Every custom manufacturer in India now owns at least two pieces of software that were supposed to fix production. An ERP that cost a few lakh and a small fortune in consultant time. And, more recently, some kind of dashboard — a screen of charts the sales rep promised would finally give you 'visibility.' Both are useful. Neither does the one thing you actually needed at eleven o'clock on a Monday morning.
Here is that thing. A customer has just changed an order. A machine has just gone down. Material that was promised for today is now stuck at the supplier. You do not need software to tell you these things happened — the WhatsApp group already did, faster than any system could. What you need is something to work out what they mean for the other twenty-nine jobs on the floor, and to do the tedious re-planning before the day falls apart. Your ERP cannot. Your dashboard will not. It will simply turn a tile red.
Three generations of factory software
It helps to see where the tools came from. The first generation was the ledger — the ERP. Its job is to record: Job 1142 started, finished, went to inspection, shipped, invoiced. It is a superb historian and a poor planner, because everything it knows is already in the past.
The second generation was the dashboard. It took the ledger's data and drew it as charts — orders at risk, machines idle, deliveries slipping. This felt like progress. For the first time you could see the whole factory on one screen. But a dashboard is a smoke alarm, not a fire brigade. It is very good at telling you that you are late. It has nothing to say about what to do next.
The third generation is the one most shops do not have yet, and it is the only one that touches the real problem. It does not just record the past or picture the present. It works out what to do — and does the boring part of it for you.
The dashboard trap
A lot of manufacturers spent the last two years buying visibility and are quietly disappointed. The screen is beautiful. The red tiles are honest. And nothing has actually changed, because seeing a problem and solving one are different jobs.
Worse, more visibility with no help can make the day feel harder, not easier. Now you can watch all thirty ways it is going wrong, in real time, and you still have exactly one planner to deal with them, one at a time, by hand. The dashboard raised the alarm and then handed you the same bucket you always had.
From reporting to reasoning
Picture the alternative. A drawing revision lands from your customer's engineering team — the bore size on Job 1136 has changed. It arrives the way these things really arrive: as a message with an attachment, not as a tidy ERP entry.
A system built for this catches that message where it lands. It understands what changed. It works out the consequence — two jobs queued to cut tomorrow use that drawing, and one of them shares a machine slot with a Friday dispatch. Then it proposes something concrete: hold the two affected jobs, pull Job 1142 forward into the open slot, and the Friday dispatch is still safe because it now finishes six hours early. The planner reads a plain-language summary and taps approve.
Notice what the human did and did not do. He did not hunt through email for the revision. He did not re-sequence thirty jobs on a spreadsheet to find out what the change touched. He did not discover at 4 pm that a promise was already broken. He made one judgment call — yes or no — on a fix that had been worked out for him, with the reasoning shown.
Acting is not the same as running unattended
This is the part every factory owner rightly worries about, so let us be plain. A system that acts does not mean a system that runs your floor on its own. It means the tedious eighty percent — catching the change, tracing what it affects, drafting the fix — is done for you, and the twenty percent that needs judgment stays with a person.
That balance is deliberate, and it is the difference between software that survives on a shop floor and software that gets switched off in a week. Three rules make it work. It proposes; you approve — nothing changes the live plan without a human tap. It always shows the reason — never a mystery re-sequence, always 'this, because that.' And every change is traceable — you can see who approved what and why, which matters the day a customer asks how a date moved.
Trust on a shop floor is earned, not assumed. The sensible path is to let a system act where the stakes are operational and reversible — sequencing, alerts, catching changes — and prove itself there before anyone extends it further. Start where it saves the most time and threatens the least, and widen the leash only as it earns it.
A Monday morning, replayed
Here is the same kind of disruption on a floor that has this layer. Nine forty a.m. A customer calls; a supervisor thumbs a message into the group: 'Job 1152, make it 40 pieces not 60, still Friday dispatch.' Within seconds the change is caught. The planner sees one line: 'Job 1152 reduced to 40 — Friday still safe, job now completes six hours early; the freed time covers the rush part that came in at 9.' He taps approve.
The office sees the new plan. The planning room sees it. The floor sees it. The operator on that machine knows before he starts cutting the second batch, not after. Nothing about how the supervisor works has changed — he still types into WhatsApp. But the message now moves the plan instead of dying in a chat nobody re-reads.
Why this shape, and why now, for India
The generic version of this story is being told at every trade show right now. The version that fits an Indian custom manufacturer has three non-negotiable features, and they come straight from how your floor actually runs.
It has to sit on top of the ERP you already have — Tally, SAP, Busy, or the system your cousin built — not replace it. Nobody is ripping out a working ledger, and nobody should have to. It has to listen where the truth actually lives, which is WhatsApp, email, phone notes and Excel, because that is where your floor reports reality and no memo is going to change that. And it has to reduce the load on the planner, not add to it — the one role you cannot easily hire for, and the one whose knowledge, as any owner who has lost a good planner knows, walks out of the gate the day he resigns.
The timing is not an accident either. The orders are arriving — the shift of global supply chains toward India is real, and it lands hardest on exactly the high-mix custom shops least equipped to absorb the coordination it demands. The shops that turn that demand into repeat business will be the ones whose planning kept pace. The rest will win the first order and lose the second on a missed date.
Reports, or decisions
That is the whole distinction, and it is worth keeping in a single sentence. Software that only reports leaves all of the work with you. Software that reasons and proposes takes the tedious majority of that work and leaves you the judgment — which is the part you are actually good at, and the part no machine has any business making alone.
- A change is caught the moment it lands, from the channel it lands in — no data entry, no new habit
- The consequence is worked out across every job it touches, not just the one in front of you
- A concrete fix is proposed, in plain words, with the reason attached
- A human approves in one tap — and everyone, office to floor, sees the same live plan at the same moment
The factories that pull ahead this decade will not be the ones with the most dashboards or the fattest ERP. They will be the ones whose software stopped merely watching the floor and started helping run it — with a person's hand always on the wheel. Recording the past is solved. Picturing the present is solved. Deciding what to do next, fast enough to matter, is the frontier. That is where the next advantage is.