The Cost-Cutting Lie Executives Feed The Press
When enterprise executives announce that they are slashing thousands of customer service jobs because "customers prefer chatbots," they are lying to you. More accurately, they are lying to their shareholders to hide a desperate, short-term margin play.
Centrica, the parent company of British Gas, axed 1,300 call center workers under the banner of modern automation. They claimed people simply do not want to talk to human beings anymore. That assertion is nonsense. Meanwhile, you can explore other developments here: Why Red Sea Panic Will Not Push Crude Oil to 100 Dollars.
Customers do not prefer automated text boxes. What customers actually prefer is getting their problems solved on the first attempt without sitting on hold for forty minutes to listen to pan flute music. When a company cripples its human support teams, creates artificial phone queues, and buries its contact numbers under seven layers of web menus, of course chat usage goes up.
You did not adopt their digital assistant because you loved it. You used it because they burned down every other bridge to human interaction. To explore the complete picture, we recommend the detailed report by Harvard Business Review.
Call this what it really is: financial engineering masquerading as operational progress.
The Illusion Of Deflection Metrics
I have watched corporate leadership teams throw tens of millions at automated support platforms over the last decade. The board pitch is always identical. Slide decks promise 80% inquiry deflection, reduced overhead, and higher customer satisfaction scores.
Here is how the trick works behind closed doors.
When a customer opens a web page, fights with a text prompt for six minutes, gets routed to an irrelevant FAQ article, and rage-quits the browser, the software logs that interaction as a "successful self-service resolution."
The metric looks clean on an executive dashboard. The software vendor collects their bonus. The chief operating officer gets to parade a lower cost-per-contact number in front of institutional investors.
Meanwhile, in the real world:
- The customer's broken boiler remains broken.
- Their anger doubles.
- Their lifetime customer value drops off a cliff.
- They quietly switch to a competitor the moment their contract expires.
You cannot run a enterprise on contact deflection alone. Deflection measures how effectively you push your customers away; it says nothing about whether you fixed their issue or earned their loyalty.
[Traditional Support Model]
Customer Problem -> Human Agent -> Solved Issue -> Retained Customer Value
[The Enterprise "Efficiency" Model]
Customer Problem -> Automated Bot -> Unresolved Frustration -> Lost Customer
Why Energy Utilities And Telecoms Get Away With It
If an e-commerce brand or an independent SaaS provider tried this stunt, they would go bankrupt within eighteen months. Churn would eat them alive.
Monopolistic or oligopolistic industries like energy utilities, telecoms, and legacy airlines operate under a completely different incentive structure. In these sectors, service quality across the entire market is uniformly poor. When British Gas breaks its service promises, where does the customer go? To another giant provider running the exact same cost-cutting playbook?
These companies execute a calculated gamble:
- Fire human staff to immediately strip out payroll costs and pension liabilities.
- Deploy cheap conversational software to absorb the incoming wave of complaints.
- Accept the drop in consumer sentiment because switching costs for the customer are high, tedious, or artificial.
- Pocket the short-term profit margin boost to keep equity analysts happy for another quarter.
It works on a balance sheet for twelve months. It is poison for a brand over twelve years.
The Economics Of Human Support Vs Digital Automation
Let us look at the actual numbers instead of public relations handouts.
Deploying basic conversational models across an enterprise call infrastructure costs a fraction of maintaining a human workforce. An offshore or onshore support agent costs between £25,000 and £40,000 annually when factoring in hardware, management overhead, utilities, and benefits. A text bot processing thousands of simple requests per hour costs pennies per session.
For routine tasks—updating a billing address, checking a balance, confirming a meter reading—automation makes total sense. Nobody wants to wait on hold to change an email address.
The catastrophic failure occurs when executives confuse simple administrative tasks with complex, high-friction problem resolution.
Imagine a scenario where an elderly customer receives a rogue energy bill for £4,000 during a freezing winter month.
An automated assistant responds with pre-programmed scripts, extracts key terms, and spits back a link to "Understanding Your Tariff." The customer panics. The software fails to register nuance, emotion, or systemic billing glitches. The problem escalates from an accounting error to a public crisis.
A trained human rep, empowered to exercise actual judgment, solves that issue in five minutes, saves the account, and protects the firm's reputation.
Automating the easy stuff is efficient. Automating the complex stuff is brand destruction.
The Real Winner Is The Enterprise Software Vendor
Who actually wins when 1,300 support workers lose their jobs to a script?
Follow the money. The winners are not the consumers, who now spend twenty minutes trying to rephrase their emergency repair requests to pass an automated gatekeeper. The winners are not the remaining human workers, who get overwhelmed with the high-stress cases the software failed to handle.
The winners are the enterprise software vendors who charge massive recurring license fees while convincing non-technical CEOs that software can completely replace human empathy.
These platforms sell the dream of zero marginal support costs. CEOs buy it because human labor is messy, expensive, and demands fair working conditions. Software does not complain, join unions, or ask for cost-of-living adjustments.
Stop Measuring Cost-Per-Call And Start Measuring Cost-Of-Churn
If you manage support operations or run an enterprise, you are measuring the wrong things if you treat customer service as a pure cost center.
The companies that build decades of brand equity treat customer support as a retention engine, not a drain on profits.
- Fix first-contact resolution metrics. If a bot requires three separate chat attempts before a customer gives up, your "resolution rate" is a lie.
- Remove the hidden barriers. Stop hiding human contact numbers behind endless self-help trees. If a customer wants to call you, let them call you.
- Reserve human talent for high-stakes interactions. Use basic automated tools for simple database lookups, but route high-friction, emotionally charged issues directly to experienced humans instantly.
- Calculate the true cost of customer loss. Compare the money saved on call center salaries against the revenue lost from customers who cancel their accounts out of sheer frustration.
Replacing experienced staff with scripted digital assistants during a customer service crisis is not innovation. It is lazy management executing short-sighted triage.
Build a service organization that actually solves problems, or watch your customers walk to the first competitor brave enough to answer the phone.
