Buying Automation vs. Building It: When Custom Integration Pays and When It Doesn’t
1. What This Resource Covers & Why It Matters
Every operation making a first or second major automation purchase faces the same fork in the road: buy a preconfigured system designed for the application, or spec a custom cell with an integrator. Both paths lead to an automated process. However, the cost, timeline, risk profile, and long-term support implications are genuinely different. Choosing the wrong path for the wrong application is one of the most consistent and expensive mistakes in mid-market automation.
This article covers the practical comparison between buying preconfigured or turnkey automation and commissioning custom integration. It addresses cost ranges, timeline differences, risk allocation, and the conditions that make each approach the correct one. The audience is operations managers making a capital automation decision with a budget between $50,000 and $500,000 and limited prior experience with either path.
This article does not cover robotics-as-a-service financing models or lease arrangements. Those topics affect how you pay for automation, not which architecture you choose. The two questions are separate.
2. What’s Actually Happening: The Market Has Changed
Turnkey and Preconfigured Automation Has Matured
The turnkey automation market has expanded significantly in the past five years. Platforms like RBTX by igus now offer pre-tested, CE-certified robotic systems covering machine tending, inspection, palletizing, and pick-and-place applications. RBTX reports that 95% of its solutions cost under $15,000, with a complete six-axis arm and cell configuration starting around $24,000. OnRobot’s D:PLOY platform, Universal Robots’ UR+ ecosystem, and FANUC’s packaged machine tending cells all represent the same trend: a growing catalog of pre-engineered automation products targeting specific, well-understood applications.
These products exist because the same applications recur across thousands of shops. A shop tending a CNC lathe with a cobot, a packaging line moving boxes to a pallet, or an inspection station checking parts against a camera-based vision system all have enough in common that a pre-engineered solution can cover 80 to 90% of the use case without custom engineering. That remaining 10 to 20% is where the decision gets more complex.
Custom Integration Has Not Gone Away
Custom integration produces a cell built specifically for the application, the specific machine models, the specific part geometry, the specific safety requirements of the facility, and the specific production flow context. The integrator handles mechanical design, controls architecture, robot programming, HMI development, safety system validation, and commissioning. The output is not a product. It is an engineered system built once for one customer.
Custom integration remains the right answer when the application falls outside what preconfigured systems cover, when the process involves unusual part geometry, multi-robot coordination, complex vision requirements, specialized end-of-arm tooling, or process integration with existing PLC systems. The cost and timeline are higher. In exchange, the specification covers exactly what the operation needs rather than approximately what a catalog product covers.
3. Side-by-Side Comparison
| Decision Criterion | Turnkey / Preconfigured | Custom Integration |
|---|---|---|
| Application fit | Designed for common, well-understood applications | Engineered for specific application requirements |
| Capital cost | $15,000–$150,000 for most standard applications | $150,000–$800,000+ depending on scope and complexity |
| Lead time | 2–8 weeks from order to delivery | 12–36 weeks for engineering, fabrication, and commissioning |
| Timeline certainty | High — product is already engineered and tested | Moderate — scope changes and integration discoveries extend timelines |
| Programming | Pre-built; operator-configurable changeover | Custom programs built for the specific application |
| Flexibility after installation | Limited — designed for one application | Higher — programs and tooling can be modified by the integrator |
| Internal skills required to operate | Low — designed for operator-level use | Moderate to high — controls and programming knowledge needed |
| Support model | Vendor product support and warranty | Integrator support contract; knowledge held by integrator |
| Risk profile | Low engineering risk; risk is in application fit | Higher engineering risk; managed by integrator’s experience |
| Scalability | Replicate the same system — fast and cheap | Each additional cell may require re-engineering |
| Best fit | First automation purchase, standard application, limited internal skills | Complex or non-standard application, multi-process integration, significant existing PLC infrastructure |
4. When Each Approach Makes Sense
Turnkey: When the Application Is Standard and Speed Matters
A shop adding cobot machine tending to a single CNC lathe running standard cylindrical parts does not need custom integration. The application is well-defined, thousands of similar cells exist, and preconfigured systems from multiple vendors cover it fully. The correct move is to select a pre-engineered package that includes the robot, end-of-arm tooling, safety guarding, and application software, validate it against the specific machine model and part envelope in a vendor application test, and deploy it. RBTX’s feasibility check service and video consultation process specifically target this validation step before purchase.
Speed is the other argument. A custom integration project for a standard machine tending application might take 20 to 30 weeks from scope to commissioning. A preconfigured system can deliver and commission in four to eight weeks. For a shop with a labor gap driving the automation decision, that timeline difference has direct operational cost implications. Every week of delay is a week of the problem the automation was purchased to solve.
Turnkey: For Replicating Proven Cells
Once a preconfigured system has run successfully on one machine, replicating it across additional machines in the same facility requires minimal additional engineering. The Dutch integrator Just Automate demonstrated this model clearly with its MoveComponents palletizing product, built on OnRobot’s D:PLOY platform and deployed at WEMAS Absperrtechnik in Germany. The first installation validated the system. Subsequent installations replicated it without re-engineering. For operations scaling automation across multiple identical machines or workstations, this replication advantage compounds across every additional cell.
