How to Assess Smart Factory Readiness Before Funding an APAC Pilot
A factory can stream equipment data and still be unprepared to act on it. A dryer may report temperature and load weight, but if the team cannot link those readings to a batch, verify final moisture, or approve a change to the drying cycle, it cannot tell whether an optimization pilot improved production.
In McKinsey’s 2025 global survey of 101 senior manufacturing executives, including respondents with most of their sites in APAC, 46% reported limitations in their data or IT/OT systems as they pursued AI in operations. That finding makes a case for checking the data path and plant workflows before pricing the pilot build. [1]
A factory is ready for a smart factory pilot when it can measure the target loss, collect trustworthy data, act on the result through an approved workflow, and verify the outcome without disrupting production.
1. Set the assessment boundary
Scope the assessment to one plant, one process, and one operating decision. Singapore’s Smart Industry Readiness Index (SIRI) assesses maturity across process, technology, and organisation. Use those dimensions to structure the review, then assess the proposed use case against specific evidence from the line. [2]
For a dryer pilot, ask whether operators can shorten eligible cycles while keeping finished moisture within the approved range. Name the equipment, product families, shifts, quality checks, and systems in scope. Agree with finance and quality on the baseline, including energy per kilogram of accepted output, cycle duration, and moisture compliance. Define how to handle rework, atypical loads, and downtime before collecting comparison data.
Record the funding decision the assessment must support. TMA’s Innovation as a Service offering covers feasibility testing, prototyping, architecture comparison, and roadmaps with budget and schedule estimates. Use discovery to determine which of those activities the plant needs before committing to implementation. [3]
2. Review process and people readiness
Walk the process during a normal shift and a changeover. Ask the operator to show how a batch starts, how exceptions are recorded, and how quality releases the output. Compare the documented procedure with what staff actually do. Include the shift handover so the proposed workflow accounts for work that crosses teams. [2]
Give each pilot decision an owner. Production owns the operating response, quality owns release criteria, maintenance owns sensor and equipment checks, and OT engineering approves access and control changes. Rehearse an exception before the build. If a recommendation conflicts with the operator’s assessment, establish who decides, how the override is recorded, and how the team investigates it. [2]
For a regional programme, agree on shared metric definitions and keep site-specific work instructions alongside them. A second plant may use the same KPI while relying on a different recipe, approval sequence, or shift calendar. Review those differences with the receiving site before budgeting replication. [1][2]
3. Review technology and data readiness
Trace an actual batch from source to decision. Identify the programmable logic controller (PLC), sensors, and any existing historian or supervisory control and data acquisition (SCADA) system. Then locate the production record and the quality result. TMA’s published dryer solution captures machine, sensor, and load-cell data to estimate moisture during the cycle. A proposed pilot can test that estimate against the plant’s approved final moisture measurement. [4]
Check timestamps, units, asset identifiers, and batch boundaries against real records. Walk through a reconnect and a product changeover. Specify where the system will detect missing readings, reconcile buffered events, and distinguish a stale recommendation from a current one. Start with approved read-only acquisition and document the network boundary between collection and machine control. NIST’s OT security guidance addresses security alongside plant performance, reliability, and safety. [5]
Use the table to record the evidence required for the proposed architecture. [4][5]
| Assessment area | Evidence gap | Evidence to release pilot funding |
|---|---|---|
| Equipment interface | Access depends on an undocumented controller connection | Approved read path demonstrated on the target asset |
| Batch context | Sensor logs and quality records use different identifiers | Sample cycles matched to the correct recipe and final result |
| Connectivity | Recovery behaviour is unknown | Disconnect and reconnect test with gaps and duplicates accounted for |
| Operator workflow | Dashboard has no response procedure | Named owner, action window, override, and disposition record |
| Control boundary | Analytics can reach control functions without a documented policy | Approved network paths, permissions, and change authority |
Lessons Learned from the Field
The distinctive problem in TMA’s dryer example is estimating remaining moisture during the cycle. The published design combines equipment signals and load-cell readings to estimate a process condition that matters to the operator’s decision. TMA describes the sensing and prediction approach. The public page provides qualitative benefits, so set energy and throughput targets from your plant baseline. [4]
This is where a low-cost offshore development center scoped mainly to software tickets can leave an expensive gap. Ask both vendors to explain how they would match an estimate to a physical batch, verify it against a quality measurement, and handle a missing signal. The engineering proposal should assign ownership across instrumentation, data, validation, and the operator workflow. [3][4]
4. Turn gaps into pilot prerequisites
Convert each gap into a work item with an owner, due date, and acceptance test. A missing batch identifier becomes a test that joins complete cycles to production and quality records. An untested recovery path becomes a witnessed disconnection exercise. Prioritise gaps that prevent safe operation or invalidate the result, then cost their resolution separately. [3][5]
Include security and privacy evidence in procurement. TMA lists ISO/IEC 27001 in its quality framework and reports incorporating GDPR checks into software quality assurance. Verify the certification scope for the proposed service. If records identify operators, agree with the site’s privacy team on access, retention, and any regional transfer requirements. [6][7]
Technical deployment pipeline
Use this proposed sequence to decide when the site can progress from assessment to supervised operation. [3][4][5]
- Baseline the process. Agree on eligible batches, loss measures, and quality limits.
