Skip to main content

Services · Service

AI Integration Consulting

Coming soon

From process discovery to live pipeline — with engineering ownership.

Request information Service scope
Glass layers rising from process map to pilot and scale, linked by a path of light.

Overview

The problem in the field. Management says "we should be using AI" — but nobody knows in which process, with which data or for what gain. Vendor demos are impressive but not connected to company data and systems. Pilots start without metrics, end with "not bad" and never reach production. Data privacy, KVKK and new AI regulation are unclear; IT does not want the cloud, while the business unit wants quick results. In an industrial company the data sits in SCADA and on paper, and AI teams do not know that world.

The Elmes approach. Discovery workshops produce a process and data inventory; scenarios are scored on impact, feasibility and risk; for each scenario an ROI model (person-hours, cost of errors, time) and a list of required data and access are written up. Infrastructure options (cloud API / on-prem / local model) are compared by data class and cost, and a compliance framework (KVKK, AI Act risk class, human oversight) is defined. The pilot is designed with metrics, run by the AI Agent Development & Integration team or the customer's own team, measured, and closed with a go/no-go decision. The scaling plan covers architecture, governance (permissions, monitoring, cost, model lifecycle), team and training, and handover.

The result. A measured pilot and a reasoned roadmap; a vendor-neutral infrastructure decision; a compliance framework; a sustainable in-house team. AI stops being a "project" and becomes a capability in operation.

CriterionManagement consultantAI product vendorElmes
Process / ROI analysisYesSales-drivenYes, validated by a pilot
Technical ownership (builds the pilot)NoWith its own productYes (with AI Agent Development), vendor-neutral
Industrial data (SCADA, PLC, devices)UnfamiliarRarelyOwn infrastructure (SCADA, telemetry, gateway)
Infrastructure neutralityYesNoYes (cloud / on-prem / local model)
Compliance framework (KVKK, AI Act)LegalLimitedTechnical framework + legal referral
Handover and in-house teamReportTraining (on the product)Training + governance + handover
A path of light advancing through glass layers for process map, pilot and scale.
Concept

This product is in development or field testing. Contact us for technical information, pilot use and a preliminary quotation.

Features

  • Process & data discovery: workshops, system review and sample data assessment; a data quality note.
  • Scenario scoring matrix: impact × feasibility × risk; out-of-scope actions (physical commands, decisions on personal data) are flagged from the start.
  • ROI modeling: person-hours, cost of errors, time, LLM/infrastructure cost; sensitivity analysis.
  • Infrastructure selection: cloud API / on-prem / local model; comparison of data residency, latency, cost and lock-in risk; recommendation for a multi-provider abstraction.
  • Pilot → production: a pilot with metrics, a go/no-go decision, governance and architecture.
  • Training & handover: separate modules for management, business units and IT/engineering.
  • Industrial focus: SCADA, telemetry and device-data scenarios; data access through the SCADA & Industrial Automation, Telemetry & Process Intelligence and Wireless Gateway / RTU infrastructure.
  • Compliance framework: KVKK data flows, AI Act risk class, human oversight and logging.
  • TargetAI Workspace setup & customization: installation of the AI Workspace platform and adaptation to the organization.
  • TargetAtlas knowledge graph data migration: moving company data into the Atlas knowledge graph of AI Workspace.
  • TargetTechnical documentation templates: ready-made document templates for the compliance framework.

Service scope

Indicative durations and pricing depend on project scope; final figures are set in the quotation.

Values marked “Target” are design targets, next-generation values or chip-vendor data; they are updated as measurement and certification are completed.

Service scope
Scope (included)

Process/data discovery, scenario scoring, ROI model, infrastructure and model selection, compliance framework, roadmap, pilot design, execution management and measurement, production rollout plan, governance, training, handover; optional AI Workspace setup (the platform is in closed early access)

Scope (not included)

Agent/software development (separately, through the AI Agent Development & Integration or Industrial Software Development services); LLM/cloud/GPU costs; legal opinions (referral and the technical framework are included); enterprise data warehouse projects (planning included, implementation separate); change management and HR processes (recommendations only); clinical decision support

Process

Discovery → Pilot → Scale

Typical duration
Indicative

Discovery 2–4 weeks; pilot 4–8 weeks; scaling 4–8 weeks; 4–16 weeks in total

Deliverables
Target

Process and data inventory; data quality note; scenario scoring matrix; per-scenario ROI model (with sensitivity analysis); list of required data and access; infrastructure comparison matrix and recommendation; compliance framework (KVKK data flows, AI Act risk class, human oversight); roadmap; pilot design and metrics document; pilot results report with go/no-go decision; scaling plan (architecture, governance); training modules and handover package

What we need from you

An executive sponsor; business unit and IT representatives (workshop attendance, about 4 hours a week); process documents and sample data (may be anonymized); system inventory and API information; cloud/data policy; users and budget for the pilot (LLM/infrastructure); decision turnaround within 1 week

