Salesforce just made it dramatically easier to deploy an agent. It did not eliminate the harder work: designing the workflow that agent runs. Those are different problems. In most enterprises, deployment was not the real bottleneck.
The pitch reads like a modernization story. The clock reads like a governance reset.
The model was never the bottleneck. The implementation was. OpenAI just spent $4 billion confirming it.
It's 2009. You're in a requirements meeting for a new SharePoint deployment. The consultant across the table has one answer for everything: InfoPath. Employee onboarding? InfoPath. Expense approvals? InfoPath. Complex multi-step workflows with business logic that has no business living in a form layer? InfoPath. They're certified, they're confident, and they're wrong about where this is going.
Who owns the field name that just changed?
Not who configured it. Not who requested it. Who owns it — as in, who is accountable when that renamed field breaks a downstream calculation that feeds a compliance report that gets submitted to a regulator?
If the answer takes more than five seconds, the platform isn't the problem. The operating model is.
April 15 is Tax Day. Fitting, because software has its own version: a complexity tax that most teams pay without ever seeing the bill.
Every manual handoff between systems. Every field re-keyed from one screen into another. Every reconciliation spreadsheet that exists because two platforms don't agree on what happened. That's the tax. And most organizations pay it so routinely they've stopped noticing the line item.
The conference room has too many people in it. Two from procurement, one from legal, three from the program office, and across the table, the contractor's team — four deep, including someone whose only job today is to run the slide deck. The proposal is spiral-bound. It's thick. The methodology section alone runs eleven pages.
Nobody in the room will remember the methodology section in six months. But the contract will still be running.
Angular versus React versus Vue. SOA Suite versus REST. Windows Phone versus everyone. Every cycle produces true believers — and every cycle leaves them behind.
Certifications are a proxy. Back in the late '90s, when career-switchers were flooding into IT from every other industry, certs offered the fast path — a way to prove baseline competence short of going back for a CS degree. They exist because the buyer can't tell the difference between someone who can deliver and someone who can't — not until the system is in production and it's too late to switch vendors.
Everyone talks about the first time they used AI to generate code like it was a revelation. Like the clouds parted. Like they immediately saw the future. That's not what actually happened for most experienced developers. What actually happened was a knot in the stomach.