Capital Investment Planning Support in Thailand 2026: Approval, BOI and Acceptance

Capital investment planning support is not the task of arranging supplier quotations in a spreadsheet. For a Thailand factory, it should establish the loss baseline, define the investment boundary, connect safety and measurement requirements to acceptance evidence, test BOI assumptions, and verify benefits after start-up. This guide sets out a practical 90-day pre-investment plan and an evidence chain from the RFP through FAT, SAT and benefit realization.

Capital investment planning support: build evidence gates, not an approval pack

An investment can fail even when the equipment runs. Automation may move the bottleneck to another process. Added safeguards may reduce the planned capacity. The inspection result may not be reproducible. A BOI incentive assumed in the business case may not apply. A machine may pass FAT but remain unstable with production material in Thailand. The project may close before anyone checks whether the financial benefit was realized.

A disciplined plan uses seven decision gates.

A gate is not another meeting. It asks whether the evidence is sufficient for a defined decision. Every open assumption needs an owner, due date and impact on cost, schedule, safety or performance. The word “approved” must not erase exceptions.

Start the automation investment plan with factory losses

Starting with a preferred robot or machine fixes the solution before the problem is understood. Begin with where a loss occurs, under which operating conditions, how large it is and how confidently it can be measured.

Separate the loss mechanisms

Distinguish unplanned stops, speed loss, defects, changeover, rework, material waiting, labor constraints, safety restrictions and inspection delays. Stratify them by product family, shift, machine, material lot and upstream or downstream condition. A monthly average can hide a chronic night-shift recovery problem or a loss that appears only on the most profitable product mix.

“Automate because labor is scarce” is not yet a requirement. Observe how many minutes are lost when a role is absent, how cycle time and defects vary with skill level, and how often secondary work interrupts the main task. The purpose is to understand process variation, not to rate individual workers.

Give every baseline a measurement definition

Availability, downtime, defect rate and labor hours can mean different things to production and finance. State whether planned stops count, whether material waiting is a machine stop, whether a re-inspected part is defective, and where relief labor is booked. Record the source, clock, missing-data treatment and approving owner.

ISO 10012:2026 was published in February 2026 and replaces the 2003 edition. It addresses requirements for measurement-management systems and confidence in measurement results. In an equipment project, that perspective extends beyond calibration stickers: the measurement process, personnel, environment, software and records all influence validity. If the pre-investment and post-investment methods differ, the claimed improvement cannot be compared cleanly.

Confirm the system constraint

Speeding up one station does not increase saleable output when material supply, downstream inspection or another operation remains the constraint. Observe work in process, queue time, blocking, starvation and quality-decision lead time across the flow. Also distinguish the purpose: more saleable output, lower overtime, better delivery reliability, reduced risk or redeployment of people. The same cycle-time improvement produces different cash effects.

Define a comparable scope for the automation investment plan

Compare more than one response against the same outcome. Options may include standard-work improvement, fixture changes, preventive maintenance, layout changes, level loading, outsourcing, retrofit, semi-automation, a dedicated machine or a robot cell.

High-mix plants must include changeover, first-piece approval, recipe control, exception handling, cleaning and maintenance—not only maximum speed. See our high-mix, variable-volume automation guide for Thailand for the related design method.

Put the boundary on one page

Show the workpiece hand-off points, good/reject/exception routes, existing-machine signals, utilities, building works, network, higher-level systems, operator paths and maintenance access. Assign each interface to the customer, machine builder, system integrator or third party. Mark unresolved responsibilities openly.

Without a common boundary, one quotation may include conveyors, guarding and data integration while another excludes them. The totals are not comparable. Boundary definition is not administrative detail; it is the basis of price, schedule and performance comparison.

Provide the operating range, not one golden sample

Describe dimensions, mass, center of gravity, surface, reflectivity, stiffness, tolerances, temperature, contamination, presentation and material-lot variation. Supply controlled samples at the acceptable limits and plausible abnormal samples. Classify future products as included in the design, accommodated by reserved flexibility, or handled through a future change order.

