Case Study of Call Transcripts

Recently, we received an email from an investor who was concerned about artificial intelligence (AI) disrupting the investing field. The investor feared that AI would narrow down the mispricing in the market, thereby reducing the alpha (outsized gains) for long-term investors. The investor acknowledged that even though long-term investing depends more on psychological aspects than technical aspects; however, looking at the pace of advancements in AI field, the day may not be far away when AI matches human psychological intelligence.

In our reply, we explained that investors should see AI not as a competitor, but as a helpful tool that can expand what they’re able to do. Currently, there are limitations to what an individual investor can achieve with her single human brain and its limited processing power.

However, AI can help an investor by becoming an extension of her analytical abilities and helping her process significantly larger amounts of information than she can otherwise read.

Why you should use AI/ChatGPT in Stock Analysis

In this regard, in the current article, we take an example of using AI, i.e. large language models (LLMs) like ChatGPT, to quickly process investing documents like conference call transcripts.

Companies hold conference calls with investors usually every quarter to discuss their financial results and on other important developments/occasions to explain their points of view. These calls are usually one-hour-long discussions, and their text transcripts are usually 15-25 pages long.

On average, listening to a conference call recording with rewinding on important parts and then taking notes takes about 2 hours. Reading the text transcript of the same conference call and writing down notes takes similar time; however, the benefit of text transcripts is that taking notes becomes much easier, as we have to simply copy important information from the PDF file of the transcript to a Word document.

So, on average, it takes about 2 hours to process a conference call transcript.

Imagine an investor analysing a company, which has a history of conducting regular conference calls every quarter for the last 15 years i.e. a total of 60 (=15*4) conference calls. If she takes 2 hours for analysis of each conference call, then theoretically, 60 conference calls will take about 120 hours (=2*60).

Assuming 8 hours of non-stop work every day, it will take about 15 days (=120/8) to finish analysing these conference calls. However, an investor would believe that doing this monotonous task for 15 days is psychologically burdensome; therefore, the pace of reading will drop over time, and the investor would lose interest in continuing with the task.

As a result, it becomes important for the investor to reduce her effort to keep herself motivated to finish the analysis of the company without losing interest and abandoning the task midway.

She can do so either by selecting/shortlisting the conference calls that she would focus on. Say, to read the last 4 quarters’ conference calls in-depth and only focus on historical conference calls of periods with significant developments like a sudden deterioration in performance, a corporate development like a merger etc.

Or she can take the help of AI tools like ChatGPT to summarise the discussion of any conference call and tell her key points. With their immense processing power, AI tools/LLMs like ChatGPT can read documents like conference call transcripts very quickly and present you with the key information. Moreover, they can keep doing it without stopping for innumerable documents one after another, which is not possible for any human being.

However, working with LLMs like ChatGPT requires an investor to keep in mind some key aspects.

Learning how to ask the right questions to ChatGPT

First, what ChatGPT will answer depends a lot on how we frame our question. Asking questions to LLMs like ChatGPT to get the desired output has grown into a specialised field called “Prompt Engineering”. Prompt Engineering focuses on how LLMs think & work and then guides investors on how they should ask their questions to LLMs so that they get the best output.

One of the best resources to learn “Prompt Engineering” is the paper, “A Prompt Pattern Catalog to Enhance Prompt Engineering with ChatGPT” written by Jules White & others (Click here to read).

This paper guides an investor to use multiple approaches to frame her questions to LLMs to get the desired output. These approaches include Persona, Visualization Generator, Recipe, Template, Question Refinement, Alternative Approaches etc.

However, if an investor does not wish to spend time reading such formal documents, then she should simply let LLM tools like ChatGPT teach her how she can best use them.

For example, let us use ChatGPT to help us use it in reading/summarizing conference call transcripts.

To proceed, an investor should open ChatGPT at the following link: https://chatgpt.com/ and log in. Thereafter, in the chat box highlighting “Ask anything” write a question asking ChatGPT to suggest the best prompt that one should write to ChatGPT to extract maximum information from a conference call transcript.

Below is a sample example of one such prompt/query that we put to ChatGPT to make it help us in this analysis.

I am a stock investor who follows a fundamental investing approach. I tend to understand the business of any company in-depth while making an investment decision.

I study conference call transcripts of companies where management discusses their performance with analysts.