Team / tools

Lead consultant (AI/systems engineer), integration engineer (ERP/SCADA), data analyst; workshop, scenario scoring and ROI model templates, infrastructure comparison matrix, compliance checklist; the AI Agent Development toolset for PoCs; Elmes DM (work management)

Pricing model
Indicative

Discovery: fixed fee. Pilot management: fixed fee (development quoted separately). Scaling: fixed fee or per person-day. Subscription: a monthly fixed-fee pool of consulting days. Training: per day

Acceptance and commitment

Deliverables are accepted against acceptance criteria (content and completeness); no guarantee of realized ROI is given, but validation through a pilot is committed. Continuity is provided through a subscription

Roadmap update
Target

Once within 6 months of delivery

Applications

  • Manufacturing
  • Logistics
  • Municipal and public services
  • Healthcare administration
  • Maintenance and quality reports
  • Technical document search
  • SCADA and telemetry interpretation
  • Predictive maintenance explanations

Sector notes. In manufacturing: maintenance records, quality reports, shift summaries and technical document search; in logistics: order/correspondence processing and planning support; in municipal and public services: request classification, document search and interpretation of water/infrastructure telemetry; in healthcare: administrative processes only — clinical decision support is out of scope. On the industrial side, the targets are engineering document generation and predictive maintenance explanations based on Telemetry & Process Intelligence and VibroBal data.

Competence: Elmes's own AI products. The consulting draws on Elmes's experience building and operating its own AI tools:

  • Agent-based software operations with Elmes DM.
  • A multi-step AI production pipeline with AutoPCB.
  • An AI-first shell on DiscOS.
  • Internal MCP-based agents.
  • The AI Workspace platform, in closed early access.

These are Elmes's own products and internal use cases; they are not presented as customer references. No customer case study has been published yet.

Compliance & documentation

Contract and intellectual property

  • Contract: consulting agreement + SOW; subscription annex. Pilot development is contracted under the AI Agent Development & Integration agreement.
  • Confidentiality: mutual NDA; process information, data and reports are confidential; use as a reference only with written permission.
  • Intellectual property: customer-specific reports, ROI models, roadmap and pilot documents belong to the customer; Elmes templates, the scoring method and workshop materials remain with Elmes.
  • Neutrality: whenever Elmes's own products (AI Workspace, AI Agent Development) are recommended, this is stated openly and alternatives are presented as well.
  • KVKK / GDPR data-flow assessment
    Data inventory and personal data flows during discovery; processor agreement with the LLM provider and cross-border transfer assessment; anonymization recommendations
    Self-declared
  • EU AI Act risk classification
    Classification into prohibited / high / limited / minimal risk; technical documentation, human oversight and record-keeping obligations for high-risk areas. Tracking of regulatory developments in Türkiye is planned
    Self-declared
  • Security: data classification, least privilege, provider data retention terms, IEC 62443 principles
    IEC 62443 principles are applied to industrial connections
    Self-declared
  • Automated decisions about individuals (HR, credit)
    Out of scope by default, or a human decision is mandatory
    Out of scope
  • Legal opinion
    We work together with your legal team
    Out of scope
  • Consulting templates (discovery form, workshops, scenario scoring, ROI model, infrastructure matrix, compliance checklist, pilot report, training materials)
    Target

Frequently asked questions

If you are not selling a product, what are you selling?

The decision and the pilot: in which process, with which data, on which infrastructure and for what gain; then technical ownership and measurement of the pilot. If development is needed, it is quoted separately under the AI Agent Development & Integration service.

Will you recommend your own AI platform?

If it fits, yes — and we say so openly; alternatives are compared as well. The infrastructure choice is based on data class, cost and lock-in risk.

Does our data have to leave the company?

No. On-prem and local model options are evaluated; if the cloud is preferred, it comes with data classification and contractual safeguards.

Do you guarantee ROI?

No. The ROI model is built on assumptions and validated by the pilot. If the pilot does not meet its metrics, a "no-go" decision is issued with written reasoning.

How long does it take?

As an indication: discovery 2–4 weeks, pilot 4–8 weeks, scaling 4–8 weeks; 4–16 weeks in total.

We are a manufacturer and our data is in SCADA and on paper. Will AI help?

That is exactly what discovery answers. Access to SCADA and telemetry data is Elmes's own field of expertise; for paper-based processes, a data preparation plan is drawn up.

What do you do about KVKK and AI regulation?

We set up a technical compliance framework: data flows, risk class, human oversight, logging. For legal opinions, we work together with your legal team.

Do you have references?

No customer case study has been published yet. Elmes's experience building and operating its own AI tools is the in-house proof.

Can our team carry on independently afterwards?

That is the goal: a governance document, team roles, training and a handover package, with optional ongoing guidance through a subscription.

Let's talk about your project

Our engineering team replies within 24 hours.

Request information +90 (312) 395 85 86

Related products