Curious Catalyst AI Prompt Engineering Session Overview
Introduction
The session began with a warm welcome to participants from various professional backgrounds—engineering consultancy, steel‑industry sales, HR, and a French defence company. After brief introductions, the host explained that the meeting was part of the Curious Catalyst series, an ISO 9001:2015‑certified corporate training house that offers industry‑focused AI training.
About the Presenter
Vertika Dehelvi was introduced as a certified advanced prompt engineer with five years of experience in MEAN‑stack web development. At Curious Catalyst she works across technology, operations, and strategic enablement, and has already automated several internal tasks. The session was positioned as a follow‑up to a previous, more theory‑heavy workshop on AI prompting.
Session Agenda
- Creating good prompts – understanding prompts and prompt engineering.
- Impact of minor changes – how small wording tweaks can drastically alter AI responses.
- Quick, reusable prompts – examples that address common business pain points.
- Elaborated use cases – detailed scenarios across functions.
- Templates & screenshots – ready‑to‑use prompt structures.
- When NOT to use AI – ethical and practical limits.
- Disclaimer – data‑privacy considerations.
Prompt Engineering Basics
- Prompt = Input that conveys your exact thought to the AI; unlike human conversation, AI cannot read minds.
- Prompt engineering is the craft of writing clear, detailed inputs so the AI can produce the desired output.
Types of Prompts
| Type | Description | Example |
|---|---|---|
| Open‑ended | Allows AI to explore freely | “Write a story about a robot.” |
| Direct | Asks a single factual question | “What is the capital of France?” |
| Instructional | Gives explicit task instructions | “Summarize the text in three bullet points.” |
Where AI Can Help in Business
- Automating routine tasks (report drafting, documentation, initial customer replies).
- Predicting market trends, analyzing customer behavior, risk prediction.
- Brainstorming and refining ideas, structuring content, creating new content.
- Optimising operations (e.g., extracting action items from meeting notes).
Building a Good Prompt
A checklist (not all items are mandatory, but inclusion improves results):
- Define AI’s role & audience – e.g., “You are a teacher, I am a student.”
- Provide a clear task and context – specify company size, industry, problem statement.
- Specify output format – numbered list, table, custom layout.
- Give examples – demonstrate the desired response style.
- Chain complex prompts – break a large request into sequential prompts if needed.
- Iterate & refine – adjust wording when the AI returns nonsensical answers.
Illustrative Examples
- Role definition: Adding “consider yourself as a teacher” turned a simple arithmetic answer into a step‑by‑step classroom explanation.
- Task & context specificity: “Brainstorm three employee‑engagement ideas for a company with 50 employees, focusing on morale, teamwork, and budget‑friendliness” produced a tailored list, whereas a vague “brainstorm three ideas” yielded a generic response.
- Output format: Requesting a table of client meeting schedules forced the AI to return a structured table, while no format request led to free‑form text.
Quick Prompt Library
| Function | Sample Prompt | Desired Output |
|---|---|---|
| Marketing | “Write three social‑media posts promoting a leadership webinar in professional, inspiring, and conversational tones.” | Three distinct posts, each matching the specified tone. |
| Sales | “Write a friendly outreach email to a potential B2B client introducing our automation services.” | A concise, friendly email draft. |
| Customer Support | “Summarize this customer complaint into five bullet points, highlighting key issue and urgency.” | Five‑point bullet summary. |
| HR Recruitment | “Compare two resumes and tell me who is a better fit for a front‑end developer role and why.” | Comparative analysis with justification. |
| Finance | “Explain this financial report in simple language for a non‑finance manager.” | Plain‑English explanation. |
| Logistics | “Write a professional email informing a client about a shipment delay.” | Polished email draft. |
These one‑liners can be copied, edited, and screenshot for repeated use.
Detailed Use Cases
- Scheduling Client Meetings
- Prompt: “You are an AI scheduling assistant. Schedule meetings for five clients, respecting working hours 9 am–6 pm, avoiding overlaps, and using low urgency.”
Result: A table with client names, dates, and times spread across the year (low urgency). Changing urgency to “high” compressed meetings into a single day.
Data Insight Generation
- Prompt: “You are a data analyst for ‘Technova’. Identify negative trends, root causes, and recommendations across finance, sales, HR, and operations.”
Result: Structured insights—key findings, root causes (e.g., financial mismanagement), actionable recommendations, and risk assessment.
Brainstorming Digital‑Marketing Ideas
- Prompt: “Act as a top marketing expert. Provide two budget‑friendly ideas for each channel (social media, paid ads, email, marketplace, SEO) for an international home‑decor brand.”
Result: Concise ideas per channel, such as short before‑and‑after reels for social media or keyword optimisation for SEO.
Prospect List Generation
- Prompt includes: target industry, region, company size, decision‑maker roles, and required fields (name, role, company, relevance, challenges, outreach angle, priority).
