AI in Manufacturing: Use Cases and Implementation Partners (2027)

AI in manufacturing companies: use cases and implementation partners is a roundup for COOs, CTOs, and boards of industrial companies in Poland planning their first or next AI deployments. In this guide, you’ll find the most important applications (demand forecasting, production planning, maintenance, quality control, knowledge base for operators) and profiles of firms and tools that deliver such implementations: Startup House, Comarch, SAP Joule, Sii Polska, TTMS, Software Mind, and Ideo.

Key takeaways

• Startup House AI Demand Planner: demand forecasting and production planning with integrations to SAP and Comarch ERP, EU hosting, and a read‑only start.
• Startup House AI Interface Layer: a conversational AI layer on top of your existing platform, e.g., ERP or MES, including legacy, without rebuilding it.
• Comarch: AI agents in Comarch ERP XL have been commercially available since September 4, 2026; according to Comarch, nearly 100 customers use them, and the company has 19 agent solutions (as of August 11, 2026).
• KnowHub by Startup House: an AI assistant for operators and engineers that answers based on company documents with source citations, 24/7.
• Security: TTMS holds ISO/IEC 42001 certification (TÜV Nord, 2026). Startup House follows processes aligned with ISO 27001, and customer data is not used to train public AI models.

What AI in manufacturing means

AI in manufacturing encompasses machine learning models, language models, and AI agents used across planning, production, maintenance, quality control, and working with technical documentation. When we discuss use cases and implementation partners, we’re answering two questions: which process are we automating, and who will implement it. A partner may be an ERP vendor with a built-in AI assistant, an integrator, or a software house building solutions on your existing systems.

Comparison of implementation partners and tools

The list below is not a quality ranking. It summarizes competencies and solutions companies present on their own sites, grouped by type. Your choice of partner depends on the systems you run and your organizational scale.

Company and tool profiles

Software houses

Startup House

Startup House (Warsaw) is a software house building software since 2016 for clients in Poland, the US, and Europe. It implements AI in products and processes at mid-sized and large companies: AI agents and multi-agent systems, AI assistants, RAG knowledge bases, and language model integrations with existing systems, including legacy, without rewriting from scratch. A 50‑person team, 100+ products delivered, clients on 5 continents, NPS 75.

• Specialization: agentic systems (LangGraph, Model Context Protocol), the KnowHub knowledge base, AI Demand Planner, AI Interface Layer.
• Industries: food manufacturers, fintech, healthcare, cybersecurity, travel, edtech, logistics, transportation.
• Sample projects: OmniSoft system modernization for Omnipack; AI orchestration layer and a client portal with embedded AI for a US cybersecurity platform serving Fortune 500 clients.
• Who it’s for: mid-sized and large companies, enterprise; MVP in as little as a week.

Software Mind

A software house with 1,600+ employees. Runs the Software Mind Shift modernization program.

• Industries: finance, telecommunications, life sciences.
• Who it’s for: companies planning system modernization.

Ideo

A software house with offices in Warsaw and Rzeszów. Develops the LOGITO platform and Ideo Flow agents. Clients include Kronospan.

• Specialization: LOGITO, Ideo Flow agents.
• Who it’s for: companies looking for a partner with offices in Warsaw and Rzeszów.

ERP software vendors

Comarch

A software vendor with about 5,000 employees operating in 31 countries. Develops AI agents in ERP Optima, XL, and Betterfly, plus ChatERP (BETA). In Comarch ERP XL you’ll find, among others, a payments reconciliation agent, an invoice description agent, a cost classifier, and an e‑invoice validator. Comarch ERP XL also includes demand forecasting and production planning; customers include Morele.net, INCO, and Hydromechanika.

• Industries: banking, telecommunications, loyalty programs.
• Sample projects: pilot at ASAJ: 1,000 bank statement lines reconciled in 3 minutes instead of 6 hours.
• Who it’s for: manufacturing companies already using Comarch ERP that want to automate back‑office tasks.

