Organisms like humans have billions of cells. If you change a few cells in one tissue you still have lots of unchanged cells with the old DNA. Phenotypes are not always under the control of a single allele at just one locus in a DNA strand. So, no, it would be hard. There may be some benefit to implanting specially changed cells with modified DNA that affect a disease process in one very specific tissue. For example there are ongoing efforts to implant modified Islet of Langerhans cells in pancreatic tissue - the hope is to cure type I diabetes.
Still I would say overall that what you describe would have to occur when the human embryo has VERY few cells in order to"fix" phenotype problems. Not likely to happen in adult humans.
You realize that just because DNA changes in the embryo or the adult, that change does not guarantee the adult will have the desired results.
In other words, if you took Isaac Newton's DNA and put it into an early embryo would you get another Physics genius? Answer: no, close to a 100% guarantee. Why? Environment has a huge effect on how genetic information is manifested. He may end up as a couch potato gamer. Nobody knows at the outset. A lot of very intelligent people seem to wind up in otherwise weird circumstances. Not how we want them to wind up necessarily.
While this represents a difficult problem I wouldn't be quite as definite about saying no, there are a number of genetic conditions that do occur based on a single gene fault and yet which have a profound effect on the whole person. There is a lot of research aimed at replacing such genes in a large number of cells using viruses as the insertion agent. However this would also be limited to very few, specific conditions. As Jim suggests most human attributes involve multiple genes and control processes that have developed due to the action of the environment, some of which might even be pre-conception. Even if we develop the technology to change the genes in a significant number of cells, we still couldn't predict the effect, my prediction, based on nothing more than the tea leaves in my cup, is that it might very well end in tears.:)
Some phenotypes may may only involve certain kinds of cells which as a population might be easier to engineer with than others.
Some blood cell phenotypes might be changeable by: removing a few of the cell type of interest, doing genetic engineering on them in a dish, use whole body irradiation to kill those kinds of cells in the body (assuming this works on the cells of interest), reintroduce the mutated cells from the dish to recolonize the body resulting in a changed blood cell phenotype.
I think the OP might ask about gene transformation or transfection techniques in genetic engineering which surely can change the phenotypic expressions of a specific gene (e.g engineering the red rose trees to give birth to a new species that has only black roses). In humans you may consider this mediated gene transferring method as an example. And I don't think these techniques are used on unknown or little known genes.