Robot Task Done Right: A Beginner Primer on Acceptance Criteria

Picture a familiar moment in workshops and warehouses. A new robot arm or delivery vehicle arrives, the integrators get it moving, and someone says, “Looks good!” A week later the operators are asking why it stops whenever a pallet is slightly crooked, and nobody can recall what “good” was supposed to mean. If you are new to robotics operations, this friendly primer introduces one habit that prevents that: writing acceptance criteria, a short statement of what the robot must do before you call the installation finished. A mini-glossary at the end covers the vocabulary.

What acceptance criteria are, in robot terms

Acceptance criteria are a list of observable conditions that must be true before a robotic task is accepted as complete. They answer the question, “How will we know this works well enough to hand over?” For a robot, that usually covers four things: what job it performs, how reliably, how safely, and what the people around it need to be able to do.

The word “observable” is important. A robot that “works well” is an opinion. A robot that completes a defined route without human help for an agreed number of consecutive cycles is a fact someone can watch happen.

Why robots especially need a finish line

Robots sit at the point where code meets floors, shelves and people, and that setting is endlessly untidy. Floors are uneven, lighting changes, boxes arrive in odd positions, and people walk through. Each surprise tempts the team to tune a little more. Tuning is valuable, but without an endpoint it never stops, and a half-tuned installation is awkward to operate and hard to maintain.

A finish line also helps the people who will work next to the machine. Operators want to know what the robot will and will not do. Maintenance staff want to know what a healthy machine looks like. Safety staff want to know which checks passed. A written statement of done serves all three.

The five parts of a robot finish line

  • The task: one sentence naming what the robot does, such as moving containers between two defined stations.
  • The performance: how many cycles, over what period, with what tolerance on position or timing.
  • The conditions: which situations it must handle, such as slightly misaligned loads, and which are out of scope.
  • The safety checks: which protective functions were tested and shown to work, with a record.
  • The handover: who can start, stop, reset and restock the machine, and what notes they have.

A worked example with a delivery vehicle

Imagine a small warehouse adding an automated guided vehicle to carry totes from a packing area to a dispatch bay. A vague plan says, “Get the vehicle running.” A finish-line version reads: the vehicle carries totes between the two named points along the agreed route, completes an agreed run of back-to-back trips without anyone touching it, stops and waits when a person steps into its path, resumes only after the path is clear, and can be restarted by a trained operator using the written procedure.

Notice how much that sentence settles. The route is named. The pass condition is a run of trips, not a feeling. The stop-and-wait behaviour is stated, and the human role is clear. When the vehicle meets those conditions, the installation is accepted. Further improvements, such as a second route, become a new task with its own sentence. For a broader introduction to how these vehicles fit a warehouse, the full walkthrough of automated vehicles in warehouse automation is a handy companion.

What about sensors?

Many robot problems trace back to what the machine can sense. If your finish line includes behaviour like stopping for obstacles, the criteria should say how that is verified: a test where an object or person-shaped marker is placed in the path and the stop is observed. For background on how sensing is built into industrial machines, see the article describing how industrial robots get their sensors built in. Understanding the sensing layer makes it easier to write conditions that can actually be tested.

Remember that safety assessments follow rules and standards that depend on your location and equipment, and they are carried out by qualified people. A friendly primer like this one can help you structure your thinking, though your site or equipment supplier will still require its own formal safety process.

Writing your own: a short routine

  1. Describe the task in one sentence, naming start point, end point and what is moved or done.
  2. Choose one measure of reliability, such as consecutive trouble-free cycles.
  3. List the awkward situations it must cope with and the ones you will exclude.
  4. Record which safety checks must be demonstrated, with the person responsible for each.
  5. Name who will operate the machine and confirm they have tried the handover notes.
  6. Add the closing line: when these are true, the task is accepted and new ideas become new tasks.

Common beginner mistakes

The first is skipping the awkward situations. Everything works on the demonstration day, when the loads are neat and the aisles are empty. A good finish line deliberately includes the untidy cases you know will occur.

The second is leaving the operators out. A robot that technicians love but that the daily staff cannot restart will be switched off within a month. Include a hands-on trial by the real operators in the criteria.

The third is forgetting the “out of scope” list. If a machine is not expected to handle damaged pallets, say so in writing, so nobody is surprised later and nobody feels the robot has failed.

How long should the observation run be?

Beginners often ask how many trouble-free cycles are enough. The right figure differs from site to site, and a number offered without context deserves doubt. A sensible approach is to base it on the real working pattern: enough cycles to cover a full shift’s variety, including a start-up, a break, a restart and the awkward loads you listed. Settle the number together with the staff who will work beside the machine, write it down, and resist the urge to adjust it after you have seen the results.

If the run fails partway, that is not a disaster. Note what caused the stop, fix it, and begin the run again. The sentence defines the target; the retries are just the journey toward it.

A glance outside robotics

The habit of ending a task cleanly is not limited to machines. A general blog such as WarOpsX, a site whose posts are organised around an operations theme, shows how the word “operations” stretches over many kinds of activity, and in every one of them a clear end makes the work easier to plan and judge.

Mini-glossary for beginners

  • Acceptance criteria: the observable conditions that must be true before work is accepted.
  • Cycle: one complete run of a repeated task, such as pick up, move and put down.
  • Commissioning: the stage where an installed system is tested and prepared for regular use.
  • Tolerance: the amount of variation allowed in a measurement, such as position or timing.
  • Gripper: the hand-like attachment a robot arm uses to hold things.
  • AGV: an automated guided vehicle that follows a set route to carry loads.
  • Protective stop: an automatic halt triggered by a safety function.
  • Out of scope: a situation the machine is deliberately not expected to handle.
  • Handover: the point at which responsibility passes to the people who will run the system day to day.

Keeping the criteria alive after handover

Once a robot is accepted, the sentence still has a job. It tells maintenance what normal looks like, and it gives a reference when something changes. If performance drops, the team can check the original conditions to see what has slipped. If the layout changes, they can see which criteria need to be re-tested. Store the document with the machine’s maintenance records so that it travels with the equipment.

Summing it up

Robots are exciting, which makes it easy to keep tinkering. Acceptance criteria give you a friendly way to say, “That is enough for this task.” Write the task, the measure, the conditions, the safety checks and the handover. Test them with the people who will use the machine. When they are true, celebrate and move on to the next job. Your future self, and everyone who works near that robot, will be grateful for the clarity.

Related Posts

Robotics Operations Skills: A Start-Stop-Continue Review

How robotics operations technicians and engineers can review a deployment period with start, stop and continue lists, and turn the findings into career skills.

Read More

Commission the Stubborn Robot Fault First: A Diagnosis

A diagnosis of why robotics schedules slip at the end and how starting each commissioning day with the stubborn fault moves discovery earlier.

Read More

Robotics Operations Today: The Rise of the Waiting-For List

A look at how multi-party robot deployments stall on unanswered requests, and why operations teams are adopting waiting-for lists with owners and dates.

Read More

Commissioning or Routine Care: Which Robot Jobs Fool You?

A friendly comparison of two kinds of robotics operations work, setting up new equipment and routine upkeep, and how far guessed hours stray from real ones in each.

Read More

Why Does the Robot Keep Stopping Here? A Five Whys FAQ

A cart halts at the same junction again and again. This friendly FAQ shows how five why questions turn a repeat robot stop into a fixable process gap.

Read More

The AGV That Kept Stopping at Aisle Seven: A Team Story

An invented warehouse team spends weeks on a robot that halts at a clear aisle, until narrating the whole journey aloud to a charging dock exposes the cause.

Read More

Leave a Reply