Custom Integration: When the Application Has No Catalog Equivalent
A shop running complex aerospace structural components that require five-axis fixturing, in-process gauging, multi-station material transfer, and a PLC interface to an existing SCADA system will not find a preconfigured product that covers the application. Custom integration is not optional in that scenario. It is the only viable path. The engineering cost is real, but it pays for a system that actually fits the process rather than a system that requires the process to adapt to what a catalog product can handle.
Custom Integration: When Existing Infrastructure Is Dense
Facilities with significant existing PLC infrastructure, safety system architecture, MES connections, or legacy machine interfaces almost always require custom integration. Preconfigured systems assume a clean integration environment. They connect to modern machine controllers through standard digital I/O or OPC-UA interfaces. When the target machine is a 15-year-old machining center with a proprietary control, or when the cell must interface with a facility safety controller that uses a specific SIL-rated protocol, the integration layer requires custom engineering regardless of how standard the robot application itself is.
5. Real-World Cost and ROI
A standard cobot machine tending package from a platform like Universal Robots or FANUC, purchased through a VAR or direct, runs $40,000 to $80,000 installed for a single machine. Installation labor costs are low because the system arrives largely configured. Commissioning time is two to five days for most standard applications. At that cost basis and timeline, payback on labor savings commonly falls in 12 to 18 months.
A custom-integrated machine tending cell for a more complex application, one that includes vision-guided part loading, multiple part program variants, an MES interface, and custom end-of-arm tooling, runs $150,000 to $350,000. Timeline from kickoff to production release runs 20 to 36 weeks. Integration and engineering labor typically adds $50,000 to $150,000 above hardware cost. Payback extends to 24 to 36 months but the system covers the actual application rather than an approximation of it. Integration and installation services in North America run $100 to $200 per hour, and project-level engineering fees commonly exceed the hardware cost on complex applications.
The cost comparison shifts further in turnkey’s favor when replication is planned. A $60,000 preconfigured system replicated across five machines costs $300,000. An equivalent custom integration project replicated five times could cost $600,000 to $800,000 because each installation carries a portion of engineering overhead even when the cell design is similar.
6. Integration Considerations
A preconfigured system arrives with an operator-facing interface designed for someone without deep automation expertise. Changeover between part variants happens through a menu-driven HMI, not through PLC programming. The vendor’s support model covers the product under warranty. For operations without internal automation engineering capability, this support model is a direct operational advantage. There is a single point of contact and a documented support escalation path.
Custom integration produces a system where the integrator holds the full engineering knowledge: the program structure, the PLC logic, the HMI design, and the wiring documentation. After commissioning, the operation depends on that integrator for any programming changes, fault diagnosis outside operator capability, and future scope modifications. This dependency is the most underappreciated cost of custom integration. Budget for a long-term support contract with the integrator from the start. Assume that the first 12 months will require integrator involvement for program modifications, process tuning, and faults outside the maintenance team’s competency.
Vendor documentation for preconfigured systems covers the product’s supported applications, machine interface specifications, and changeover procedure. It does not cover integration to non-standard machine controllers, facility-specific safety requirements, or process context that falls outside the designed application envelope. Discovering that the application falls outside the product’s actual capability during installation rather than during a pre-purchase application test is the most common and most expensive failure mode in turnkey automation.
7. Common Mistakes When Choosing
The most common mistake is choosing turnkey because the price looks right without validating that the application actually fits what the product covers. A preconfigured machine tending system designed for modern CNC machines with standard Ethernet interfaces does not integrate cleanly with a 2008 Mazak with a proprietary MAZATROL control. Discovering this at installation rather than during a pre-purchase feasibility check produces cost overruns that eliminate the price advantage that made the turnkey option appealing in the first place. Run the application test before purchasing. Both RBTX and Universal Robots’ application engineering teams offer this service specifically to prevent this outcome.
The second common mistake is choosing custom integration as the default, perceived-safer option for an application that is actually well-served by a preconfigured product. Operations managers with little automation experience sometimes associate higher cost and longer timelines with higher quality. In standard applications, that assumption does not hold. A preconfigured cobot tending system for a standard lathe is not a compromise. It is the right tool for the job, and spending $200,000 on a custom-integrated equivalent buys a more complex system that requires more internal expertise to sustain with no additional production benefit.
A third mistake is underestimating support dependency in custom integration. A custom cell is an engineered artifact. The integrator knows how it works. The operation does not. Without an extended support contract and a commitment to developing internal knowledge of the system, the operation becomes permanently dependent on the integrator for any modification or fault resolution. Build internal capability or budget for ongoing integrator support. Leaving both undefined at project close is a operational risk that compounds over time.
8. Key Questions Before Committing
- Has the proposed preconfigured system been validated against the specific machine model, part envelope, and material handling context through an application test, not just a product demonstration on reference parts?
- Does the custom integration scope include all integration points, specifically the machine controller interface, the safety system architecture, and any MES connection, and has each been priced as a defined line item rather than a contingency?
- For turnkey systems, what happens when a new part variant falls outside the product’s supported changeover range, and who handles the modification and at what cost?
- For custom integration, what is the support contract structure after commissioning, and what does the first 12 months of post-commissioning integrator involvement cost at the quoted rate?
- If the operation plans to replicate this cell across additional machines or workstations in the next 24 months, does the chosen approach support replication efficiently, and what is the estimated cost of each additional cell under that same approach?