Most steel plant owners learn what went wrong last month. These seven numbers, seen every morning, tell you what is going wrong today.
It is 9 in the morning. The night shift has handed over, the first trucks are already at the weighbridge, and your phone has rung four times. Somewhere in the plant, the numbers that decide this month's margin are being written in a register you will see in three weeks.
Most steel plant owners we meet run on instinct and the month-end statement. Instinct is valuable. The statement is late. By the time it shows that power per tonne went up, the extra units are already on the electricity bill.
The seven numbers
These are the numbers that, seen every morning, change how a plant is run. Not a dashboard with fifty charts. Seven lines.
- Production by unit, yesterday. Sponge iron from each kiln, billets from the furnace, bars from the mill. Planned against actual.
- Coal and ore per tonne of sponge iron. The single biggest cost in a DRI plant, and the one that drifts quietly.
- Units of power per tonne of billet. If it rises two days in a row, something changed: the charge mix, the lining, or the way a shift runs the furnace.
- Billet-to-bar yield. Burning loss, mill scale and end cuts. Small percentages, large money.
- Stock that matters. Finished goods by size, and critical raw material in days, not tonnes.
- Dispatch and orders. What left the gate yesterday, and what is promised for today and tomorrow.
- Money. Collections received, dues past their terms, and the three dealers who owe the most.
Why you do not get them today
The numbers exist. They are scattered. The kiln operator writes them in a logbook. The lab writes the ore and coal analysis on a sheet. The weighbridge prints slips. Accounts enter bills a week later. Nobody is paid to bring them together every morning, so it happens at month end, when the accountant has to.
The problem is not people. It is the typing. Every one of these numbers has to be copied from paper into something, and copying is the first thing that slips when the plant is busy.
What good looks like
A good morning in a well-run plant looks like this: the owner reads one message, sees that power per tonne in furnace two was higher than usual on the night shift, and calls the shift in-charge before the day shift repeats it. The conversation takes three minutes. Without the number, it never happens.
You can start this without any software. Ask each unit head to send the same three numbers on WhatsApp at 8 am, in the same order, every day. Within a month you will know which numbers you actually look at, and which units struggle to send them on time. That is your map.
Where AI helps
The hard part is not the report. It is getting the readings in without anyone typing them twice. This is where AI has become practical for a steel plant:
- A supervisor speaks the shift log in Hindi at the end of the shift, and it becomes the entry.
- Weighbridge slips and lab reports are read from a phone photo, so every truck's weight and quality is recorded the moment it arrives.
- An agent compares each shift with its normal band and tells the shift in-charge the same day when power per tonne or yield drifts.
- At 9 am, a morning brief arrives with the seven numbers, and you can ask a follow-up in plain words: "Why was yield low on the mill yesterday?"
A person still decides. The AI only makes sure the decision is made on today's numbers instead of last month's.
Where to start
Do not start with all seven. Start with the one that costs you most, usually power per tonne or coal per tonne, and get it in front of you every morning. Once that habit holds, add the next.
If you would like to see how this would work in your plant, we have set out the features for steel plants, from sponge iron to coil, or you can simply tell us which number you would want first.