/reference/kibana/connectors-kibana/gemini-action-type

Which documentation set does this change impact?

ESS and serverless

ESS release

applies to multiple versions

Serverless release

Feature differences

API docs impact

None

Prerequisites, privileges, feature flags

No response

[benironside]

Current plan as of meeting with Mike last week is to move the security LLM connector content into the new AI section being created in this PR, and to link to it from the connector guides. This circumvents the issue that some of the connectors don't have a 1:1 correspondence with the guides that describe how to use them (the OpenAI connector is used for both OpenAI and Azure OpenAI, with different requirements and separate guides).