Humans of Support is our interview series featuring the people building and running customer support teams. Each episode brings a different perspective from the industry.
For this episode, we’re talking to Aden Ritz, USA Growth Partner at VirtuHire, where he focuses on retention, expansion, and partnerships, and works closely with AI and automation tooling (Claude, OpenAI, Zapier, n8n) as part of how VirtuHire staffs and supports lean US teams with remote talent.
Aden has sat on both sides of the customer relationship closing deals and then owning the account afterward and that vantage point shapes almost everything he has to say about support: that most of what determines whether a customer stays happy was decided before the first ticket was ever opened, that automation doesn’t shrink support teams so much as it concentrates the hard work onto the people left, and that the most dangerous thing in support right now is a dashboard that looks fine while the relationship quietly falls apart.
Here’s the full interview with Aden Ritz below.
You've sat on both sides — closing deals and then owning the accounts afterward. What did sales teach you about support that most support leaders miss?
That the customer's opinion of you is set long before they ever open a ticket.
Aden Ritz, USA Growth Partner, VirtuHire
In sales you learn exactly what a client was promised, which is usually more specific than what support ever gets told. Then you own the account and watch a support team answer a question competently while completely missing what the person was actually worried about. The ticket gets closed. The relationship does not improve.
Most support leaders optimize the ticket. The thing they miss is that support is where the promise made during the sale either holds up or falls apart, and nobody on the support side was in the room when the promise was made. If your support team cannot see what was sold, they are working blind and their metrics will still look fine.
What support metric is overrated, and which deserves more attention?
Overrated: First response time.
It is easy to measure, easy to game, and a fast reply that resolves nothing just moves the frustration later in the day. Teams hit the SLA and lose the customer anyway.
Underrated: How often a customer has to explain their situation more than once.
Repeat explanation is the single most reliable predictor of a churn conversation, and almost nobody tracks it. It exposes broken handoffs, context that never got passed between tiers, and tickets closed without being resolved. Every other metric hides those.
What's in your stack that you'd fight to keep, and what would you replace tomorrow — and why?
Keep: Anything that preserves context across a handoff.
The tool matters less than whether the next person can see the whole history without asking. Most of the damage in support is done at handoffs.
Replace: Any tool where the reporting is prettier than the workflow.
There is a category of software that produces beautiful dashboards for work, which makes it harder to do. If your team has built spreadsheets alongside the tool, the tool has already lost, and the dashboard is measuring the wrong thing.
You implement AI and automation workflows alongside the staffing at VirtuHire. Does automation reduce the number of people a client needs, or change what those people do?
It changes what they do, and the honest version is that it usually does not reduce headcount.
The pattern we see placing staff with lean US teams is that automation takes the volume that never needed a person: order status, password resets, the tier-one layer. What is left is denser. Every remaining ticket is one the automation could not close, which means it is a judgment call, an exception, or an upset customer.
So the job gets harder, not lighter. Teams that automate tier one and then hire cheaper for what remains tend to go backwards, because they automated the easy half and then staffed the hard half like it was the easy half.
What's a support workflow you've automated for a VirtuHire client that you'd point to as the cleanest win?
The cleanest wins are boring: the handoff between an automated first touch and the person who picks it up.
Where it works, the automation collects everything the human will need before the human arrives, so the customer never repeats themselves and the agent opens the ticket already knowing the account, the history, and what was tried. Where it fails, the automation just delays the same conversation and the customer starts over, which is worse than no automation at all.
I would rather point at that than a volume number, because volume numbers usually reflect deflection, and deflection is not the same as resolution.
Where's the line for you — which support tasks should an AI agent never touch?
Three, and they are the same category underneath.
Anything where the customer is already angry. A bot cannot de-escalate, it can only be one more thing that does not listen.
Anything involving money moving the wrong way. A billing error, a double charge, a refund that did not arrive. People experience those as a trust problem, not a process problem, and the answer has to come from someone who can be held responsible.
Anything where the correct answer is that the company got it wrong. An AI will not concede that cleanly, and a hedged apology from a machine reads as a company hiding.
What's the most overhyped claim you hear about AI in support right now?
That it deflects some large percentage of tickets.
The number is usually real and usually meaningless. Deflection counts the customer who did not reach a human. It does not distinguish the person who got a good answer from the person who gave up. Those look identical in the dashboard and completely different in the retention numbers six months later.
AI: Boon or Bane?
Boon, and it is not close. But it is a boon the way a power tool is: it makes a competent operation faster and a sloppy one worse, faster.
AI applied to a support function that already has clear ownership and clean handoffs is transformative. Applied to one that does not, it industrialises the mess and gives you a dashboard saying everything is fine.
Where do you think human support still matters in five years, no matter how good AI gets?
Wherever the customer is frightened, angry, or about to leave.
An AI resolves a shipping delay or a password reset faster than any person will. But once someone feels dismissed, a bot compounds it, because what they want is evidence that a human considered their situation. The cases stay stubbornly human: high-value escalations where the client expects a relationship and not a ticket number, emotionally loaded situations like a billing error or a missed order before an event, and anything where the right answer needs judgment no decision tree covers.
Those are not edge cases. They are the small number of interactions that decide whether the account renews.
If you had to describe great IT support in one sentence, what would you say?
The user gets on with their day and never finds out how much had to happen for that to be true.
AI Agent
Asset Management
Knowledge Base
Automations
SLAs
Customization
Ticket Assignment
Reports & Analytics
Approval Management
Helpdesk Security
Explore All Features
Banking and Finance
Healthcare
MSP
Government 
Non Profit
IT Services
Customer Support
Human Resources
Facilities
