Executive Series
by Rod Edwards
Posted on July 2, 2026
Here’s something I never expected to say: email might be your cheapest channel.
For most of my career, that sentence would’ve gotten you laughed out of a contact center ops review. Everyone learned the hard way that email wasn’t the cost-saver it looked like on paper. You’d push customers off the phone and onto email, thinking you were saving money. Then you’d watch your cost-per-contact go up because of all the back and forth: collecting information, probing for details, re-clarifying, re-responding. An email chain that took four days and six exchanges to resolve the same thing a three-minute phone call could’ve handled didn’t save you anything.
So the playbook became: steer volume to voice where it makes sense, tolerate email because customers want the option, and accept that it’s going to cost you more. That became conventional wisdom, and stayed conventional wisdom for a long time.
And it wasn’t just an ops problem. Customers weren’t thrilled with email either. Waiting hours—sometimes days—for a reply on something they needed resolved wasn’t a great experience. The asynchronous nature of the channel cut both ways: convenient for non-urgent stuff, genuinely frustrating when something actually mattered. So nobody was fighting hard to keep email around.
But AI has flipped it.
The Operational Reality
When you put generative AI on your email channel—not just to draft templates, but to actually handle the response workflow—the economics change fast. Fix a mistake in your AI’s output once, and it’s fixed in every similar email from that point forward. Response time collapses. Quality stays consistent. The back-and-forth that made email expensive starts to disappear because you’re getting it right the first time, more often.
The math is pretty simple once you run it. When AI handles most of the email interaction digitally, the cost per contact drops well below what a phone call costs. The channel that was quietly the most expensive one you had starts looking like your cheapest.
And when AI is handling your email, it responds in seconds, removing the lag that kills CSAT for email workflows.
Companies that see this early start actively routing customers toward it. They can offer faster service over email because the AI responds quickly and accurately. Customers try it, realize it’s actually pretty convenient for routine stuff, and the loop feeds itself. That’s a different posture than “tolerate the email queue.”
On the operational side, this isn’t magic. You still have to run it well.
There are two ways this plays out. For defined, high-volume case types—order status, refund requests, account updates—AI can handle the full interaction from intake to resolution, no agent required. For everything else, AI supports the agent: faster drafting, better quality, less after-call work. Both are worth pursuing. The economics just look different.
The primary metric for email has always been emails per hour. Handle time gets messy for email because the workflow isn’t linear. Most brands, even ones running solid CRMs, can’t calculate accurate handle time for the channel. So you work with what you can measure: hours worked minus shrinkage, divided by emails handled. Benchmark your agents, identify your best performers, make sure quality is holding up, and you’ve got a baseline.
The simplest version of AI in this workflow is faster drafting. If your agents can produce a response more quickly, your emails-per-hour goes up. That’s it. There’s no reason not to be doing this. Customer service programs that aren’t using AI for email drafting in 2026 are operating with one hand tied behind their back.
We’ve been running templates in contact centers forever. You can still use them. But have the AI fill the templates out. You get the consistency and specific language your ops team has approved, and the AI does the actual work. The agent does a quick proofread, adds a human touch where it matters, and it’s done.
The more interesting layer is what happens beyond templates. For the right case types, AI reads the customer’s email, pulls the relevant account details, applies the appropriate policy logic, and sends a complete, accurate response. Fast, consistent, and auditable. For cases that need a human, the same capability gives agents a fully formed starting draft instead of a blank screen or a library of 40 canned responses that don’t quite fit. Either way, the heavy lifting is done.
AI Solves the Quality Layer, Too
The quality consistency piece is real. In any contact center, you’ve got agents who write naturally clear, professional responses, and agents who struggle with written communication—second-language speakers, people who are great on voice but not as comfortable in writing, people who are just moving too fast to polish. Traditional answer: training, style guides, templates. Those help but they don’t solve the real-time problem.
AI writing support solves it. An agent types a quick, functional response, hits refine, and what comes out matches the brand. Same information. Same resolution. But the customer gets the version that sounds like the brands they chose to buy from.
There’s also something worth noting about how customers respond to AI in text channels specifically. What we see is that they’re more accepting of automation in writing than they are on voice. A bot on a phone call can feel impersonal fast, but that same level of automation in an email or chat? Customers mostly don’t mind, especially when the response is quick and accurate.
Email was expensive because of structural cost and customer experience reasons that AI largely removes. Once those constraints are gone, the whole channel strategy changes fundamentally. The companies that benefit will be the ones who actually adjust their approach rather than just using AI to do the old workflow slightly faster.
If you’re still steering customers away from email by default, it might be worth running the numbers again.
Rod Edwards heads product, tech, and ops as Laivly’s Chief Operating Officer.
