There was a recent discussion in the Customer Experience Management group on Linkedin in response to an article by Ondra Synek entitled Why AHT is dead and how to do better.  https://www.linkedin.com/groups/74793/74793-6115418212273242112?trk=hb_ntf_COMMENTED_ON_GROUP_DISCUSSION_YOU_COMMENTED_ON#commentID_6123188289483661312
Let me start by recapping the author's position.  This is easily and best done using his own words from his summary:
  • Talk to your customers as long as possible and try to keep them more engaged
  • Keep measuring AHT and other efficiency metrics, but for management purposes only
  • Emphasize customer-centric KPIs such as Customer Satisfaction
  • Analyze customer interactions and use the results for continuous customer service improvements
  • Make your customers happier and they’ll be more loyal
  • Remember: Act like a human and your customers will love you
In the discussion that followed, there were a few opinions to the contrary. But most I think generally agreed with the author that 1) AHT was an important operational metric (key to resource planning, for example) but useless as an agent metric and 2) that business should not worry about AHT and should almost exclusively focus on CSAT.
Well, let me get this out of the way.   I completely disagree.  
To color that a bit more, I completely disagree with the first point, and I probably completely disagree with the second point, but can't fully commit to that position without understanding the business, how customers define value, make buying decisions, etc.  
In short, my two themes here will be that I think the original article and subsequent discussion did not have enough of a General Manager perspective and that it framed the role of the agents and how they deliver value to customers and deliver value for the business incorrectly.
Let's start with a thought experiment and take a short walk in a GM's shoes.   You walk into a meeting with the CEO and CFO to start the budgeting process for next year.  The business unit you lead had a great year last year and they tell  you so.  Despite that great year, they need more from your part of the business...the stock price is flat and every business unit and every function is getting squeezed next year.  They need your revenue to grow by 10% though the industry average is only 5% and they need you to become even more profitable. As you leave their office you hear them joking about how "no good deed goes unpunished!"  You try to muster a smile while your mind scrambles to figure out how you are going to achieve those daunting objectives.
Marketing already wants to increase their budget, maybe you will increase that even more to get those sales, but that is going to put greater pressure on the other parts of the business to cut costs.  If you  leave some features out of the product it will enable you to cut costs by redeploying development resources, but you wonder with less new features, will customers want to buy the new product at the increased rate you need to achieve your growth targets?  IT is advocating spending money on infrastructure...you will need to spend it in the next three years anyway and there is some short-term risk of outages which, if they occur, would have a huge negative effect on customers.  Can you delay that spend to cut costs this year? Is it worth the risk?
You see some options, but you know there are some very tough trade-offs ahead. As you are thinking through all this, your call center leader comes by and says, "We decided to change our strategy.  We are going to start "talking to customers as long as possible" to try to engage them and make them super happy.  We're not completely sure, but our guess is that this will mean at least a 15-2o% increase in AHT maybe more, which means we will need more agents in order to maintain our service levels (% answered in X time) because we don't want to increase customer frustration and the abandonment rate.  And the extra agents means, of course, that we will need to open another location to accommodate the additional heads."
For a moment, you allow yourself to fantasize about various forms of public execution, but you remember something you read in a leadership book and you decide to give your call center leader the benefit of the doubt.  You respond calmly, "Sounds interesting.  Can I ask what the business rationale is for this substantial, increased investment in service?  How do you think this will help us grow revenue and be more profitable...this coming year?"
The reply you get is, "Zappos does it and it's worked out pretty well for them."
Alright, a bit provocative.  And if it was too much, I apologize.  But I hope it illustrates a couple nuggets of uncomfortable truth.  
"Talking to your customers as long as possible" is not free.
First, "talking to your customers as long as possible" is not free.  While a lovely idea, you have to see that who-could-argue-with-it stance for what it really is: an investment decision.  And, it's an investment decision that the GM has to weigh against all the other investment decisions from all the other parts of the business that are also making their case on how their increased spending will help grow revenue and/or make the organization more profitable.  
There are many factors that create loyalty and drive sales.  Service is one of them.  But, depending on the business, service may not be the most effective way to delight customers and grow.  Don't get me wrong, talking to customers is always important, but call center leaders should not delude themselves into thinking they are at the front of some line for who best delivers value to customers.  Think like a GM.
Second potential nugget in our GM thought experiment:   there is nothing wrong with making investments and big bets, but please show me the business where you get to make big bets without a business case so I can short their stock.  
