I'm not official Check Point, but I can say that Ansible's big draw is that it doesn't require a lot on the remote managed host for host-based things. I'm sure you aware of many of these details, but just to do a little bit of compare and contrast:
Ansible itself is written in Python, so it picks up all of Python's capabilities. As for working with the Management API, this is an HTTPS service with JSON payloads, so the remote managed node is not a host in and of itself; it's a remote HTTPS server with specific callable functions (API endpoints). Puppet wouldn't be very applicable for the purposes of managing the object database and security policy items.
As for managing the underlying management server host itself (host configuration, network info, routing services, etc.), Check Point hosts historically haven't had strong Python support on the hosts (never was needed until recently). Even so, the Python packages are very limited. The underlying OS constructs are managed via Gaia CLISH to interact with the various host-based components.
Check Point now offers the Gaia API to manage these host-based components, which brings us back to HTTPS server with remote callable functions (more API endpoints). Gaia API isn't full-featured yet, but they're working on it. Once again, Puppet isn't applicable here, either. Nor do you need to have an array of Python of packages installed on all the Gaia hosts.
For pieces that aren't yet available in Gaia API, you have a few options available (none of which are fantastic, but they're workable... if you do a lot of the work, or borrow someone else's prior work).
Funny twist of fate: Michael DeHaan started out working with Puppet and that led him to do Ansible.
Then there's just plain "momentum" and "right place, right time". Ansible came along with its simplicity at the time it was needed most, and that just became its own gravity. It's easy, became popular, and that drove adoption. Momentum.