ExpertCallers CEO Antony P. Gregory spent 26 years selling seat-based contracts. He now runs a two thousand-person operation and sells against the model that built it, and thinks smaller service businesses are next in line.
Most founders running an agency, a dev shop, or a consultancy don’t think of themselves as being in the outsourcing business. Antony P. Gregory would argue they are, and that the reckoning arriving in his industry is arriving in theirs next, just more slowly.
Gregory has worked in outsourced customer support since 2000. Writing in Forbes India in July, he called the split now running through his sector the Great BPO Divide – companies that sell labor on one side, companies that sell judgment on the other, and a five-year window to pick.
I asked him what that divide looks like to a founder whose proposals still quote hours.
Me: You’ve Framed This As An Outsourcing Story. Why Should A Founder Running A Small Agency Care?
Gregory: Because the mechanism has nothing to do with outsourcing specifically. It has to do with what you’re charging for. For thirty years my industry sold hands – move the work to where bill by the hour or the seat, add margin. That worked because the unit of value was a person doing a task. Automation deleted that assumption.
If your proposals quote headcount, hours, or seats, you are selling the same thing I was selling. The only difference is that the pressure reached me first, because my work was documented, measurable, and high-volume. That’s the profile software absorbs first. It isn’t unique to BPO, we were just standing closest to the door.
Me: Is The Argument That Ai Is Overhyped?
Gregory: The opposite, and I want to be clear about it. When software can resolve a routine customer contact for a few cents, that’s not hype, that’s a price. If you handle high volumes of genuinely repeatable work and your documentation is clean, you should be automating that tier, and plenty of companies with exactly that profile are moving too slowly.
The problem isn’t that it doesn’t work. The problem is that it works on the repeatable half of what you sell, and most firms haven’t repriced the other half.
Me: How Does A Founder Know Which Side Of The Line They’re Standing On?
Gregory: Apply one test to each service on your list. Could a competent new hire do this correctly after reading your process document?
If yes, you’re selling labor, and your ceiling is set by whoever bids lower – increasingly, that isn’t a person. If doing it correctly requires knowing which exception matters, which client will escalate, which technically-correct answer is commercially the wrong answer, that’s judgment. That’s what belongs on the invoice.
And if nobody at your company can say what share of the work falls into each bucket, start there. That’s a documentation problem. Those are much cheaper to fix than pricing problems.
Me: What Does Automation Actually Leave Behind?
Gregory: The hard part. The furious enterprise account, the complex claim, the case that’s already failed one resolution attempt, the customer deciding whether to stay. Those don’t get absorbed, because the work isn’t executing a step, it’s deciding what matters.
That’s not labor. That’s judgment, and judgment is the one thing you can’t buy for pennies. But here’s what founders miss: your average cost per unit of work goes up after automation, not down, because what’s left is the difficult remainder. If you’re still billing that remainder at the blended rate you set when easy work subsidized it, you’re now losing money on your best people and calling it efficiency.
Me: You’ve Written About An “Invisible Invoice.” What Is It?
Gregory: Every cost the labor model doesn’t bill for. When you pay by the hour, your vendor earns the same whether the work worked or not, a resolved complaint and an unresolved one bill identically.
A customer calls with a problem. The agent closes the ticket fast. On paper it’s excellent: resolved, short handle time, good score. But the problem was closed, not solved. A few months later that customer quietly buys from someone else. Nobody sends you an invoice saying “you just lost a customer.” It arrives late and disguised, in next quarter’s churn. Sell activity and you optimize the cheap number while the expensive one compounds in the dark.
Me: So What Replaces Headcount Pricing?
Gregory: In practice, a hybrid. A fixed fee that keeps the operation running, plus an incentive tied directly to the result the client actually cares about – retention, resolution, conversion, whatever the real objective is. When your partner only wins if you win, the client has finally bought the right thing.
I’m not claiming hourly work disappears. Plenty of procurement still buys hours and headcount deliberately, because outcomes are hard to write into an RFP, and there’s margin in it. But it has become the floor rather than the ceiling – the work you take to keep the lights on, not the work that decides what your company is worth in five years.
One condition, and founders get this wrong out of eagerness: only accept outcome terms for work where your judgment actually moves the outcome. Take a resolution-linked contract on a queue that runs itself and you haven’t repositioned, you’ve discounted. Take it on the escalations, the ambiguous cases, the accounts deciding whether to renew, and you’re being paid for what you’re good at, probably for the first time.
Me: What Changes About Hiring?
Gregory: The question stops being “should a human be in the loop” and becomes “where is a human worth the most?” Fewer people, in harder places.
Most service businesses recruit people who follow steps reliably, because that’s what volume work needs. What’s left after automation needs a different profile – someone who can sit with ambiguity, decide without a script, and be trusted with a client relationship on a bad day. Same job title, different job. You can’t retrain your way there in a quarter, so start hiring for it before you need it.
Me: What Would You Tell Yourself In 2015?
Gregory: That the labor contract was never the business – it was the entry point. Low-cost capacity is how you win the first engagement. Judgment is the only reason anyone keeps you after the third. I’d have started charging for the second thing much earlier, while the first thing was still paying the bills.
That’s the window most founders are in right now. It doesn’t stay open.
Author Bio
Purnendu Bala writes on AI Governance, technology, and go-to-market strategy. His work has appeared in The Next Web, Search Engine Journal, Medical Economics and The Financial Brand.