Do not make zero labor the objective

Automating every jam recovery, replenishment, sampling task, cleaning step and tool change can sharply increase cost and technical risk. Separate normal flow from exception flow and decide which exception frequency justifies automation. When people remain in the process, specify safe access, decision information, training and ergonomics as requirements.

Build a capital expenditure approval case with TCO and benefits

A simple equipment-price-divided-by-labor-saving calculation omits the cost of becoming operational and staying operational. Build total cost of ownership across the selected evaluation period.

TCO cost elements

• Concept and detailed engineering; mechanical, electrical, controls, robot, vision, safety and inspection equipment
• Building, foundation, power, air, network, fire and environmental work
• FAT, packing, transport, import, installation, SAT, training and ramp-up support
• Spares, consumables, calibration, maintenance agreements, software, licensing and cybersecurity
• Energy, utilities, planned service, modification and obsolescence response
• Production samples, trials, parallel production, old-equipment outage and initial yield loss
• Eventual removal, relocation or renewal, less defensible residual value

Use a transparent formula:

TCO = initial investment + ramp-up cost + operating and maintenance cost + change cost + expected downtime loss − residual value − confirmed incentive effect

Do not deduct a hoped-for BOI benefit in the base case. Use a no-incentive base and add a separate scenario only when eligibility, timing and evidence have been checked.

Prevent double counting

Labor benefit differs among actual headcount reduction, avoided recruitment and redeployment. Capacity benefit is realized only when demand and the system constraint support additional sales. Check whether defect reduction and material saving describe the same effect, or whether availability and overtime reductions overlap.

One useful formulation is:

annual net benefit = verifiable labor/overtime effect + margin on saleable added output + defect/rework reduction + maintenance/energy reduction − added operating and maintenance cost

An illustrative calculation, not a market quotation

The following assumptions demonstrate the method only. They are not market prices, wage data, an incentive forecast or a guaranteed result.

• Initial investment: THB 12.0 million
• Ramp-up, training and initial spares: THB 1.5 million
• Verifiable annual gross benefit: THB 4.0 million
• Additional annual operating and maintenance cost: THB 0.8 million
• Annual net benefit: 4.0 − 0.8 = THB 3.2 million
• Simple payback: (12.0 + 1.5) ÷ 3.2 = approximately 4.22 years

Test lower benefits, delayed ramp-up, higher maintenance and a capacity gain that does not become sales. Calculate NPV and IRR using the company’s own rules for discount rate, tax, depreciation, currency and funding.

Make page one sufficient for an executive decision

Page one should state the business problem, purpose, boundary, required outcome, investment range, base and downside cases, major risks, decision requested and next gate. Technical detail belongs in appendices. Executives need to know what outcome is being purchased, which uncertainty is accepted and when the company can revise or stop—not every component feature.

Treat BOI automation incentives as a verified scenario

Thailand BOI/OSOS reported that, in the first half of 2026, the Smart and Sustainable Industry initiative received 132 applications valued at about THB 17.2 billion (the Thai release gives THB 17,158 million) for machinery upgrading, digital technology, automation and robotics. These are application statistics. They do not demonstrate that a reader’s project will be approved or receive the same treatment.

Check the current Thai page and the project facts

The current Thai-language BOI automation page publishes an automotive-industry upgrading measure for applications from the first working day of 2026 through the last working day of 2027. It describes a three-year corporate-income-tax exemption capped at 50% of qualifying automation and robotics investment; the cap may rise to 100% where machinery supporting the domestic automation-machinery industry meets the published 30% threshold. It excludes land and working capital and also lists machinery import-duty exemption.

However, the English page still displays older material with a 2025 deadline. Confirm the eligible activity, qualifying expenditure, domestic-link calculation, application timing, order and start-up constraints, forms and post-approval reporting against current BOI announcements and with BOI or qualified advisers for the specific project. This article is not a tax, legal or investment-promotion determination.

