An Enterprise Sal

A day-in-the-life vignette in the tradition of Mark Weiser''s "Sal" — the day-in-the-life he used to close his 1991 essay on ubiquitous computing. Our Sal is a Fortune Global 500 CIO in 2028. Nothing dramatic happens. That, we think, is the point.

Sal Reyes wakes up on a Wednesday morning in October. The house is warm without being warm; the thermostat has stopped needing her opinion about the temperature. Her phone shows two notifications on the lock screen — one from her son''s school confirming a permission slip her calendar had already pre-signed, one from her audit committee chair asking whether she is still available for the 4pm working session. Sal replies to the audit committee chair with two words, gets out of bed, and starts the coffee.

She has been Chief Information Officer of Alton Materials for six years. Alton is a Fortune Global 500 specialty chemicals manufacturer, sixteen plants globally, tenth-largest producer of engineered polymers in North America. She reports to the CEO. She sits on three committees. She was, five years ago, the second person in the company to say the word autonomous out loud in a leadership meeting. Nobody had asked her to say it. She said it because she had read a scenario report by a publication she still subscribes to.

At the kitchen counter she opens the tablet that lives on the charger next to the coffee machine. She does not use it often. It is there for a specific pattern of moment — three cups of coffee before the day''s first meeting, twenty minutes of quiet, no laptop. The tablet is showing the platform-health card her team publishes to any device she authorizes. There are two indicators on it that she looks at every morning. The first is the count of autonomous decisions Alton''s workflows produced in the previous twenty-four hours; that number is 47,821, a number that stopped meaning anything to her about a year ago the way a mile count on the odometer stops meaning anything after a while. The second indicator is the count of escalations that were opened in the same window and are still open. That number is 4. Sal reads the four titles.

Three of them are workflows escalating to their business owners: a supplier variance workflow at the Freeport plant, a cash-application workflow in the North American controllership, a customer-order status workflow flagged by a strategic-account team. All three are already assigned. All three have owners who will handle them today. Sal is not the owner of any of them. She notes them for context and moves on.

The fourth escalation is the one she needs to touch. A workflow at the Kingsport plant flagged an atypical yield deviation last night. The escalation criterion was met — the yield was 2.1 standard deviations below trailing average — but the workflow could not identify the cause, which is the specific condition that routes to Sal''s platform team rather than to plant operations. Not because Sal''s team knows anything about yield. Because when a workflow can''t explain itself, the correct response is to inspect the workflow, not to inspect the plant. Sal reads the workflow''s own diagnostic notes: the reasoning core had processed the batch record correctly; the validator had confirmed the deviation; the router had pulled the feedstock file; the drafter had proposed three plausible causes; the escalator had, correctly, decided none of them cleared the confidence threshold. The workflow had done what it was supposed to do. Something in the world was new.

Sal sends a Slack message to her chief architect. "KP yield workflow at 4:07am. Handled correctly. Investigate causes on plant side. Update by 2pm." Her chief architect responds with a thumbs-up before her second cup of coffee is finished.

She showers, dresses, walks the dog. On the walk she thinks about the Kingsport escalation not because it is unresolved but because she likes thinking about it. Six years ago she was drafting quarterly board updates about AI Center of Excellence adoption metrics. Now she has a workflow that runs at 4am, decides it is confused, and pauses. That difference is the difference between the previous decade and this one.

At 8:15am she is in her car. The car checks her calendar and offers to route through the office parking garage that is closest to her first meeting. She agrees. The car informs her that the same route will bring her past a coffee shop where her assistant has, based on Sal''s pattern for Wednesdays, already ordered a second coffee. Sal accepts. Somewhere between the third stoplight and the freeway on-ramp Sal is asked no questions and given no notifications; the car is aware of everything it needs to be aware of and has decided none of it requires her attention. This is what her home tablet feels like too, and her office phone, and her calendar. She has stopped thinking of them as different systems. They are, in some sense, the same system, spread across her life.

She arrives at the office at 8:47am. Her platform team''s morning stand-up is at 9:00. She walks into the platform bay — six desks, four engineers, two workflow managers, a whiteboard with the current sprint on it — and reads the whiteboard while they finish their coffee. The whiteboard is boring. That is Sal''s read of the platform team''s job: making the whiteboard boring. Two items are in progress: a governance-fabric upgrade for the European operating region, and an integration to a new supplier''s ordering system. The Kingsport workflow escalation has already been assigned to Ravi, who is doing a walkthrough of the workflow''s diagnostic trace on his screen. Nobody is panicking. Nobody has any reason to.

