Solution

The image in the task seems to be irrelevant, at least on the surface. From the word document, you can extract a table that contains 10 rows and 3 columns. The columns are named "horizontal", "vertical", and "thing". Half of the rows contain images in the first two columns, and the rest are empty. All of the values in the "thing"-column refers to one of the primary colors red, green, or blue. The images that can be found in some of the rows each represent some number. Each pair of numbers have some theme (roads, bus lines, elements, albums, card games), and represents a pixel in the image. Going to the respective value for the referenced pixel and color, you will find values that happen to be possible to interpret as ASCII for the letters S?N?TA??O?. Using some pattern matching method, you'll find that the only English word that fits is sanitation.