Do not wait until after purchase ordering

Add fields for potential eligibility, evidence required, owner, enquiry date, source of response, filing deadline, ordering constraint and approval assumption. Identify early whether the equipment BOM, supplier origin, local-industry linkage, contract and payment breakdown may be needed. Keep the approved application scope synchronized with equipment revisions.

Use three scenarios

Show A: no incentive; B: a conservative case under verification; and C: the case after qualifying conditions are confirmed. If only C makes the investment viable, state the risk of ordering before approval and define a stop, phase or reduction option.

Put safety into capital planning, not the contingency budget

Safety affects layout, speed, access, capacity, maintenance, training and acceptance. It is a design condition, not an accessory added after approval.

ISO 12100:2010 remains the published general reference for machinery risk assessment and risk reduction. ISO identifies a successor draft as under development; the draft must not be represented as a published replacement. ISO 13849-1:2023 provides a design methodology for safety-related parts of control systems. Determine the applicable standards, Thai legal requirements, customer rules and export-market requirements for each project.

U.S. OSHA 1910.212 is a useful reference for general machine guarding against point-of-operation, nip-point, rotating-part and flying-material hazards. It is a United States regulation, not Thai law. Thailand compliance must be determined with competent local specialists.

Assess work scenarios, not only hazards

Cover normal production, replenishment, changeover, cleaning, jam clearing, manual operation, teaching, tool change, maintenance, power restoration and abnormal-product retrieval. Record who enters where, residual energies, reasonably foreseeable misuse, protective measure and validation method.

The RFP should request more than “complies with safety standards.” Require the risk-assessment assumptions, residual risks, safety-function list, required performance, guarding layout, mode transitions, energy-isolation concept, validation plan and handover to signs, procedures and training. A change to a safety function must update design, software and FAT/SAT evidence.

Convert quality and measurement into accept/reject rules

“To drawing,” “good accuracy” and “zero defect” are not testable acceptance requirements. Define the characteristic, datum, method, equipment, environment, sample, decision rule, record and responsible approver together.

ISO 5459:2024 addresses datums and datum systems. ISO 1101:2017 addresses geometrical tolerances of form, orientation, location and run-out. A drawing requirement and the physical or mathematical method used to establish and measure it must be consistent, or the supplier and customer can obtain different results.

ISO 14253-1:2017 addresses conformity and nonconformity decisions, including values close to specification limits while taking measurement uncertainty into account. Agree the decision rule in the contract. For critical characteristics, evaluate the measurement system, fixture repeatability, environment, operator effect and data processing.

Design the inspection process as part of the investment

A faster machine merely creates a queue when measurement cannot keep pace. Decide between full and sample inspection, inline and offline checks, measurement time, masters, calibration, abnormal segregation, re-measurement authority, retention and upstream-system interfaces. When inspection feedback automatically adjusts the machine, assess the risk that an incorrect measurement could create a run of defects.

A practical 90-day pre-investment plan

This model does not promise a purchase order in exactly 90 days. It is a way to reduce pre-approval and pre-RFP uncertainty in controlled stages. Extend it for major civil works, regulatory review or long-duration trials.

Days 1–15: lock the loss and measurement baseline

• Name the sponsor and owners from production, engineering, quality, maintenance, EHS, IT/OT, procurement and finance.
• Observe the target and adjacent processes; measure stops, defects, labor and changeover.
• Record definitions, source, missing data, uncertainty and approval.
• Confirm the system constraint and demand basis.
• List options including “do nothing,” with early stop criteria.

Outputs: problem statement, loss baseline, measurement plan and stakeholder map.

Days 16–30: compare concepts and boundaries

• Compare at least three responses against the same outcome.
• Map workpiece variation, exceptions, changeover, maintenance, data and safety boundaries.
• Evaluate rough capacity, layout, utilities, building, IT/OT and staffing impact.
• Conduct a preliminary BOI eligibility and sequencing check.
• Open a technical, safety, schedule and organizational risk register.

Outputs: concept scorecard, boundary drawing, rough layout and risk register.