I want ChatGPT to read the transcript and give me all the major learnings in a well-classified manner in bullet points without missing out on any critical information.

Please prepare a prompt for ChatGPT that I can write along with uploading the transcript file, which can produce a good output.

Please make it a reusable template.

Please note that the more context you give to ChatGPT while asking any query, the better the output will be.

To the above query, ChatGPT produced the following prompt, which can be used straightaway by uploading the PDF file of a conference call transcript.

ChatGPT Prompt Template: Conference Call Transcript Summary for Fundamental Investors

I am a long-term stock investor who follows a fundamental investing approach. I study businesses in-depth before making any investment decisions.

I have uploaded a company’s quarterly conference call transcript where the management discusses financial performance, business operations, outlook, risks, and answers analysts’ questions.

Please carefully read the transcript and summarize the key takeaways in a clear, structured, and comprehensive format. Do not miss any important points.

Organize your output under the following sections:

1. Financial Performance Highlights

• Revenues, profit/loss, margins, cash flow, cost structure, etc.

2. Business Segment Updates

• Performance by segment (products/services/regions), including notable growth or decline

3. Strategic Initiatives & Growth Plans

• Capex, expansion, new launches, M&A, R&D, digitization, sustainability efforts, etc.

4. Management Commentary on Industry & Macro Trends

• Sector outlook, demand/supply, regulatory impacts, inflation, currency, etc.

5. Key Challenges & Risks Highlighted

• Headwinds, uncertainties, regulatory concerns, raw material issues, etc.

6. Analyst Q&A – Noteworthy Questions & Management Responses

• Important questions, direct answers, evasive responses, tone of management

7. Any Forward Guidance or Outlook Provided

• Quantitative or qualitative guidance, near-term expectations, confidence level

8. My Notes (summary of investor-relevant insights, optional)

• Overall impression of business performance and management credibility

Use bullet points for each section. Maintain objectivity and include all major facts and numbers shared in the call.

If there are recurring themes or concerns, highlight them. If management avoids answering a question directly, note that too.

I will be using your summary to assess business quality and management credibility, so be as thorough and accurate as possible.

Now, we have a ready prompt that we can copy and paste into the “Ask anything” chat box of ChatGPT in a new chat and upload the PDF file of the conference call transcript of any company.

However, before putting this prompt to use, we wanted ChatGPT to use another of its useful features called “Deep Research”, which enables ChatGPT to search numerous public sources to extract relevant information and present it in its output to make it more valuable.

So, we asked ChatGPT to add features of the Deep Research tool by simply asking it:

Can you use chatGPT’s deep research tool in this process?

As a result, ChatGPT revised its earlier prompt and added the following section to enable it to use Deep Research in its analysis of conference call transcripts:

Use ChatGPT’s Deep Research to provide external verification or context for important claims or data points made in the call. For example:

• If the company claims market leadership or industry growth, check recent sector data or competitors.
• If they cite macroeconomic or regulatory factors, verify with publicly available data.
• If they mention new product launches, see how competitors are responding or if any media has covered it.

For each section, clearly separate “Company’s Words” vs. “External Context” so I can distinguish between what was said in the transcript and what has been validated or expanded through research.

Be detailed and objective. Highlight any discrepancies, overstatements, or red flags found during validation.

I will use this to assess not only business performance but also the credibility of management and the external reality of their narrative.

Ready ChatGPT Prompt to Analyse Conference Call Transcripts

As a result of asking the above two questions, we now have a good prompt that we can use as a template for analysing any conference call transcript. The complete prompt is given below, which you can also use in your analysis. You may download it in Google Docs/MS Word format from here.

ChatGPT Prompt Template: Conference Call Transcript Summary for Fundamental Investors

I am a long-term stock investor who follows a fundamental investing approach. I study businesses in-depth before making any investment decisions.

I have uploaded a company’s quarterly conference call transcript where the management discusses financial performance, business operations, outlook, risks, and answers analysts’ questions.

Please carefully read the transcript and summarize the key takeaways in a clear, structured, and comprehensive format. Do not miss any important points.

Organize your output under the following sections:

1. Financial Performance Highlights

• Revenues, profit/loss, margins, cash flow, cost structure, etc.

2. Business Segment Updates

• Performance by segment (products/services/regions), including notable growth or decline

3. Strategic Initiatives & Growth Plans

• Capex, expansion, new launches, M&A, R&D, digitization, sustainability efforts, etc.