Output: Table of prospects plus bullet‑point insights; if no contacts exist, the AI returns “ideal prospect types.”
CRM Automation
- Prompt: “You are a CRM intelligent assistant. Convert this meeting transcript into structured CRM documentation (summary, pain points, proposed solution, buying stage, stakeholders, objections, timeline, next steps).”
Output: Ready‑to‑paste bullet points for CRM entry.
Business‑Analysis Prompt
- Prompt: “You are a business analyst AI. Analyze the attached data, highlight risks, suggest opportunities, and provide actionable recommendations in bullet form.”
Output: Insight summary, risk list, recommended actions.
Marketing Strategy Template
- Prompt asks for 3‑5 ideas per channel, with objective, key message, execution approach, and expected outcome.
- Flexible fields (product name, category, features, target audience) can be added or removed.
When Not to Use AI
- Sensitive or confidential data – only use tools officially permitted by your organization.
- Human judgment required – AI cannot replace decisions that need ethical, legal, or nuanced evaluation (e.g., criminal assessment, hiring based on “spark”).
- Garbage‑in, garbage‑out – inaccurate input data leads to misleading results.
- Out‑of‑date knowledge – AI models are trained on static datasets and may lack the latest information.
Ethical & Practical Disclaimer
Sharing proprietary data with an unapproved AI tool can raise legal and ethical issues. Participants were reminded to use only company‑approved platforms for any data‑sensitive tasks.
Tools Mentioned
- ChatGPT (general purpose, limited to eight attachments).
- Zoho Recruit, Zoho CRM (recruitment and CRM).
- Otter AI (note‑taking).
- HubSpot Breeze, Copy.ai (content creation).
- Amazon Kendra, Gamma AI, Draftbit, Streak AI (various automation and presentation needs).
Upcoming Session
The next Curious Catalyst session is scheduled for 18 January, 11:00 am–12:00 pm, titled “AI‑Powered Automation: Moving Beyond Rule‑Based Systems”, led by AI strategy consultant Ankit Kali.
Q&A Highlights
- Scheduling flexibility – specifying months (e.g., “January”) or quarterly cadence in the prompt adjusts the meeting distribution.
- Resume ranking tools – beyond ChatGPT, specialized AI recruitment platforms (e.g., Zoho Recruit) can rank multiple CVs.
- Prompt length concern – participants were advised to use ready‑made templates and screenshots rather than typing long prompts each time.
- Government tender analysis – no specific tool was recommended, but participants were directed to explore AI note‑taking solutions like Otter AI for related tasks.
- Content moderation – discussion on misuse of AI‑generated content (e.g., explicit images) underscored the need for regulation and human oversight.
Closing Remarks
The facilitator reiterated that the AI toolkit, handouts, and session recordings are available via the WhatsApp group, YouTube channel, and LinkedIn links shared in the chat. Participants were encouraged to stay curious, apply the prompt templates, and join the next session on 18 January.
End of article.
The session emphasized that clear, role‑defined prompts dramatically improve AI output quality, while offering ready‑to‑use templates that span marketing, sales, support, HR, finance and logistics. It demonstrated practical business use cases such as scheduling, data insight generation, prospect list creation and CRM automation, showing how AI can streamline routine tasks and support decision‑making. Ethical boundaries were highlighted, warning against using AI for confidential data, legal judgments or when input quality is poor. Participants were provided with resources, tool references and a roadmap for continued learning through upcoming sessions and shared community channels.
Takeaways
- Clear, role‑specific prompts and explicit output formats lead to more accurate and useful AI responses.
- A library of short, reusable prompts can be applied across multiple business functions to automate routine tasks.
- Ethical guidelines stress avoiding AI for confidential data, legal judgments, and situations where input quality is unreliable.
- Practical use cases illustrated how AI can handle scheduling, data analysis, prospect generation, and CRM documentation efficiently.
- Participants were directed to official tools and community channels for ongoing support and future training.
Frequently Asked Questions
Who is Curious Catalyst on YouTube?
Curious Catalyst is a YouTube channel that publishes videos on a range of topics. Browse more summaries from this channel below.
Does this page include the full transcript of the video?
Yes, the full transcript for this video is available on this page. Click 'Show transcript' in the sidebar to read it.
| “What is the capital of France?” | | Instructional | Gives explicit task instructions | “Summarize the text in three bullet points.” | ### Where AI Can Help in Business - Automating routine tasks (report drafting, documentation, initial customer replies). - Predicting market trends, analyzing customer behavior, risk prediction. - Brainstorming and refining ideas, structuring content, creating new content. - Optimising operations (e.g., extracting action items from meeting notes). ### Building
Good Prompt
Helpful resources related to this video
If you want to practice or explore the concepts discussed in the video, these commonly used tools may help.
Links may be affiliate links. We only include resources that are genuinely relevant to the topic.