Days 31–45: write the URS and acceptance method together

• Give every requirement an ID and a verification method: inspection, analysis, demonstration or document review.
• Make capacity, quality, safety, availability, changeover, recovery, data, training and service measurable.
• Draft FAT/SAT locations, materials, samples, decision owners and retest rules.
• Connect measuring equipment, datums and decision rules to quality characteristics.
• Assign customer supply, integrator supply, third-party work and open interfaces.

Outputs: URS revision one, traceability matrix and FAT/SAT strategy.

Days 46–60: create the business case

• Calculate TCO, annual net benefit, payback, and NPV/IRR under company rules.
• Model demand, ramp-up, effect, service, currency and BOI sensitivities.
• Separate cash from non-cash benefits and avoided hiring from actual reduction.
• Show no-investment risk and options to phase, reduce or stop.
• Ask for a defined budget, authority, deadline and next gate.

Outputs: approval paper, financial model, sensitivities and risk response.

Days 61–75: issue a comparable RFP

• Give candidates the same URS, boundary, workpiece data and scoring rules.
• Share clarifications fairly.
• Standardize cost breakdown, exclusions, assumptions, schedule, payment, warranty, service and change rates.
• Score technical approach, project control, safety, service and TCO.
• Demonstrate limit samples and abnormal recovery, not only normal running.

For robot projects, our Thailand robot system integrator selection guide explains the related supplier evidence in detail.

Days 76–90: freeze contract gates and execution criteria

• Baseline requirements, boundary, price, schedule and assumptions together.
• Set design-review and approval points for long-lead items, software, cyber, safety and quality.
• Define a change request, impact analysis and rules for work before approval.
• Link FAT, shipment, SAT, production handover and final payment to evidence.
• Schedule benefit checks at suitable points such as 30, 60 and 90 days after stable start-up.

Outputs: selection record, contract attachments, master schedule and benefit-verification plan.

Link RFP, FAT, SAT and benefits in one evidence matrix

An RFP requirement must remain traceable through design and operation.

FAT is a shipment gate

Check drawings, BOM and software versions; normal and endurance performance; changeover; limit workpieces; sensor, communications and power-recovery faults; safety functions; reject segregation; backup restoration; documents and training. Agree start and stop points, warm-up, sample count, exclusions, statistic and measuring equipment in advance. A momentary fastest cycle is not evidence of sustainable capacity.

Classify open items by severity and connect them to shipment permission, corrective date, retest and payment hold. A meeting note that says “fix later” may become expensive once access, parts and people are limited at the site.

SAT is the site-integration gate

SAT includes the Thailand site’s floor, utilities, network, temperature, upstream and downstream equipment, production material, operators, maintenance and shifts. Separate an unmet customer prerequisite from an equipment nonconformity, while assigning both an owner and deadline.

Production handover can require stable operation over an agreed period, named product coverage, quality evidence, punch-list status, completed training, spares, backups, drawings, code, licences and service contacts. SAT acceptance and final benefit realization are separate gates.

Benefit verification joins operations and finance

At agreed checkpoints, use the original metric definitions. More output is not realized benefit without demand. Fewer operators produce different cash effects depending on overtime, agency labor, hiring and redeployment. Reconcile quality gains with material use, rework, defects escaping to customers and customer-complaint records.

When the target is missed, distinguish equipment defect, unmet prerequisite, operating learning, demand change and measurement difference. Set a deadline to choose tuning, added investment, operating change, scope reduction or contractual remedy.

Governance for capital investment planning support

Capital projects are not the sole responsibility of production engineering. Include an executive sponsor, operations, engineering, quality, maintenance, EHS, IT/OT, procurement, finance and the person responsible for BOI or tax coordination. Do not make everyone approve everything; define who prepares, reviews, approves and receives information for each deliverable.

Use weekly reviews for decisions rather than status narration. Distribute the question, required evidence, options, recommendation and impact of delay before the meeting. An issue log should contain the decision needed, owner, deadline, impact and next evidence—not only a symptom.

