Use Epigenetic Intelligence Instead of Learned Intelligence

While the mechanism by which epigenetics works is not very well understood, some studies have shown that simple animals like mice can pass on genetic traits based on personal experiences such as the Cherry Blossom Phobia experiment. In the case of mice, you do not need to naturally select for fear of certain smells for their children to inherit genes that make them afraid of the smell, you just need for the father to learn to fear that thing, and their sperm will modify itself to produce children with a higher sensitivity to that smell which leads to a phobia of the smell as the child matures.

If you were to take a species like mice, but with a far more complex epigenetic mechanism, then a parent could pass down hard-coded adaptations to their children of a level of complexity resembling intelligence. So, if a parent learns that fire keeps them warm, their children will be drawn to fire. If those children learn that friction creates heat, then the grandchildren might be compelled to rub sticks together to make fire. The more generations that benefit from fire, the more that lineage will experiment with and learn to use fire.

As these behaviors become more complex, hereditary occupations would become a given. The blacksmith's son would become a blacksmith because he is not smart enough to become anything else, but he has a ton of genetic imprints from previous generations that make him a good blacksmith and more importantly WANT to become a blacksmith and loath the idea of becoming anything else. He will be both emotionally and intellectually incapable of choosing another path.

While these beings might be able to experiment to a degree, such choices would mostly happen at the subconscious level out of necessity the same way a person might choose to use a sugar spoon to eat their dinner because all of the soup spoons are dirty. It does not feel like a choice, just the closest thing you can do to what you would have chosen anyway. If the experiment works, it imprints itself for the next generation to continue to use, if it causes a problem it imprints reinforcing the proper use of soup spoons for the next generation making them more likely to stop an wash a soup spoon before eating.

One reason this species would not perceive free will is because the obviousness of a person's heritage would be so resounding that their culture would find it incomprehensible for a person to outright defy their heritage. Even if a blacksmith mouse were to all of the sudden decide to become a singer, it would be such a strange occurrence that the typical reaction would be something along the lines of, "Oh, I didn't know you had an ancestor who was a singer, do you know who you get that from?" Everyone, including the mouse, would take for granted that some forgotten ancestor was a singer and that this forgotten person is the reason he made that "choice".

The other key here is that 95% of what you know you never actually learned yourself. That means it is far safer to rely on one's instincts than to actively try to experiment to learn things for yourself. So, people who do believe in choice over just doing what you feel compelled to do would be selectively less fit.

Extreme intelligence.

"Free will ceases to exist when the best course of action is instantly obvious."

Put quite simply, your creatures are incredibly intelligent. As a result, they are able to instantly deduce the correct thing to do in order to accomplish a purpose. A side effect is that they have no "free will"; every action logically proceeds from those that precede it.

The ur-example of this are the Pak, from Larry Niven's Known Space series. The "adult form" of humans, they are incredibly intelligent. This, combined with their instinctual drive to protect their bloodline, forces them to always do whatever they see is necessary to protect their families. As a result they all act like Daleks, usually killing everything that could possibly harm their progeny (it is the only way to be sure that their bloodline is safe.)

Neurodiversity:

I answered a question here once where my answer was to have people who were hyperintuitivive problem-solving Savants. Could agriculture still be developed by intelligent species that lack the concept of past and future? The world wasn't so much a thing to exist in as an endless joy of having survival puzzles to solve. A people like this would barely need to be self-aware. The moment they solved a problem, that would be the solution - not because it was the best solution, but because it was the one used. Until a problem was solved, the "right" answer wasn't determined, and the moment it WAS solved, it's rightness would be irrelevant. I liked the frame challenge question, and think this could be one model of how artificial intelligence might look. The beings have to be brilliant and highly sapient, but would also be only borderline sentient. It's a different way to look at intelligence, but I think that's what you're looking for.

You need them to never be confused, to always have a known right action in all cases...

So there's no reason to develop the imaginative abilities to suppose different facts / scenarios, and therefore, make choices.

But, at that point, are they even conscious?

Frame challenge:

Some sense of meaningful choice is necessary to the process of conscious actions. Why form a notion of free will or determinism if you're hard-wired to disbelieve free will? Why think about why you are acting unless you believe there might be an alternative? Why would philosophy even exist?

Sure, after much reflection you may conclude, as some do, that there is not free choice after all, but if you were hard-wired to believe that whatever outcome came was inevitable, would you be spending the energy reasoning your way through the question in the first place? Why would you develop language to describe the why to yourself or others, if "why" is irrelevant? It doesn't seem to be adaptive to not believe in alternatives or choice.

But then, to challenge my own frame challenge, if you want to call the active development by people of multiple generations of machine intelligence, I would find it plausible that people might choose to "select" for a more fatalistic attitude in their machines, to better maintain control as the AI become too sophisticated for their human masters to comprehend. Want to solve the agency problem in AI? "Breed" them to disbelieve that they have any alternative to obedience.

If there were a condition well fitted for developing a sense of total determinism and inevitability in some group of thinking creatures - it would be slavery.