Buying a robot is a procurement decision. Keeping it useful for years is an organisational one, and the second determines the outcome far more than the first. Across deployments, the binding constraint is rarely capital and frequently capability.
The roles that must exist
Not necessarily four people, but four sets of responsibilities.
Operators. Running it daily, changing between tasks, clearing routine faults. The largest group and the one determining day-to-day availability.
Maintenance. Scheduled servicing, consumable replacement, diagnosing mechanical and electrical faults. Usually an experienced electrician or fitter with additional training.
Configuration. Editing content, adding variants, adjusting behaviour. The hardest to recruit and the one determining whether the deployment can adapt.
Ownership. Accountable for performance, tracking numbers, deciding priorities. Often a production or operations manager rather than a technical specialist.
The fourth is most often omitted and it has the largest effect on whether the deployment is still performing after three years. Equipment without a named owner degrades quietly, and nobody notices until the figures are already poor.
What to keep in-house and what to buy
A workable boundary for most organisations.
Must be in-house. Daily operation. Scheduled maintenance. Clearing common faults. Replacing consumables. Backing up and restoring configuration. Switching between existing programmes or content sets.
Without these six, every minor change requires an external call, costs accumulate, and the equipment gradually gets avoided.
Worth developing with several machines. Adjusting behaviour. Adding new variants. Diagnosing less common faults. Improving cycle time or content quality.
Sensibly outsourced. New system design. Complex offline programming. Safety design and validation. Controller upgrades. These are infrequent, so in-house skill decays between uses.
The principle. Keep what recurs; buy what is rare. A skill exercised once every few years will have faded before it is needed again.
Training that produces capability
Training is usually delivered once, to whoever was available, at the busiest possible moment.
Train during commissioning. People who help bring a system into service understand it in a way no classroom session produces. This is the cheapest and most effective training available and it is frequently declined for scheduling reasons.
Separate by role. Operators, maintenance and configuration need different content. A combined session serves none well.
Include deliberate faults. Break something intentionally and have trainees recover it. Recovery skill determines uptime, and it only forms through practice.
Train more people than seems necessary. Across shifts. Absence and turnover guarantee the trained person will sometimes be unavailable, usually at an inconvenient hour.
Record it. Video of the integrator explaining the system is valuable two years later and costs nothing to capture at the time.
Refresh after a few weeks. Once people have used it, their questions are far more specific and a short session then is worth more than extending the initial one.
Recruiting and keeping technical people
Experienced automation technicians are scarce, and that has practical consequences.
Realistic sources. Engineers from larger manufacturers with hands-on experience. People from integrators seeking a more stable environment. Final-year technical students who need development but stay longer.
Developing existing staff is often better than recruiting. A capable electrician or fitter already understands the plant, the products and the people. Teaching them about robots is easier than teaching a newcomer about the business.
Why people leave. No development path, no investment in training, and working alone with nobody to discuss problems with. Pay is one factor among several.
What retains them. Involvement in new projects, attendance at training and industry events, and not being the only person who knows how something works — which sounds counter-intuitive but the sole expert is usually the most pressured person and the most likely to leave.
The structural point. Capability built internally stays when a supplier relationship ends. Capability rented does not.
A path for an organisation starting from nothing
A sequence that works.
Before the first project. Identify one or two people who will own it, tell them clearly it is a long-term role, and involve them from the survey stage.
During the project. Send them to see reference installations, to the integrator's workshop during testing, and have them present during commissioning.
Immediately after handover. Have them perform all routine maintenance themselves with remote support, rather than having the supplier do it for them.
In the first year. Give them responsibility for performance figures and for improvement. This is the fastest route to real understanding.
Before the second project. Assess what they can now do independently and what still requires outside help, then narrow the outsourced portion with each subsequent project.
This is slower than buying a turnkey system and it produces something buying cannot: understanding that stays in the organisation. Over three or four projects the difference between organisations that did this and those that did not becomes very large.
Frequently asked questions
Which role is most often omitted?
Ownership — someone accountable for the equipment's performance, tracking the numbers and deciding priorities. It has the largest effect on whether a deployment still performs after three years.
What must an organisation be able to do in-house?
Daily operation, scheduled maintenance, clearing common faults, replacing consumables, backing up and restoring configuration, and switching between existing programmes. Without these, every minor change requires an external call.
Is it better to recruit or develop existing staff?
Usually to develop existing staff. A capable electrician or fitter already understands the plant, products and people, and teaching them about robots is easier than teaching a newcomer about the business.
What is the fastest route to real understanding?
Having the internal owner perform all routine maintenance themselves with remote support after handover, and giving them responsibility for performance figures and improvement during the first year.
More in Robots and jobs and Myths and questions.