Sal stands through the stand-up. Twelve minutes. Ravi reports the Kingsport workflow will be reviewed and re-shipped by end of week if the cause turns out to be plant-side; if it is workflow-side, he flags the postmortem, which will run the following week. The two workflow managers, Priya and Elena, review the four escalations from the overnight window; three of them belong to business functions and are already being handled. Priya notes that the Freeport controller has requested a modification to her cash-application workflow and is planning to make it herself; Sal asks whether Priya needs to be in the loop, and Priya says no. The controller has been building her own workflows for eighteen months. The change is a routine rerouting of an escalation criterion. Sal thanks her and leaves the platform bay by 9:14am.

At 10am Sal is in a room with three other C-suite executives, prepping for the following week''s audit committee. The topic is the annual attestation on autonomous workflows — the framework she and her chief architect wrote in early 2027, which the audit committee now expects to see updated quarterly. The chair likes the framework. The general counsel likes the framework. The CFO, six months ago, asked whether they could shorten it. Sal had said no. The general counsel had said yes on her behalf. The framework got shorter. Nobody was worse off.

She spends the meeting listening more than she talks. The chair asks about the Kingsport escalation. Sal explains it in one paragraph. The chair asks whether that pattern — a workflow flagging itself as uncertain — happens often. Sal says: forty to sixty times a month across the sixty-eight workflows they run. The chair asks: and every single one gets a human review? Sal says: every single one. The chair thanks her and moves to the next topic.

At noon Sal has lunch with the chief operating officer of the North American plants. They eat in the executive dining room; she has the soup. He tells her that his plant managers are pushing for autonomy on maintenance scheduling — a workflow they do not currently run — and asks whether Sal''s platform can support it. Sal says yes. He asks how long. Sal says her platform team is not the constraint; the plant maintenance leaders would need to author the workflow with the platform team advising, and the platform team''s advising bandwidth this quarter is a few weeks. He nods and asks who on his side would need to be the sponsor. Sal names three people. He writes them down.

After lunch Sal has a review of the third-quarter platform-health report she will present to the board in November. There are twelve slides. She does not care about most of them. She cares about slide four, which shows the yield-per-employee delta between Alton and its two nearest competitors over the trailing eighteen months. Alton is 1.6 times either of them. The number is not surprising because it has been slowly assembling itself for eighteen months. Slide four is the point of the board update. Everything else is throat-clearing.

At 4pm the audit committee working session runs its normal length. At 5:30 Sal has thirty minutes with her chief architect to walk through the yearly platform roadmap. He shows her a rough draft. She points at three items. He nods and rewrites in front of her. At 6pm she closes her laptop.

At 6:04pm the four escalations from the morning have been reduced to two. The Kingsport workflow investigation is moving. The customer-order status escalation has resolved. Sal knows this because when she puts her phone in her bag it briefly buzzes with a summary. She does not look at it. She trusts the summary. She would look if the number were higher.

She drives home. The house is warm again without being warm. Her spouse is cooking. There is a story from a colleague on her phone she had been meaning to read. She sits down and reads it. It is a long piece in a business publication about how her industry''s operating margins are, quietly, diverging. The piece names six companies. Alton is one of them. The piece does not name Sal. She has not spoken publicly about any of this in over a year. She has an offer to appear on a national business program next month and she has already declined it. She does not feel the need to explain her business to anyone who did not ask.

At 10:15pm she closes her tablet. Twelve autonomous workflows enter their overnight cycles across Alton''s sixteen plants. Two of the workflows will flag themselves as uncertain before Sal wakes up. Both will be routed correctly. Both will be waiting for her in the morning card, and she will look at them, and she will act on them the way she acted on this morning''s, which is quietly and without ceremony.

None of what happened today, from the outside, looks like a technology story. That, Weiser would have said, is the sign that the technology has arrived.

This vignette is a companion piece to The Autonomous Enterprise Is Weiser''s Third Wave and to Enterprise Autonomy 2028. Read the scenario report at enterpriseautonomy.ai/report. More editorial at enterpriseautonomy.ai.