A word on romanisation : while an important transformative evolution, it didn't necessarily meant a radical change of identity and culture, but generally a more or less quick creolisation of provincial peoples. Brittania and Pannonia are obvious exemples of original provincial romanities which didn't implied a latinisation but let with a lot of native features under Roman influence but not wholly disappearing. Even the more complete exemples of deep romanisation in Gaul "only" led to a radical transformation of local Gaulish identities into local Roman identities.

Latinization of Roman province was a multiform evolution, which its hard to adress trough a single cause, but nevertheless relied importantly in Roman Italy, Gaul and Spain on both a colonial network and the adoption of Latin or Roman rights by local native elites, who undergo Roman ervegetic policies creating both the frames for a "Roman-way-of-life" (which in Gaul pre-existed to the conquest), cultural romanisation (creolisation of cults, etc.) and eventually political romanisation (as in, identitying oneself as Roman). In Gaul and Spain, Latin Rights were quickly adopted at the provincial scale in the Ist century CE by imperial decision.

While Roman Africa didn't escaped this rough evolution, it remained more a local matter and slowly evolved until the late Ist century CE, and mostly in Proconsularis (altough colonies or municipes could be found in Mauretania and Tripolitania). Now usage of both Institutional and Vulgar Latin are well attested on the coast of North-Eastern Africa, so this situation (especially with the really important economic role of Africa in western Romania) didn't impaired a regional latinization on the coast : by the Vth century, coastal Africa was essentially latinized.

We know that St. Augustine needed a translator to address part of his flock which seems to indicate a native linguistic survival (maybe neo-punic, probably Berber) which attest the limits of this latinization but on the other hand, Gaulish probably survived in significant enough pockets in the same time.

The difference with Gaul and Spain is that, besides the coast, the inner and outer hinterland remained larger Berberized with a varying (but important) degree of Romanisation : there was not a clearly defined limit between "inner" (as in provincial) Mauri and their "outer" counterparts, which strengthened both a lesser latinization, but probably allowed a deeper range of romanisation as well, we'd have there something comparable to Britto-Romans with a probably latinized elite and a more or less creolized population.Long story short, the events of Late Antiquity demonstrate this attraction/repulsion between Mauri (Romanized Berbers) and Africani (Romans of Africa) which settled a complex dual relation that the Byzantine reconquest might have prevented to resolve in the form of a Berber-Roman ensemble. By the time of the Arab conquest, the African Romance speakers certainly survived on the coast (and the conquests might have fueled some phonologic influences on Hispano-Romance evolution).

Givin the lack of sources, it's hard to adress how African Vulgar Latin evolved : what we have (some scholar accounts, latin loanwords in Maghrebi Berber and Arab) seems to indicate Afro-Roman might have been tied to Sardinian language, with similar phonologic evolution. It's quite possible, giving the complex relationship between Africans and Berbers, that you might have rather a broad linguistic ensemble of Afro-Romance rather than an unified language.

The Arab conquest, which led to an important Arab presence in north Tunisia, had the direct consequences to force the decline of African-Romance speeches (like it did after the conquest of Spain quickly after it happened) but we know it still survived unti the XIIth century as the communautarian language of local Christians (like Mozarabic in Spain).There's still mentions after that, but it's not really clear when it died out.

Afro-Romance is certain to have existed, from African Vulgar Latin spoken in Eastern Roman Africa, probably surrounded by Berber speeches which still keep trace of Afro-Latin influence. Possible, as Dalmatian language, it was never really well rooted down outside broad coastal Africa but the presence of mixed Mauri-Africani ensemble point that there wasn't such a clear limit between both.What prevented its modern survival was the political and cultural dominance of Arab which put it in a diglossic at best relationship as it happened elsewhere in the medieval western/central Arabo-Islamic world, not unlike what happened with pre-Roman population on the coast.

Sources

- Les Maures et l'Afrique Romaine, Yves Modéran

- Late and Vulgar Latin in Muslim Spain: the African connection, Roger Wright

- Vandals, Romans and Berbers, Andrew H. Merrills

- The regional diversification of Latin, J.N. Adams

​