Skip to content
GullySales
← All articles

Notes for owners · Sales management

Pilot project or full order: how to close a first automation sale

A pilot lowers an automation buyer's risk only when success is defined first and the rollout is priced before it starts. When to offer a pilot, a phased order or a full order, and how to price each.

The GullySales team · Updated 17 Oct 2026 · 6 min read

Automation engineers inspect a guarded robotic production cell
On this page
  1. Pilot, phased order or full order
  2. Why a hardware pilot costs nearly as much as the first unit
  3. Let the rollout look cheaper than the pilot
  4. Where the pilot runs, and who runs it
  5. Imagine a die-casting unit in Waluj
  6. When to refuse a pilot
  7. Put the repeat price in your next pilot proposal

The plant head likes your robot cell and wants to try one on a single machine before ordering for all six. That can work. It works only when success is defined before the pilot starts and the price of the other five is already on paper. Offer a pilot when the plant has identical stations and the doubt is about the process. When the doubt is about you rather than the process, offer a phased order: one purchase order, released in stages, each opened by acceptance of the last. For a single machine, the pilot is the order.

Pilot, phased order or full order

Ask first: what would the pilot teach? If neither side can name it, it is a delay with a robot in it.

Offer a pilot when the plant has several identical machines and something about the process is genuinely unknown. Castings that vary from lot to lot, parts that tangle in the bin, a gripper new to this shape or a crew that has never worked beside a robot all qualify. The pilot should be able to change the design of units two onwards.

Offer a phased order when you have done this application before and the buyer's worry is your firm, your support or the plant's own team. A pilot teaches nothing here. A phased order gives the buyer the same exit, with a commitment behind it.

Ask for the full order for a single machine or line, for a standard application with references the buyer can call, and for a plant that has bought from you before.

Software pilots fail in their own ways, set out in why RPA pilots stall, and the pass mark for a trial is in machine vision and robotics demos. This post is about the commercial decision around a cell on a shop floor.

Why a hardware pilot costs nearly as much as the first unit

Buyers expect a small pilot to be cheap, but the first cell carries all the engineering: the layout, the programme, the gripper and fixture trials, the safety design. Units two to six reuse it.

For example, with illustrative figures: a first cell carries ₹10 lakh of design, programming and trials on top of ₹22 lakh of hardware and commissioning, so it costs ₹32 lakh. Each repeat cell needs only the hardware and its commissioning, say ₹22 lakh. The pilot is the expensive unit, whatever its size.

Let the rollout look cheaper than the pilot

The common mistake is to discount the pilot to get in. The engineering is then recovered on the repeat cells, so the rollout price per cell comes out higher than the pilot's. The buyer sees a price rise at the moment of decision.

Do it the other way round. Price the pilot at its real cost, with the lower repeat-cell price in the same document. Credit part of the pilot fee against the rollout if it is ordered by an agreed date. And write that the pilot cell stays where it is and becomes unit one.

The format of a one-page pilot letter, with the proof agreed and an "if you do not go ahead" line, is in selling IoT devices to businesses. It works for a robot cell with the repeat price added.

Where the pilot runs, and who runs it

Set these conditions in writing before installation:

  • The machine. Ask for an average one, not the newest press or the line the best operator runs. The owner will offer the showpiece.
  • The operators. The plant's own people run the cell on every shift, nights included.
  • Your engineer leaves before the measured period begins; a cell that runs only with your engineer beside it proves nothing.
  • The cell's position. Build it where it can stay, and say so. Moving it later is a priced change.
  • A changeover and a maintenance stop fall inside the measured period, not before it.
  • The measure comes from the plant's records: the production register and the downtime log, not your controller's screen.

Imagine a die-casting unit in Waluj

Imagine an integrator offering ladling and extraction robots to a pressure die-casting unit in Waluj, Chhatrapati Sambhajinagar. The unit has six identical machines making aluminium housings. The details are illustrative. The owner asked for a pilot. Keen to get in, the integrator priced it close to hardware cost and installed it on the newest machine, the one the owner shows visitors, with its engineer on site throughout. The pilot cell cost the integrator about ₹32 lakh to build and was sold for ₹24 lakh.

The cell ran well while the engineer was there. After the engineer left, the night shift could not clear a jammed extraction and went back to manual. Then the integrator quoted the other five cells at ₹26 lakh each, more than the pilot, because the engineering had to be recovered somewhere. The owner read it as a price rise, and the project stalled for two months.

The integrator went back with a phased order and one page showing what the engineering had cost. The proposal covered all six cells, released in phases. The pilot cell on the newest machine stayed where it was and became unit one. Phase one was cells two and three, on two of the older machines, with two operators trained from each shift. The measured period would start only after the integrator's engineer had left. The repeat-cell price was fixed at ₹22 lakh including commissioning, and phase two, cells four to six, would open on acceptance of phase one. The owner signed for all six, with the right to stop after any phase. The integrator's loss on the first cell would be recovered only if the plant went through every phase.

When to refuse a pilot

Say no, politely, in three situations. The buyer wants the pilot free and will not name who decides. The plant expects you to remove the cell if there is no rollout, which means lending it your capital. Success depends on something only the plant controls, such as casting quality, and the plant will not fix it. Offer a visit to a running site instead, as described in the demo, the visit and the reference call.

Put the repeat price in your next pilot proposal

Before you send the next pilot proposal, work out the real cost of the first cell and the price of each repeat, and put both on the same page. The steps before this one, from the problem page to the site visit, are on the industrial automation and robotics page, and making the pilot a stage with rules for entering and leaving is sales process design. If you want your next first sale planned with you, tell us what you need.

Questions

Questions owners ask.

Should a pilot be free if the buyer is a well-known plant?
No. A free pilot has no budget owner at the plant, so nobody is obliged to decide anything at the end. Charge for it and credit part of the fee against the rollout. A known name is worth asking for permission to describe the project later, not a free cell.
What if the pilot works but the plant wants changes before the rollout?
Expect it, and price it as a change, not a favour. Lessons from the pilot are the reason it was run. Write them into the rollout as a revised specification, so units two onwards are built to the better design and the pilot cell is brought up to match.
Can the pilot cell be moved to another machine later?
Only if the scope says so. A cell built around one machine's door, height and part chute needs rework to move. If the plant wants that freedom, design for it and price it. Otherwise the pilot cell stays where it is and becomes unit one of the rollout.
How long should a pilot run before the decision?
Long enough to cover every shift, at least one changeover and a stretch with none of your engineers on site. Set the decision date in the pilot terms before the cell is installed, not after it is running, or the pilot simply carries on with nobody deciding.

From the blog

More notes for owners.

Your next practical step

Tell us what you are trying to grow, and we will read it before we call.

Eight short steps, mostly taps. Skip anything you would rather answer on the call.