4. Management Commentary on Industry & Macro Trends

• Sector outlook, demand/supply, regulatory impacts, inflation, currency, etc.

5. Key Challenges & Risks Highlighted

• Headwinds, uncertainties, regulatory concerns, raw material issues, etc.

6. Analyst Q&A – Noteworthy Questions & Management Responses

• Important questions, direct answers, evasive responses, tone of management

7. Any Forward Guidance or Outlook Provided

• Quantitative or qualitative guidance, near-term expectations, confidence level

8. My Notes (summary of investor-relevant insights, optional)

• Overall impression of business performance and management credibility

Use bullet points for each section. Maintain objectivity and include all major facts and numbers shared in the call.

If there are recurring themes or concerns, highlight them. If management avoids answering a question directly, note that too.

Use ChatGPT’s Deep Research to provide external verification or context for important claims or data points made in the call. For example:

• If the company claims market leadership or industry growth, check recent sector data or competitors.
• If they cite macroeconomic or regulatory factors, verify with publicly available data.
• If they mention new product launches, see how competitors are responding or if any media has covered it.

For each section, clearly separate “Company’s Words” vs. “External Context” so I can distinguish between what was said in the transcript and what has been validated or expanded through research.

Be detailed and objective. Highlight any discrepancies, overstatements, or red flags found during validation.

I will use this to assess not only business performance but also the credibility of management and the external reality of their narrative.

Analysis of Conference Call Transcript of Deepak Nitrite Ltd using ChatGPT

Let us now use this prompt in ChatGPT on the latest available conference call of Deepak Nitrite, which took place on February 21, 2025, and contains 23 pages (click here).

On uploading the conference call transcript and asking the above prompt, ChatGPT within a few seconds provides the below output highlighting key points from the conference call as well as cross-verifying them from other sources like performance of peers (Aarti Industries, Navin Fluorine etc) and other public sources like industry outlooks from Corteva, Fortune Business Insights and disclosure of UPL Ltd.

ChatGPT highlighted that the performance of Deepak Nitrite is poorer than Aarti Industries and Navin Fluorine, indicating higher exposure to input costs and the impact of the shutdown of its plant.

1) Financial Performance Highlights

Company’s Words:

• Revenue for Q3 FY25: ₹1,924 crore (↓5% YoY, ↓6% QoQ).
• EBITDA: ₹190 crore (↓40% YoY from ₹318 crore).
• EBITDA Margin: 9.9%.
• PAT: ₹98 crore; PBT: ₹135 crore.
• Export Revenue: ₹279 crore (15% of total); Domestic Revenue: ₹1,645 crore (85%).
• Advanced Intermediates: Revenue ₹552 crore (↓18% YoY); EBIT margin: 3%.
• Phenolics: Revenue ₹1,366 crore (↑ marginally YoY, ↓5% QoQ); EBIT margin: 9%.
• Balance Sheet: Net worth at ₹5,222 crore (consolidated), with ample headroom for funding future capex.

External Context:

• Compared to peers like Aarti Industries and Navin Fluorine, Deepak Nitrite’s YoY decline in EBITDA was steeper, indicating higher exposure to input cost pressures and plant shutdown impacts.
• The Indian chemicals sector has seen modest recovery signs by Q1 2025 as per ICIS data, especially in phenol demand.

2) Business Segment Updates

Company’s Words:

• Advanced Intermediates: Hit by global destocking and weak agrochemical demand. Recovery signs observed late Q3.
• Dyes & Pigment Intermediates: Stable demand, improved market share, margin pressure from raw materials.
• Phenolics: Production impacted by annual maintenance; demand strong; India nearly self-reliant in phenol.
• Domestic demand improving; export dispatches resumed in December to EU, non-EU.

External Context:

• Agrochemical demand in CY2025 is expected to rebound globally with 5–6% growth (source: industry outlooks from Corteva and UPL).
• Dyes segment supported by pending anti-dumping probes in India.

3) Strategic Initiatives & Growth Plans

Company’s Words:

• Projects:
  • Nitric acid complex commissioning moved to Q1 FY26; expected to yield ₹70–80 crore in annual margins.
  • MIBK/MIBC and acetophenone plants delayed to H1 FY26.
  • New R&D center near Vadodara to be completed by Q4.
  • Polycarbonate resin project expected by FY28 – targets import substitution.
• Capex Strategy: Fully integrated assets, localized supply chains, storage infra (e.g., for ammonia).
• R&D Focus: Expanding into specialty chemicals and new solvent applications.

