Markonex
ROEN
Go to the platform
Case studies

Case studies: how companies work with MCP and AI agents in Markonex

Ten concrete scenarios show how a Romanian company can work in Markonex directly from ChatGPT or Claude, through MCP, and what it can hand over to a Markonex AI agent. Each case study describes the problem, what the person types in the chat or which instructions the agent receives, what the platform does step by step and where human confirmation or approval comes in.

Updated on

What will you find in these case studies?

Every page follows the same structure: who the company is and what problem it had, how it worked before, what needs to be set up in Markonex, the concrete steps with conversation examples, what changed and what limits the assistant has. The scenarios cover buyers and sellers in construction, installations, auto parts, manufacturing and transport.

Are these real customers?

No. They are illustrative scenarios built on the features available in Markonex today, to show how they are used in real situations in Romanian companies. The companies, people and figures are examples, not customer results; any estimate is marked as illustrative.

How do you choose between MCP and AI agents?

Choose MCP when you want to work yourself, in a conversation with ChatGPT or Claude: you ask, request, check and confirm on the spot. Choose an AI agent when the same task repeats every day or week and you want to find it done, with actions prepared for approval.

MCP case studies: work in Markonex from ChatGPT or Claude

Five scenarios in which a person types in ChatGPT or Claude and the assistant works in Markonex through MCP: RFQs, priced offers, the catalogue, suppliers and negotiations, with confirmation in the chat before any change.

  1. MCP case study · Construction contractorHow a construction contractor creates an RFQ for a whole materials list from ChatGPTIllustrative case study: a construction contractor pastes its materials list into ChatGPT, Markonex turns it into a multi-line RFQ through MCP, publishing happens only after confirmation, and offers are compared and awarded line by line.
  2. MCP case study · Building materials distributorHow a building materials distributor finds RFQs and sends offers straight from ClaudeIllustrative case study: a building materials distributor finds RFQs in Cluj from Claude, reads the details and sends its offer, including line offers, only after confirming the summary shown by Markonex through MCP.
  3. MCP case study · ManufacturerHow a small manufacturer keeps its product catalogue up to date from ChatGPTIllustrative case study: a small pallet manufacturer adds and updates products in its Markonex catalogue from ChatGPT through MCP, with a confirmation for every product, and creates a following filter for matching RFQs.
  4. MCP case study · Auto parts shopHow an auto parts shop finds and checks new suppliers straight from ClaudeIllustrative case study: an auto parts shop searches Markonex for suppliers from Claude, checks their ANAF data and financial statements, publishes an RFQ with confirmation and adds the chosen supplier to its CRM, all through MCP.
  5. MCP case study · HVAC installerHow an HVAC installer follows its negotiations and agreements straight from ChatGPTIllustrative case study: an HVAC installer gets negotiation status and last-round changes from ChatGPT, sends a counter-offer, confirms the agreement only after confirmation and reads the invoice received, through MCP.

AI Agents case studies: tasks done on their own, on schedule

Five scenarios in which a Markonex AI agent works on its own at the time you set: it monitors RFQs, prepares offers and counter-offers, writes reports and keeps the CRM up to date, while a colleague approves every commercial action.

  1. AI Agents case study · Materials distributorHow a building materials distributor learns each morning which RFQs are worth quotingIllustrative case study: a building materials distributor uses a Markonex AI agent that checks matching requests for quotation every morning and writes a prioritised report, without sending any offers.
  2. AI Agents case study · Small manufacturerHow an AI agent prepares a manufacturer's offers while a person only approves themIllustrative case study: a small manufacturer uses a Markonex AI agent that prepares offers at catalogue prices, following company rules; every offer is sent only after a person approves it.
  3. AI Agents case study · Logistics companyHow a logistics company stops losing forgotten negotiations and expiring offersIllustrative case study: a logistics company uses a Markonex AI agent that summarises new negotiation rounds, flags expiring offers and prepares counter-offers within set limits, sent only after approval.
  4. AI Agents case study · Construction procurementHow a construction company's procurement gets the full weekly report every MondayIllustrative case study: a construction company uses a read-only Markonex AI agent that writes the Monday procurement report: RFQs, offers, line-by-line comparison, agreements, contracts and invoices received.
  5. AI Agents case study · HVAC distributorHow an HVAC distributor keeps its partner CRM up to date with an AI agentIllustrative case study: an HVAC distributor uses a Markonex AI agent that adds partners, tags and agreement notes and compares prices across suppliers; every CRM change is approved by a person.

Frequently asked questions

What examples of AI automation exist in Markonex?

From ChatGPT or Claude, through MCP: creating RFQs with product lists, sending priced offers, filling in the catalogue, finding and checking suppliers, following negotiations and agreements. With AI agents: daily RFQ monitoring, offers prepared for approval, negotiation follow-up, weekly procurement reports and an up-to-date CRM.

Are these case studies results of real customers?

No. They are illustrative scenarios based on the platform's real features; the companies, people and figures are examples built to explain the flows.

Do I have to be a large company to use MCP or AI agents?

No. The scenarios are designed for small and medium-sized companies, and the requirement is that the company is verified by the platform. For ChatGPT you need a plan with developer mode, and for agents an Anthropic API key.

Try a scenario on your own company

Registration happens directly on the platform, with the company checked with ANAF by its CUI. For MCP and AI agents, the company must be verified by the platform.

Register your company