Ammonoids belong to the cephalopods, and hence to the same group of mollusks as modern octopods, squids and nautilus. Indeed, it is generally accepted that ammonoids were more closely related to octopods and squid than nautilus. As such, we can safely take as a starting assumption that those features shared by modern cephalopods were also present in ammonoids, such as a muscular siphon for propelling the animal, and an array of arms or tentacles surrounding a central mouth. But how many tentacles did ammonoids have? Squid and octopods have eight or ten arms, but nautilus have many more, about ninety. Because nautilus bear a superficial resemblance to early cephalopods in retaining an external shell, it has been tempting to assume that they are more primitive than octopods and squid, but there are good reasons to believe that the supernumerary tentacles of nautilus are a derived peculiarity of that group. Arm development in cephalopod embryos begins from ten original buds in both nautilus and squid, with these buds becoming divided in nautilus (Klug & Lehmann 2015), suggesting that the lower number could be the more primitive. With ammonoids on the squid line rather than the nautilus line as mentioned above, it seems likely that they retained the primitive arm number like their sister group. In their review of preserved ammonoid soft-tissue remains, Klug & Lehmann (2015) noted that there is only a single known fossil ammonoid (going by the memorable name of GSUB [Geosciences Collection, University of Bremen] C5836) that might include preserved arm tissue, but the area in question shows little more than a tarry smear. Trace fossils have been used to argue for a low tentacle number in orthocerids, a group of Palaeozoic cephalopods commonly believed to include the ancestors of both ammonoids and squid, but again the evidence is not enough to be conclusive.
If we do presume that ammonoids had a squid- or octopus-like number of tentacles, can we then interpret ammonoids as basically a squid in a coiled shell? This may be the most common representation of such animals:
Unfortunately for Akane's purposes, ammonites may not have provided much in the way of good eating. Whereas the fossil record of ammonoid tentacles themselves is next to nonexistent, we do have a bit more evidence about the arrangement of an ammonoid's mouthparts. Living cephalopods usually have a hardened beak at the opening of the mouth, with the ribbon-like radula sitting directly behind it. The majority of tearing and crushing of food is done by the beak; the radula mostly functions to pull food particles back into the gullet. In basal ammonoids, the beak was more or less similar to that of a recent cephalopod, but in the derived ammonites* it became quite modified. Ammonites possessed a broad structure near the opening of the body chamber that is called an anaptychus or aptychus according to its configuration (though just to confuse matters, the term 'aptychus' seems to sometimes be used to cover both types). An 'anaptychus' was a single chitinous, semi-circular plate; an 'aptychus' was a calcified, bivalved arrangement. The aptychi were not directly attached to the main shell and may commonly be found as isolated fossils. Examination of aptychi that have been preserved still in their original body chamber has lead to the widely held conclusion that they represent a modification of the original lower jaw of the beak. Meanwhile, the upper jaw became reduced and weakened in ammonites with aptychi (Tanabe et al. 2015).
*A quick explanation about 'ammonoid' versus 'ammonite': 'ammonoids' are a particular group of shelled cephalopods that first appeared during the Devonian. 'Ammonites' are a particular clade within the ammonoids including most of the Mesozoic species (a small number of non-ammonite ammonoids survived into the Triassic). So all ammonites are ammonoids, but not all ammonoids are ammonites.
Because they often have a similar configuration to the opening of the ammonite's shell, the aptychi have often been interpreted as functioning as an operculum for when the animal retracted itself into the body cavity, presenting a tough barrier to any would-be predator. Certainly the reduced upper jaw meant that they could not function as a beak to bite into food (though some Late Cretaceous ammonites did exhibit a re-enlargement of the upper jaw and may have regained their bite). However, if aptychi functioned as opercula then the tentacles of ammonites could not have sat in quite same arrangement as in modern cephalopods. They could not have completely surrounded the mouth because then they would have prevented the operculum from closing. Perhaps some of the lower tentacles were lost, or perhaps the base of the circle became divided. Some authors have argued that aptychi were jaw structures only, with no operculum function, but I confess I find it difficult to understand their purpose in that case.
That most ammonoids were not subjecting their food to strenuous chewing is also indicated by the structure of the radula: where known, the majority of ammonoids had radulae with high, slender teeth more suited to grasping than rasping (Keupp et al. 2016). The overall indication is that most ammonoids were probably micropredators, feeding on small plankton such as crustaceans; where possible stomach contents have been identified in ammonoid fossils, they have also supported this conclusion. The modern nautilus has a similar diet, and ammonoid arms possibly did resemble nautilus tentacles in being short and slender rather than long and muscular (though at least one author has discussed the possibility of ammonoid arms being expanded into broad fans for the capture of plankton). The Late Jurassic ammonite Aspidoceras had a much more robust, powerful radula than is known for other ammonoids but may provide something of an exception to prove the rule: its stomach contents are dominated by the pelagic crinoid Saccocoma, suggesting that it was still a planktivore even if it was tackling tougher prey than its relatives (Keupp et al. 2016).
So to sum up, ammonoids probably had only a small number of tentacles, no more than ten at the most. They were probably slight affairs, suited for sweeping small or poorly motile food objects out of the water rather than grabbing and manipulating struggling prey. A planktivorous habit for ammonoids would also seem to fit with their predominance when they were around; after all, there's no shortage of plankton in the sea.
Keupp, H., R. Hoffmann, K. Stevens & R. Albersdörfer. 2016. Key innovations in Mesozoic ammonoids: the multicuspidate radula and the calcified aptychus. Palaeontology 59 (6): 775–791.
Klug, C., & J. Lehmann. 2015. Soft part anatomy of ammonoids: reconstructing the animal based on exceptionally preserved specimens and actualistic comparisons. In: Klug, C., et al. (eds) Ammonoid Paleobiology: From Anatomy to Ecology pp. 507–529. Springer Science.
Tanabe, K., I. Kruta & N. H. Landman. 2015. Ammonoid buccal mass and jaw apparatus. In: Klug, C., et al. (eds) Ammonoid Paleobiology: From Anatomy to Ecology pp. 429–484. Springer Science.