JonW's answer is a good one, and addresses the issue of whether a textual equivalent should be provided for the infographic as a whole, but I don't think it answers the real question you are asking here, which is about how detailed that description should be, and in particular, whether incidental graphics in the infographic should be described.
Generally speaking, the guidelines for alternative text for any image is that the text should focus on providing a succinct textual semantic equivalent of the graphic. For example, images that are fancy rendered text typically ("New!" in a fancy font with shading) have an alt text of just the text and don't describe the rendering; an image of a VCR-style play button is typically "play" instead of "right pointing triangle", and images of purely cosmetic images typically have empty alt text.
More details on how to create good ALT text for images can be found at this (WebAIM) and various other sites.
Applying these guidelines to infographics suggests that you should only describe the graphics if they are actually (a) conveying information that is (b) additional to the information conveyed by the text.
This means you can ignore any 'theme' graphics elements right off the bat.
If a graphic element is redundant with respect to text, you may be able to ignore that also. For example, in the case of text such as "80% of infographics are uninteresting" that has a pie-chart showing what 80% looks like, the pie chart is redundant with respect to the text, so does not need to be described separately: the text already covers that information. (However, if the graphic would cause a sighted user to infer particular emphasis on some point, ensure that the text also reflects this.)
From what I've seen of infographics so far, most of the graphics on them tends to fall into one of these first two categories, and while it adds to visual appeal, doesn't add additional semantic information, so doesn't need additional text.
The tricky cases with infographics are those cases where graphics are actually conveying information that is not conveyed in the text alone: these tend to be line graphs or comparative charts, and need to have a textual description of their own that conveys the information that a sighted user would see. This is one case where you need to take the goal of the chart into account, and come up with a description that makes sense in that context. For example, rather than describing a line chart as an exact series of numbers, describe the point it is illustrating: "sales revenue peaks in 2001, and falls to half two years later".
Or, long story short, your view on this (versus your colleague's) is the right approach. And a for what it's worth, one place where your colleague's level of detail might be appropriate is if the infographic was appearing on say a design blog site that was actually discussing the layout of the specific infographic itself: context matters a lot in terms of how images should be described. But in the usual case, the text should stick to the message(s) of the infographic, not the presentation.