ChatGPT and Codex training for consulting and trade fair companies in Düsseldorf

Proposals, reports and the tools in between

A consulting firm writes a proposal to win a project and a report to close it. A trade fair company writes exhibitor letters, stand proposals and a report after each fair. These documents are long, follow a pattern and carry details that belong to one client only.

ChatGPT drafts such documents fast. The risk is a sentence from another client’s proposal, a number without a source or a promise nobody agreed to. The training shows your staff how to draft with ChatGPT and how to review what it wrote.

Next to the documents run the internal tools, such as a script that builds the report from a spreadsheet or a small app for stand bookings. Often one person wrote the tool, and nobody else changes it. Codex can add what the tool lacks, under rules.

The ChatGPT track: proposals and reports

• A project per service line that holds your method descriptions and the past proposals your company has released for reuse.
• The client’s question restated in the client’s words, as the first section of a proposal.
• Report chapters from interview notes and data, with each figure traced to its source.
• Slide texts from the report, for the meeting with the client.
• The review sheet below, used on each draft.

The Codex track: internal tools

• An AGENTS.md that names who uses the tool, which data it touches and how to run it.
• Tests for the tool’s current behavior before Codex adds a feature.
• A README and a short operating guide, written by Codex and checked by the developer.
• Sample files in place of production data in each test.
• Access limits for the agent, as described on AI agents under the same rules as humans.

A review sheet for drafts written with ChatGPT

Each question needs a yes before a proposal or a report leaves the company.

1. Does the first section state the client’s question in the client’s words?
2. Does each number have a source in your records?
3. Are names, dates and prices checked against your files?
4. Is the draft free of text from another client’s proposal or report?
5. Are the people in the team section available for the project?
6. Has your company agreed to each promise in the text?
7. Has a person who knows the client read the whole draft?

Who I am

I am Michael Wutzke, an IT architect for data platforms in Frankfurt. In several due diligences, for buyers and for sellers, I examined and prepared the documents on the quality of IT systems and teams. I have written and reviewed contracts, licensing agreements and terms of use for business development and IT. Earlier in my career I worked on exhibitions and annual reports for companies. Today I build software with Codex and Claude for a client and teach at Claude Hacker House. I am also interested in open-source AI models that a company runs on its own servers. Details: Career stages.