Yeah - i remember it mentioned in the Air-Land Battle concept from 70's - 80's but Google is not readily supplying a reference for that.
Found an AU article on the subject though that supplies a good analysis on the difference between BAI and AI, basically saying that BAI is in support of friendly forces engaged but does not require detailed coordination ala CAS where AI is more upstream, preventing enemy forces, support or logistics from being brought to bear.
Article: https://www.airpower.maxwell.af.mil/airchronicles/apj/apj90/spr90/2spr90.htm
For this next generation mission driving the need for a Light to Scalable capabilities platform IMO, the doctrine has to be written to give this more footing, LAAR seems to get dismissed as a niche capability but it is not. This is an on-going, established requirement again IMO that is demonstrated from our 15+ years of COIN / LIC which is now morphing into Grey Zone Combat Operations.
Not exactly persistent ISR with kinetic effects capability if called, we've got that covered with Tier II RPAs, MQ-9 and successors.
Not exactly traditional Attack as it is longer in duration and target development (usually) than receiving a call for fires message and delivering effects with the main concern being friendly deconfliction, currently covered with several platforms, hopefully with an A-X dedicated platform in the future.
The new mission is air operations conducted on a repetitive not persistent scale, tailored effects as required, usually in permissive environments but capable of up to low+ / moderate-, network and comm focused for dynamic collaboration while capable of independent operation from C2 and with a flexible logistical footprint to allow for operational flexibility.
That's just my musings but looking at that, you come back to a two crew manned platform with room for growth, basically Scorpion.