INDUSTRY PLAYBOOKS / Industry Playbooks

Fitting general AI capability into a specific industry

The same model behaves completely differently in different industries: what data it must reach, which permissions it must pass, and how costly a wrong answer is. We currently put most of our effort into intelligent manufacturing and higher & vocational education.

01 / FOCUS AREAS

Intelligent manufacturing · Higher & vocational education

Both share the same traits: knowledge-intensive, fast staff turnover and expensive failure — and AI can produce measurable improvement in all three.

Focus area

Intelligent manufacturing & industrial equipment

Common problems

  • Equipment manuals, process documents and past work orders are scattered across systems; frontline staff have to ask a colleague to find anything
  • Experienced engineers' knowledge is hard to capture; new staff ramp up slowly and after-sales response depends on individual ability
  • Export markets push ESG, packaging and carbon-data requests onto the factory as questionnaires, which is costly to answer manually

How we approach it

  • Equipment and process knowledge base: manuals, drawing notes and work orders become searchable Q&A
  • After-sales and channel technical-support assistant: frontline self-service, with complex cases escalated to a human
  • Export compliance data preparation: automatic collection and verification of supplier ESG and packaging compliance data

Typical deliverables

Equipment & process knowledge baseAfter-sales support assistantStructured records and reportingCompliance data preparation
Focus area

Higher & vocational education

Common problems

  • Courseware, practical-training manuals and teachers' expertise are scattered; student questions concentrate on a few teachers
  • Equipment operating procedures are hard to demonstrate repeatedly — one person explains, many watch — and effectiveness depends on space and staff
  • Course evaluation and teaching outcomes rely on manual collation, and experience is hard to carry across cohorts

How we approach it

  • Course knowledge base and teaching Q&A assistant: authorisation tiered by programme, course and class
  • Practical-training guide: equipment procedures turned into step-by-step, queryable guidance
  • Teaching analytics: aggregate frequent questions and weak spots to feed back into course improvement

Typical deliverables

Course knowledge baseTeaching Q&A assistantPractical-training guideTeaching analytics dashboard
02 / OTHER AREAS

Four more areas we also work in

Our delivery experience in these areas is equally mature, and they make good first pilots for AI.

Cross-border trade & overseas services

Common problems

  • Regulations and market-access requirements change quickly across markets, and teams struggle to keep up
  • The same compliance material has to be redone for every client and every market
  • Files sit in mailboxes and on personal computers — when someone leaves, the knowledge goes with them

How we approach it

  • Regulatory and market-access knowledge base: search by market and category, with citations to the clause
  • Compliance material automation: turning source data into disclosure working papers as a process
  • Past projects distilled into reusable templates to cut repetitive work

Typical deliverables

Regulatory knowledge baseMaterial automationTemplates and ledgers

Retail & consumer goods

Common problems

  • Product information, promotion policies and store operations manuals are voluminous and change often
  • Customer service and shop-floor teams are stretched, and answers are inconsistent
  • Marketing content is slow to produce and hard to keep consistent across languages and markets

How we approach it

  • Product and operations knowledge base: one shared source of truth for stores, support and distributors
  • Shopping assistant and customer-service assistant: connected to WeCom / official accounts, self-service first, human second
  • Content generation assistance: multilingual drafts per brand guidelines, with a human final review

Typical deliverables

Cross-channel service assistantMultilingual content generationQuality and performance dashboard

Energy & utilities

Common problems

  • Safety and operating procedures run to many clauses, and on-site lookups and execution are error-prone
  • Operational data and text records are separate, and fault reviews depend on manual collation
  • Strict requirements on data residency and access isolation

How we approach it

  • Procedure and policy knowledge base: natural-language search with clause-level citations
  • Operations assistant: recommendations and rationale grounded in the asset register and past events
  • Deployment in the region you specify, with authorisation tiered by role

Typical deliverables

Knowledge base and Q&AOperations assistantPermissions and audit framework

Professional services & consulting

Common problems

  • Project material and industry experience are poorly reused — knowledge leaves with the people
  • Proposals and reports involve a lot of repetitive writing, and documentation stretches delivery
  • Training new joiners depends on apprenticeship and does not scale

How we approach it

  • Experience knowledge base: project documents, cases and templates become reusable assets
  • Proposal drafting assistance: first drafts from templates and past cases, finalised by a human
  • Knowledge transfer: tacit experience turned into searchable Q&A and checklists

Typical deliverables

Knowledge asset managementContent drafting assistanceTraining and handover tools
03 / COMPLIANCE PLATFORMS

Two compliance scenarios we have turned into products

For manufacturers expanding overseas, the first obstacle is rarely technical — it is preparing and evidencing compliance data. We have turned these two high-frequency scenarios into platforms.

PLATFORM 01AI + blockchain

Supplier ESG Compliance Platform

ESG data collection, verification and attestation

For companies with cross-border supply chains: AI collects, cleans and structures supplier ESG data scattered across questionnaires, emails and spreadsheets, while blockchain attests the key figures — so green compliance no longer depends on chasing Excel files by hand.

  • Automatic triage of supplier ESG questionnaires, answer-quality checks and gap prompts
  • Alignment of emission and energy metrics, anomaly detection and trend analysis
  • Key data attested on-chain: traceable provenance, tamper-resistant records
  • Disclosure working papers and compliance reports for CSRD, CBAM, EUDR and similar requirements
PLATFORM 02AI + regulatory knowledge base

EU Packaging Compliance Platform

PPWR readiness, gap analysis and documentation

For packaging and consumer-goods companies exporting to the EU, we turn the PPWR text into an executable checklist: starting from your packaging list and BOM, we compare recyclability, recycled content and reduction requirements item by item, and output a gap list, technical documentation and an annual maintenance ledger.

  • Packaging list and component BOM, with weight, material and structure recorded line by line
  • Gap analysis and remediation advice on recyclability, recycled content and empty space
  • Regulatory text and source register, traceable to clause level and ready for review
  • Assembly of DoC technical files and EPR registration material for each market

These platforms provide data organisation and gap-analysis capabilities. They do not constitute legal advice or a compliance conclusion; a formal conclusion requires professional review against the actual documentation.

04 / WHERE TO START

Three common starting points — we suggest the first

If you are unsure where to begin, we usually suggest an internal knowledge assistant: short cycle, low risk, and the fastest way to see real results.

01

Internal knowledge assistant

The safest starting point. The data is inside the company, the users are employees, and mistakes can be caught by a human. Ideal as a first AI project to prove value.

  • Controllable investment, results in 2–6 weeks
  • No exposure to external compliance risk
  • Easy to extend horizontally once proven
02

External service assistant

Q&A and support for customers or channel partners. The value shows up directly in headcount and response time, but the bar for answer accuracy and content safety is higher.

  • Needs stricter knowledge boundaries
  • Human handover should be retained
  • Requires a content-safety policy
03

Compliance & process automation

Embedding AI in existing processes: material review, compliance data collection, report generation. The clearest return, but it needs deep integration with existing systems.

  • Processes and owners must be clear
  • Depends on how open your system interfaces are
  • Start with a single-step pilot
05 / NEXT STEP

Tell us your industry and scenario, and we will map a path

We will first judge whether it is feasible under your data and permission constraints, then propose a phased plan — no guarantees we cannot keep, no exaggeration.

Book a discovery callExplore AI capabilities