Discovering which mutations are important is hard, but once you know what mutation to target, finding it is relatively straightforward and is done by software. You can actually test for 500K+ specific mutations for a few hundred dollars with a gene chip, you dont even need sequencing if you know what you're looking for. But to discover what mutations matter you need a lot of training data including expensive / hard to access tumor samples, and high quality clinical data showing how patients with different mutations fared.

The process of selecting cells to engineer is pretty complicated and labor intensive. Basically you have to get a patient to a clinic, draw blood (or biopsy), process the sample for shipment, ship it to a lab (often have just a few hours to get this done), then you do a number of steps to isolate the specific cell types you want. A lot of this is manual but isn't really novel, the techniques are pretty well established and straightforward, some probably could be automated but im not sure how much, and it would be expensive to automate. I think a few large manufacturers probably invest heavily in automation but most small companies or academic labs probably cant afford it