I love Zappos.   My wife and I use them all the time...OK, she more than I...and the call center interactions have always been positive.  But I am so weary of hearing the Zappos' approach breathlessly described in panacea-like terms.  
Yes it works brilliantly for them.  But look at their business.  Zappos doesn't make any shoes.  Service is their offering!  For Zappos and Cabella's and many more, the interaction with the call center is the penetralia of the customer buying decision and more important, at the heart of the decision to keep using them.  The question every GM and call center leader has to ask is whether call center interactions are at the heart of their customers' buying and repeat buying decisions.
The question every GM and call center leader has to ask is whether call center interactions are at the heart of their customers' buying and repeat buying decisions.
Is the call center critical to the buying decision and loyalty decision for DirecTV customers?  Do you defect to Comcast and give up your NFL Sunday Ticket and the ability to watch any football game because of a kerfuffle with a call center agent?  
Are call center interactions what leads to repeat purchases for United Airlines customers?  Are you really not going to fly direct and are you willing to pay more because they didn't make enough small talk and didn't engage with you enough the last time you called?
If you are going to propose loving customers to death and increasing handle time and increasing staff and tying up capital in facility leases and real estate, you'll get laughed out of the room without presenting a bullet-proof business case.  You have to have data you can eat, not hypotheses, not...well if this happens then hopefully that will happen, which could lead to this...causal probability chains. Your business case has to demonstrate, unequivocally, that the call center experience influences loyalty and word of mouth and ongoing buying decisions.  
And for our GM thought experiment, you'd have to be able to demonstrate that those service interactions influence buying decisions 1) in the near term, and 2) more than the other investment trade-off decisions the GM has to make.  But that might still not be enough.  Even with a strong business case, you still might get laughed out of the room:  investment decision meetings are very tough rooms.
Even with a strong business case, you still might get laughed out of the room: investment decision meetings are very tough rooms.
Extending our GM thought experiment, what is more likely to happen to the call center leader is that she is not only not going to be allowed to increase costs, she is going to have to lower them...substantially.  Under that directive, do you still want to raise AHT indiscriminately?  Is it even possible to do that and still cut your costs as you will be required to do?
A wiser direction might be to do a thorough analysis of the customer segments and call drivers and ask how does the call center create business value for this customer segment on this call type?  Is it to handle the call politely but as efficiently as possible or is it to delight the customer because this customer is so valuable or because this call type has a huge influence on buying decisions in demonstrable ways?  Or is it better to not talk to customers with this issue at all and just handle the call in the IVR?
If you have a valuable customer segment where hand-holding is critical, then by all means, give those customers the TLC they want and need.  Or if you have a critical call type, like a customer calling in to explore what piece of software is right for them, by all means, talk to them as much as you need to.  There is a good chance however you are going to have to offset those added costs somewhere else.  And some of those lower value call types identified in the previous analysis might be an option.  
If the customer is calling to activate their cell phone, change their address, make a reservation, do you still really think talking to them as long as possible is a good investment, especially in light of the other investments in added talk time you are making and in light of the cost-cutting mandate?
Modern cars are technological miracles and you and I are both willing to pay more for certain features, comfort, and safety.  
Why doesn't your car have titanium lug nuts?  Why doesn't it have a gold plated catalytic converter?  Because those "features" raise costs without influencing buying decisions.  There is near zero ROI for offering them.  The same is true for the bulk of call center interactions for certain businesses. And it is also true for certain call types even within businesses where service is very important to customer loyalty.
The GM perspective might have moved some of you who were in the "damn AHT, full speed ahead on CSAT" camp to at least slightly modify your position. But I am guessing others are not so easily convinced.  Let me get a little more tactical.
What happens when Mary the Rep and the peers in her pod decide they are going to gab away, keeping their customers engaged and happy?  Well, if the budget is fixed (which it always is) and we can't add staff and we are loathe to cancel training and lunches again, then dozens or hundreds or thousands of customers start to twist in the wind in the call queue.  What is happening to the satisfaction of the customers on hold while Mary et al. gab away?  
What is happening to the satisfaction of the customers on hold while Mary et al. gab away?  
Moreover, which has a bigger effect on CSAT:  a short wait time or extra TLC during the call?   I don't know the answer and in fact, I don't think there is one that is independent of the business, the product, industry standards, how customers define value and make buying decisions.  
