I have to disagree with some of the implications of the article. Older people (baby boomers or older) like my parents or mother and father in-law immediately look for a phone number when they encounter an issue because they "don't trust computers to be as up-to-date as a real, live person" (summarization). This tendency seems to decrease as a person becomes faster at looking up information online, but even my father that's in IT will still reach for a phone if he can't figure out what he's looking for within a minute. For myself, the resistance to calling is because of the dreaded wait time in a call queue.
It should also be noted that while the statistics unequivocally show a trend upwards in use of non-phone support channels, this still leaves roughly 40%+ of interactions by phone at the moment. The other important implication is that when customers do happen to call you, it will increasingly be because all other sources failed first which means your call center should be moving towards higher-skilled problem solvers than mostly tier 1 / level 1 support agents that might as well be replaced with an AI but aren't because voice systems are expensive for the results in conversation.
My fundamental point is that you while you should understand general trends, you should validate against your organization. AARP probably should not think about dropping half their phone agents exactly, but maybe they should revisit the idea in another 4 years as the general population becomes more educated and the technology-handicapped pass on.