- Prove the data path. Verify controller access, batch matching, and recovery from interrupted collection.
- Replay historical cycles. Compare recommendations with known outcomes across representative recipes and load sizes.
- Run in shadow mode. Record recommendations alongside normal production decisions and have operators review discrepancies.
- Trial supervised use. Give authorised staff control over interventions and record outcomes, overrides, and exceptions.
- Review the investment. Compare results with the baseline and price the work required for another site.
5. Use readiness to stage investment
Release funding in stages. Fund discovery when the opportunity is credible and evidence is incomplete. Fund the pilot once the data, ownership, security, and measurement prerequisites pass. Fund replication once quality and finance accept the results and the receiving site verifies its interfaces, procedures, and support arrangements. [1][3]
For dryer optimization, compare similar batches and report energy and cycle time alongside moisture compliance. Cost sensors, integration, commissioning time, and ongoing support. Agree up front on the minimum result that would justify another site and the conditions that would trigger a redesign or stop. [3][4]
TMA reports 29 years of software engineering, 4,000 engineers serving customers in 30 countries, and experience delivering mission-critical programmes involving hundreds of engineers. Its telecom practice also describes building customer R&D centres. Those capabilities give buyers a basis to evaluate TMA for both the initial factory implementation and a longer-term R&D partnership covering integration, validation, and platform evolution. [6][8]
FAQ
What should a smart factory readiness assessment produce?
Produce a scoped use case, measured baseline, data and integration map, named owners, tested prerequisites, and costed funding decision. Tie each prerequisite to evidence the plant can verify. [2][3][5]
Can a plant with legacy equipment qualify for a pilot?
Yes, if the target asset has an approved collection path and the team can link readings to production and quality records. Verify access, signal quality, and recovery on the actual machine. [4][5]
When should a pilot move beyond shadow mode?
Move to supervised use after quality approves the comparison results, operators can handle exceptions, and OT engineering signs off on permissions, recovery, and any proposed intervention. [4][5]
How can buyers evaluate a smart factory vendor’s reliability?
Request relevant plant references, named OT and data engineers, and a demonstration of how the team handles missing signals or interrupted connectivity. Verify certification scope and ask who owns commissioning, recovery, and ongoing support. [3][5][6]
What should trigger a hold on pilot funding?
Hold funding when the plant cannot establish a credible baseline, match data to the target process, approve access, or name an owner for the intervention. Fund the specific work needed to resolve those gaps. [2][3][5]
Start with a funding decision the plant can defend
Bring one target line, a sample of its operating data, and the people responsible for production and quality to a TMA Innovation readiness workshop. Use the session to define the pilot boundary, identify the evidence still needed, and cost the next funding stage. [3]
References
[1] McKinsey — From pilots to performance: How COOs can scale AI in manufacturing, December 2025.
[2] INCIT — Smart Industry Readiness Index.
[3] TMA Solutions — Innovation as a Service.
[4] TMA Innovation — Dryer Machine Optimization.
[5] NIST — Guide to Operational Technology Security, SP 800-82 Rev. 3.
[6] TMA Solutions — Company Overview and Engineering Capabilities.
[7] TMA Solutions — CSR Report 2024, Security and Compliance.
[8] TMA Solutions — Telecommunications Engineering and R&D Centres.