But I do know this, you should have an answer before you let agents talk as long as they want to, because my experience is that you ignore the steam coming out of the ears of customers in the queue to your peril.  You'll get fired for a failure to lower a stubbornly high abandonment rate before you'll get fired for failing to raise middling CSAT scores.
Finally, one thing that was particularly disappointing to me about the original article and the discussion was how the agents' jobs were framed in a kind of either/or way. Either they handled the call efficiently or they talked away trying to engage and delight customers.
I'm not sure if my first thought experiment was helpful or not, but let me try another.  You have two agents.   Sally takes a transaction inquiry call for a credit card company and talks to one customer for 30 mins...polite, sweet, answers all the questions accurately, but goes beyond that and also makes small talk, laughs with the customer, establishes real rapport, and gets a 7 out of 7 CSAT w/ Agent rating after the call.  
During that same 30 mins, Jane takes 6 transaction inquiry calls, is polite, answers the questions, asks if there is anything else she can help with and closes the call.  She averages 6 out of 7 on post call CSAT w/ Agent ratings.
Some of you are howling that this is a ridiculous and an extreme example, but I'll argue here that it is not.  
What if instead of this narrow...it's either CSAT or AHT...view, we reframed the agent's job.  What if we saw agents as little "factories" whose job was to "produce" happy customers.  There will of course be differences in the quantity and quality of the output from each "factory",  but I think we can agree that ideally our little agent factories are producing lots and lots of happy customers, every day.  
At the end of the day, if Agent 1 produces more happy customers each day than Agent 2, on this one measure alone, isn't Agent 1 more valuable to your business? This is not a trick question.
How does an agent produce more happy customers?   She has to think about both CSAT and AHT.   She has to think about the "state" the customer she is talking to is in.  Is the customer in a hurry?  Agitated?  Lonely?  Is the customer confident the issue is resolved?  
How does an agent produce more happy customers?   She has to think about both CSAT and AHT.
She has evaluate her assessment and decide if doing a little more hand-holding will produce a significantly happier customer or if it would be better for the business if she ended this call, rescued another customer from the queue, and attempted to produce a second happy customer.
When I ran a call center, this is exactly what we did.  We told agents their job was to produce 1) more, 2) happy customers and we paid them based on how well they did just that.  We measured calls per staffed hour, which is, of course, AHT (A rose by any other name...) and paired that with a measure of CSAT. Agents who produced more happy customers earned significantly more money than agents who took a lot of calls but had mediocre ratings and also more than agents who took few calls but rang the bell on CSAT.
Seen in this light, an agent is like anyone in management who has the challenge of making customer/shareholder trade-offs.  Our agents had to make a decision, in the moment, if more talk time would lead to more satisfaction or if they had done what they could and would be better off trying to make a different customer happy.  
These are not easy decisions... how to balance short-term, long-term, customers, shareholders, more touch for one customer vs. touching many customers...not easy for management, not for agents, not for anyone.  But the reality is your agents are making these decisions right now.  
Either/or is not what is needed. What is needed is guidance on how to balance both.  Agents need training, not just on the product and how to answer service questions and how to be polite, but how to make these....customer vs. shareholder, this customer vs. the next customer...trade-offs.  
As an aside, the agents who do this well in my view should be paid more...a lot more...than the average agent.  They are an incredible asset to the business and the added salary expense has a demonstrable ROI.
Mercifully, let me close.  I probably tried to cover too much ground here with too many disparate perspectives.  And if in doing so, I confounded the debate, I apologize.  
Having been down it more times than I care to admit, I am well aware of what the Road to Hell is paved with, but my intentions really were good.  
My hope is that now:
  • CSAT and AHT are not pitted against each other in an either/or fashion and any sense of moral high ground about this matter is, well, a little bit lower
  • Changes in call handling procedures are seen as investment decisions...that must be 1) viewed in the context of other decisions on how to best deliver value to customers...value that influences buying decisions, and 2) supported with ROI calculations tailored to your business and customers...not based on what works for a different business with a different customer value chain
  • The role of agents is framed in a more nuanced way...where CSAT and AHT are both seen as essential customer(broadly defined)-centric KPIs
  • There is more clarity around the need to provide a new kind of training and support to agents...training that will help them do a better job of making the, in the moment, business decisions they confront every day.

There was no intention that this be the last word in this discussion.   So please weigh in on what you agree with and where you think I went afield.  It helps all of us learn.