Set concept, intermediate and final design reviews appropriate to the project. Review mechanical, electrical, control, safety, quality, maintenance and data together. After approval, evaluate every change for its effects on risk assessment, drawings, software, testing and training as well as cost and schedule.

Common failure modes

Treating data integration as a task for after equipment selection

Equipment-data design cannot be postponed if the pre-investment baseline, FAT/SAT results and post-production benefits are to remain in one traceable flow. During requirements definition, decide at what granularity machine state, product, variant, lot, alarm, quality result, operator action, recipe and time will be linked. Specify not only tag names but also units, sampling interval, state definitions, time zone, retention, behavior when data is missing and the data owner.

A single stop signal, for example, cannot distinguish a breakdown from material waiting, quality waiting or a safety stop, so it is weak evidence for explaining investment results. Conversely, collecting more high-frequency data than the decision requires increases network, storage, analytics and security costs. Work backward from the business questions and collect the evidence needed to answer them.

The IT/OT responsibility boundary should cover the machine PLC, edge device, factory network, server, cloud, MES or ERP, user administration and remote maintenance. Decide whether a communications loss stops the machine or allows local buffering and later transmission, and how timestamps that drift will be handled. Verify not only that a backup can be created but also that it can be restored from separate media during FAT or SAT.

Remote access can improve serviceability, but permanent connections, shared accounts and supplier privileges without expiry create risk. Requirements should address connection approval, multifactor authentication, least privilege, activity logs, permitted time windows, emergency disconnection and revocation when personnel change. Equipment cybersecurity must be decided together with safety, availability and maintenance design, not delegated to information technology in isolation.

Reducing training to an operator briefing

Handover requires different competencies for operators, team leaders, maintenance technicians, quality personnel, process engineers and IT/OT staff. Define training by role and cover normal operation, product changeover, recognition of abnormalities, safe stopping, the permitted scope of recovery, escalation, backup, inspection and records. Confirm competence by practical demonstration or understanding checks, not only attendance.

Training material must use language and expressions understood on the factory floor and match screen names, alarm numbers, drawings and part numbers. Even an accurate translation is unusable for recovery if the machine display is available only in another language, the procedure is obsolete, or the training video shows a different software revision. Make the final document list, revision, storage location and update owner part of the handover criteria.

Experience will diverge by shift after initial training, and personnel, products and software will change. Develop in-house trainers and include periodic refreshers and change-triggered training in operations. When only a few experts know how to recover from abnormalities, they become a single point of failure for the investment benefit.

Tie contract and payment terms to evidence gates

An equipment contract can define milestones such as advance payment, design approval, procurement of major components, FAT, shipment, SAT, production handover and warranty completion. Payment should be tied to the deliverable and acceptance evidence at each milestone, not merely to calendar dates. No universal payment percentage or warranty term is appropriate; these depend on project risk, commercial practice and legal review.

Separate acceptance from the warranty start date

Shipment approval, transfer of title, transfer of risk, site acceptance, production handover and warranty commencement do not necessarily occur on the same date. Damage in transit, delayed customer construction, long storage after FAT or partial operation can make an ambiguous start date contentious. Map the contractual terms to project gates and define which evidence establishes each date.

For warranty scope, distinguish covered defects, consumables, customer modifications, misuse, material conditions, remote diagnosis, site visits, travel, holiday response, replacement parts and software corrections. “Free repair” alone does not define downtime exposure or a recovery deadline. Separate acknowledgement, start of diagnosis, site arrival, temporary recovery and permanent corrective action, and confirm the parts and access rights required to achieve them.

Design change management before requesting quotations

At contract baseline, freeze the requirements, responsibility boundary, cost breakdown and schedule at the same revision, then issue a numbered change request for deviations. One record should track the reason, requester, affected documents, technical, safety, quality and data impacts, cost, schedule, retest, training and spares. If urgent work is authorized verbally, define a deadline for completing the record.