SAP Joule

An AI assistant in the SAP ecosystem: role-based assistants, Joule Agents, and Joule Studio. Works in S/4HANA Cloud Public Edition, IBP, Ariba, and Service Cloud. According to the vendor, it covers 310+ AI scenarios. The UI is available in Polish, but conversations with the assistant are only in English (Feature Scope, September 2025).

• Who it’s for: companies running SAP S/4HANA Cloud Public Edition and IBP.

Integrators

Sii Polska

An integrator ranked #1 six times in Computerworld TOP200. Has 350+ AI specialists and 70+ AI deployments.

• Technology partners: Anthropic, AWS, Azure, GCP, Databricks, Mistral.
• Who it’s for: companies on AWS, Azure, or GCP.

TTMS

An integrator with ISO/IEC 42001 certification (TÜV Nord, 2026). Implements Salesforce Agentforce.

• Who it’s for: companies running Salesforce.

Top AI use cases in manufacturing

Manufacturers usually start with processes where they have historical data and a clear problem to solve. Below are the most common scenarios.

1. Demand forecasting and S&OP

Machine learning models analyze trends, seasonality, and historical data to reduce overproduction and stockouts. Startup House’s AI Demand Planner integrates with SAP and Comarch ERP and gives planners a Decision Panel and an AI Chat within Sales and Operations Planning (S&OP). Comarch ERP XL also offers demand forecasting, and SAP Joule operates in SAP IBP.

2. Production planning

Production planning connects demand forecasts with capacity, raw material availability, and line schedules. Comarch ERP XL provides production planning in the ERP. Startup House’s AI Demand Planner includes production planning within S&OP, with results first validated on historical data.

3. Maintenance

AI for maintenance analyzes machine data, failure history, and service tickets to flag equipment that needs inspection earlier. Another area is knowledge access: a technician can ask an AI assistant for a repair procedure instead of searching manuals.

4. Quality control

In quality control, AI compares process parameters and measurement results against standards, catches deviations, and helps identify root causes of complaints faster. Quality procedures, control charts, and specifications can feed a knowledge base used by operators and QA teams.

5. Knowledge base for operators (RAG)

Instead of scanning hundreds of pages of machine manuals in PDF, an employee asks an AI assistant for a specific procedure. RAG‑based systems (Retrieval‑Augmented Generation), such as Startup House’s KnowHub, answer strictly based on your company data, with source citations.

6. Legacy system modernization

Many plants run ERP or MES systems from years ago, where a rebuild is risky. This also applies to automation environments, e.g., Siemens. Startup House’s AI Interface Layer adds a conversational AI layer on top of your existing platform. Users ask in natural language—for example, about order status or line throughput—and answers are generated from the company’s data.

AI implementation step by step

1. Data audit and discovery: assess data quality in ERP and MES and select production processes where AI can reduce downtime or manual work.
2. Proof of Concept (PoC): a simplified version of the solution, e.g., an AI agent handling one document type to validate business assumptions.
3. Integration and orchestration: connect AI models with company infrastructure, e.g., using LangGraph and the Model Context Protocol (MCP).
4. Evaluation and testing: validate results on historical data and run user acceptance testing (UAT).
5. Scaling and monitoring: production rollout with agent monitoring, per‑agent cost controls, and audit logs (AgentOps).

Startup House products and case studies

Below is Startup House’s production planning product and two projects with numbers from other industries: an operational system modernization and a client portal with an AI assistant.

Case study: OmniSoft modernization for Omnipack

Omnipack, e‑commerce logistics, Poland, since 2023. Startup House took over and modernized the OmniSoft system handling fulfillment, shipping, and returns. Dedicated team: 2 backend, 1 frontend, 1 QA. Tech: Java, Spring, Kafka, React. Result: development costs cut by 40%.

