I wanted an overview of my ‘Ask better. Work smarter. ChatGPT in business’ training programme. Rather than simply writing that the course was excellent, I asked AI to analyse its content against the market and published research. This article documents that exercise. The resulting AI-generated assessment is not an independent evaluation or a verified promise of business results.
Stage 1: a considered input
The input matters. My requests went beyond ‘compare this with AI training in Estonia’ and asked it to:
- Research the Estonian AI-training landscape and compare it with my programme’s broad focus on in-company use .
- Consider the training’s potential return on investment (ROI), referring to productivity research associated with MIT and Harvard Business Review .
- Suggest a market price by comparing public training fees and expert day rates.
ChatGPT Deep Research produced a three-page analysis with source references. It was an AI-generated assessment, not an independently commissioned review.
Stage 2: turning the analysis into different formats
I then asked NotebookLM to convert the report into several formats as an exercise in efficient content production:
- Podcast: An interview-style audio discussion to make the material conversational.
- Video slides and visuals: Summary slides with the main points and figures.
- A blog article
The exercise shows how one carefully prepared analysis can become three content formats.
Created with: NotebookLM Audio Overview. The original experiment’s AI-generated audio is in Estonian. Its assessments are not an independent review or a current quotation.
Created with: NotebookLM Video Overview. The original experiment’s AI-generated video is in Estonian. The text below clarifies the research context and assumptions.
Training programme: ‘Ask better. Work smarter. ChatGPT in business’
Created with: NotebookLM Reports
The following is an edited version of the original AI-generated training overview. It describes the intended programme content. Indicative prices are not a current quotation, and the research findings are not measured outcomes of my training.
The programme’s core idea
‘Ask better, work smarter’ treats effective AI use as a developing workplace skill. The aim is not only to introduce a tool, but to help employees apply ChatGPT systematically to tasks where the usefulness of its output can be assessed.
The programme presents ChatGPT as a work assistant. A clear objective and relevant context help, but do not guarantee a correct answer. You provide the objective and boundaries; AI helps create, explain and structure content, and a person checks the result.
The programme therefore focuses on practical application and reviewing the result, with benefits assessed against the team’s tasks.
What does the programme focus on?
AI-training courses vary in their scope and practical relevance. This programme focuses on applying the tools within a company’s day-to-day work.
• Practice before abstract theory Training is built around workplace examples and techniques participants can try themselves: writing clearer requests, integrating ChatGPT into workflows and handling recurring tasks. The intention is to develop skills they can continue using after the session.
• A detailed focus on prompt design The programme develops a structured approach to prompt design. Participants move beyond trial and error towards a repeatable method, learning how the information they provide shapes the result and what still requires checking.
• Business context The focus is in-company use across tasks, rather than only a broad strategic overview or one area such as marketing. Examples and exercises are adapted to the team’s business context.
The programme’s value is best assessed through the practical skills each participant can use afterwards.
Main topics and skills
The programme combines principles of using ChatGPT with practical exercises. The table summarises the intended learning areas and skills.
| Area | Topics and skills |
|---|---|
| Foundations | Distinguish model and tool capabilities, including terminology such as LLM, LRM and LAM. Choose a suitable tool and apply critical fact-checking rather than assuming every output is reliable. |
| Writing precise requests | Structure requests with an objective, facts, constraints and output format. Ask AI to help clarify or improve a request, then assess whether it covers the task. |
| Working with ChatGPT | Create and manage custom assistants for recurring tasks. Use appropriate modes, such as Research and Agent mode, for more complex, multi-step work where available. |
| Personalisation and efficiency | Configure Memory and personal preferences deliberately, considering context, consistency and the information appropriate to retain. |
| Privacy and security | Apply data-protection principles and review settings before sharing sensitive business information. Understand technical limitations, including specialist file formats such as DWG, to avoid unsupported assumptions. |
These skills are intended to support productivity, work quality and experimentation. Their effect needs to be measured in the organisation’s own workflows.
Potential benefits for the business
Training is an investment in employees’ ability to use AI appropriately. The following are potential benefits, not assured returns.
Time required. In Shakked Noy and Whitney Zhang’s 2023 study published in Science, 453 college-educated professionals completed occupation-specific writing tasks. Access to ChatGPT reduced average task completion time by 40%. This is evidence of potential benefits for particular tasks, not a measured result of my course or a forecast for every employee.
Work quality. The same study reported an 18% increase in assessed output quality. MIT’s explanation of the research notes that the tasks did not require precise factual accuracy or detailed knowledge of a company or customer. Workplace use also needs time for fact-checking and corrections. These figures do not demonstrate the impact of this training programme.
• Experimentation and creativity Reducing repetitive steps may leave more attention for developing ideas and solving problems. Training includes trying alternative approaches and exploring ‘what if’ scenarios, followed by a review of their usefulness.
• More informed use of technology The programme aims to improve confidence and digital skills while covering confidentiality, permissions and data protection. Knowing when not to share information or rely on an answer is part of effective AI use.
To assess the return, compare the time and quality of a selected task before and after adopting a new workflow. Include tool costs, learning time and review. A general study of AI does not replace that measurement.
Format and indicative pricing in the original analysis
The following records the format and price assumptions used in the original AI analysis. They are not a current price list or a verified market comparison. The scope and cost of a specific session are agreed separately.
• Duration: 2-4 hours, depending on the company’s needs and required depth.
• Group size: The analysis proposed 2-8 participants to allow individual attention and discussion of specific work questions.
• Indicative base price in the analysis: Approximately €800 plus VAT for one group session; not a current quotation.
Pricing comparison in the analysis: The AI-generated report used a group-session range of EUR 650-950 and a public-course fee of EUR 239 per person. I am not confirming those as market rates here. Dividing EUR 800 among eight participants gives EUR 100 per person, but that arithmetic does not establish equivalent course content or the superiority of one offer.
What the analysis assumed was included: Preparation, delivery and sharing the training slides. The actual scope and materials need to be agreed in the quotation.
For the current approach and options, explore AI training for teams. The summary of 22 survey responses from my 16 October session provides separate participant feedback from that particular survey.
Trainer
• Name: Anti Soosaar
• Phone: +372 58 200 900
• Email: anti@nulleuri.ee
Website: Anti Soosaar AI training