Defect correction, different interpretations of a requirement, customer-requested change, changed site conditions, and legal or standards compliance can carry different cost responsibilities. Define the classification, decision authority, deadline for impact assessment, customer approval period and limited exceptions for starting work before approval. The goal is not to prevent change; it is to preserve alignment between the investment purpose and acceptance evidence while change is controlled.

Copying a supplier proposal into the approval paper

A supplier proposal is optimized around its own boundary and assumptions. The customer’s loss baseline, adjacent processes, financial rules, BOI facts and maintenance capability still need independent evidence. Normalize every proposal to the same URS and boundary.

Treating the lowest quotation as the smallest investment

Adding excluded civil work, safety, inspection, system integration, training, spares and production ramp-up can change the ranking. Compare TCO for the same outcome, not the first invoice.

Designing acceptance after approval

If no verification method exists, the finished design may be impossible to test. Assign FAT/SAT evidence while the URS is being written.

Assuming BOI approval in the funding plan

A published measure does not establish project eligibility. Keep a no-incentive case and confirm the filing, ordering and implementation sequence before committing.

Closing the project on the first production day

Temporary expert support can make early results look strong. Define stable-operation criteria and a benefit-review date; have finance verify realized effects.

FAQ about capital investment planning support

When should a factory request planning support?

Before the equipment specification is fixed—ideally while the loss baseline and concepts are being compared. Support after quotation can still normalize RFP and acceptance evidence, but the room to change the solution and boundary is smaller.

How many quotations are required for an automation investment plan?

The number alone does not create fairness. Candidates must answer the same URS, boundary, operating conditions and acceptance rules. For a genuinely single-source technology, strengthen the decision with open cost breakdown, independent review and comparison with a non-purchase alternative.

What is the most important number in capital expenditure approval?

Not one payback figure. The important chain is baseline, TCO, annual net benefit, downside case and the condition under which a benefit becomes cash. Naming the metric owner and method also enables post-start verification.

Can every factory use a BOI automation investment incentive?

No. Measures have eligible activities, investment content, periods, qualifying costs and implementation conditions. Confirm the current announcement, forms and the particular project with BOI and qualified tax or legal advisers before ordering on that assumption.

Does FAT approval justify full payment?

Not automatically. Some evidence can only be produced after site integration. Align payment milestones with FAT, shipment, SAT, production handover, document completion and punch-list closure, subject to project risk and legal review.

Can ROI be guaranteed?

Market demand, operations, material, maintenance and exchange rates change. Separate contractual equipment performance from investment benefits that also depend on customer conditions. Manage the latter with sensitivity analysis and post-start verification.

Summary: make the investment verifiable before ordering

The value of capital investment planning support is not a polished approval document. It is an evidence system that joins the problem, concept, safety, quality and measurement, BOI scenario, TCO, RFP, FAT/SAT and benefit realization. A 90-day staged plan reduces uncertainty while requirements remain changeable. Traceable evidence allows management to approve, revise or stop the project with more confidence.

When the baseline owner, measurement method, acceptance responsibility and benefit-review date are defined in advance, the decision record survives personnel changes. Keep updating what is fact and what remains an assumption at every gate. That discipline separates buying equipment from solving a business problem and makes capital decisions more repeatable.

TOMAS TECH can support Thailand factories from loss-baseline definition and concept comparison through equipment and integrator requirements, FAT/SAT and production-data integration. If you want to clarify a project boundary before issuing an RFP, please use our contact page.

Primary references

• Thailand BOI: first-half 2026 investment applications
• BOI OSOS: H1 2026 investment summary
• Thailand BOI: current Thai automation page
• Thailand BOI: English automation page
• ISO 10012:2026: measurement management systems
• ISO 12100:2010: machinery risk assessment and risk reduction
• ISO 13849-1:2023: safety-related parts of control systems
• ISO 5459:2024: datums and datum systems
• ISO 1101:2017: geometrical tolerancing
• ISO 14253-1:2017: conformity decision rules
• U.S. OSHA 1910.212: general machine guarding