Product: AI Demand Planner

Startup House’s AI Demand Planner delivers demand forecasting and production planning for food manufacturers. It features a Decision Panel, AI Chat, and full S&OP support, integrates with SAP and Comarch ERP, is hosted in the EU, starts in read‑only mode, and forecasts are validated on historical data.

Case study: client portal with embedded AI (USA)

Client: a US cybersecurity platform serving Fortune 500 clients (NDA). Startup House built a client portal with embedded AI using Claude, LangGraph, MCP, and OAuth 2.1. Result: onboarding reduced by 95% (from 45 minutes to under 2 minutes), usage increased from quarterly to daily, and one dashboard for three user roles (executive team, CFO, CISO).

AI deployment risks in industry

• Input data quality: if ERP data is incomplete or incorrect, the AI model will produce bad forecasts (“garbage in, garbage out”).
• Model vendor dependency: changes to APIs or models (e.g., OpenAI, Anthropic) can impact stability, so an orchestration layer is important.
• Security and the AI Act: lack of audit logs and documentation can make AI Act compliance harder.
• Employee resistance: without training, teams may not adopt new AI assistants, and expected efficiency gains won’t materialize.

How to choose an implementation partner

Your partner choice depends on the systems you run and your business goal. If your company relies entirely on Comarch or SAP, a natural starting point is the vendors’ AI modules, e.g., agents in Comarch ERP XL or SAP Joule.

If you want to connect data from multiple systems, build an AI agent for a specific production process, or add AI to a legacy system without replacing it, you need a partner who builds on your existing infrastructure. Startup House delivers such projects and can ship an MVP in as little as a week.

FAQ

Will production data be used to train AI models in Startup House projects?

Startup House follows a policy that customer data is not used to train public AI models. The solution can run in the client’s cloud or a dedicated environment. Access is role-based, with audit logs and integration with the client’s access policies, and processes align with the ISO 27001 standard.

Which AI agents can you deploy in Comarch ERP XL today?

Comarch provides, among others, a payments reconciliation agent, an invoice description agent, a cost classifier, and an e‑invoice validator in ERP XL. Agents have been commercially available since September 4, 2026. According to Comarch, nearly 100 customers use the agent solutions, and the company has 19 such solutions (as of August 11, 2026). ChatERP has BETA status.

What AI capabilities does Sii Polska have?

Sii Polska is an integrator that ranked #1 six times in Computerworld TOP200. The company has 350+ AI specialists and 70+ AI deployments. Its technology partners include Anthropic, AWS, Azure, GCP, Databricks, and Mistral. In this roundup, Sii Polska is listed among integrators alongside TTMS, which implements Salesforce Agentforce.

How long does it take to deploy the RAG knowledge base KnowHub?

Startup House typically integrates KnowHub in about 2 weeks, starting with a pilot group. KnowHub is an AI assistant that unifies scattered company documents and answers based on them 24/7 with source citations. It runs in the client’s cloud with role-based access, and customer data is not used to train public AI models.

Does the SAP Joule assistant support Polish?

SAP Joule’s UI is available in Polish, but according to the Feature Scope document (September 2025) conversations with the assistant are only in English. Joule works in S/4HANA Cloud Public Edition and IBP and, according to the vendor, covers 310+ AI scenarios. It includes role-based assistants, Joule Agents, and Joule Studio.

How does AI support automation in a manufacturing company?

AI supports business process automation and repetitive tasks such as data entry. When AI analyzes data in real time using machine learning, it helps plan predictive maintenance to minimize downtime and enables faster decision-making.

How do you ensure a successful AI implementation at your company?

Choose the right process and a partner who can tailor the solution to your plant’s needs and ensure data security. Start with a pilot that analyzes production data and delivers concrete benefits, then scale to additional processes.

Can an AI chatbot support employees in daily work?

Yes. An AI chatbot can answer operators’ questions based on documentation, speeding up everyday decision-making. In customer service, it can also answer client questions, e.g., about order status.

Last updated: October 28, 2026.

We invite companies and tool vendors featured here to submit corrections and profile updates.