External Context:

• The polycarbonate market is projected to grow at ~5% CAGR (Fortune Business Insights), with strong India demand due to EVs and electronics.
• Deepak’s planned backward integrations (e.g., ammonia storage) align with global best practices for margin protection in volatile raw material markets.

4) Management Commentary on Industry & Macro Trends

Company’s Words:

• Raw material costs remain stubborn (e.g., toluene, xylene, caustic lye).
• Forex volatility influenced import behavior.
• Indian demand robust across phenol end-uses: pharma, auto, furniture.
• Non-integrated global phenol capacities are operating at borderline profitability, especially in SE Asia and Europe.

External Context:

• India’s chemical industry is benefiting from China+1 supply shifts.
• Global benzene/propylene prices have been volatile; however, demand for phenol-based products in India remains firm.

5) Key Challenges & Risks Highlighted

Company’s Words:

• Delays in project commissioning (e.g., nitric acid, MIBK/MIBC).
• Agrochemical segment faced weak demand and capacity idling.
• Margin erosion due to high raw material costs and plant shutdowns.
• DASDA (dye intermediate) hit by dumping; Government has initiated investigation.
• Operating deleverage and unabsorbed project costs affected profitability.

External Context:

• Rising global scrutiny on Chinese exports has led India to start anti-dumping probes, aligning with the industry’s protectionist trend.

6) Analyst Q&A – Noteworthy Questions & Management Responses

Key Questions:

• Impact of shutdown on phenolics EBIT.
• Confidence in recouping Q3 volume/margin loss in Q4.
• Clarification on ammonia sourcing and nitric acid capex ROI.
• Strategic rationale for acetophenone and fluorination investments.
• Concerns on project execution delays and ROCE implications.

Management Responses:

• Yes, margins were hit due to a combination of lower volumes, spread compression, and shutdown costs.
• Confident of Q4 improvement; normalization expected between Q1–Q2 FY26.
• Building ammonia storage to de-risk supply volatility.
• Capex delays driven by engineering enhancements, not market weakness.
• Return on capital to improve as new assets mature; emphasized long-term integration strategy.

Tone: Candid but cautious; acknowledged issues without evasion; showed confidence in medium-term recovery.

7) Any Forward Guidance or Outlook Provided

Company’s Words:

• Q4 will be “meaningfully better” than Q3.
• Normalization expected by Q1–Q2 FY26.
• Phenol segment to stabilize by March with reduced import impact.
• 15–20% volume recovery anticipated in Agro segment over medium term.
• All polymer chain assets (e.g., polycarbonate) to be commissioned by FY28.

External Context:

• Sector-wide expectations also point to stabilization in H1 FY26.
• Deepak’s forward integration into higher-value segments aligns with margin expansion trends in specialty chemicals.

8) My Notes (Investor-Relevant Insights)

• Business Performance: Q3 was a low point, but management has outlined clear recovery levers. Asset utilization and cost optimization are key near-term triggers.
• Management Credibility: Honest about execution delays, detailed on project status. Strategic clarity on long-term integration is reassuring.
• Red Flags: Margins compressed severely; execution slippage needs monitoring. Any further delays could hurt return metrics.
• Positive Signs: Timely investments in capacity, sustainability, and R&D position Deepak for long-term leadership. Demand trends in India remain supportive.

From the above summary, an investor can quickly gain insights into the performance of Deepak Nitrite in Q3-FY2025 that its revenue as well as profits declined due to plant shutdowns, weak agrochemical demand, and high raw material costs. Its Phenolics business was affected by imports and shutdowns. Management expects recovery from Q4, normalisation by Q1–Q2 FY26, and margin gains from commissioning of delayed projects like nitric acid and MIBK/MIBC. The company plans to continue with its strategic capex in polycarbonates and solvents for FY28.

Along with the summary of the discussion, ChatGPT has also provided her with a validation of what management has communicated to investors. In case there is a discrepancy between what management states and what is happening in the outside world, then it can be an indication for the investor to be cautious and increase her due diligence.

Therefore, with the help of ChatGPT, instead of spending 2 hours to read the whole transcript, an investor can get the essence of the transcript within 5-10 minutes. Moreover, if in the above summary, an investor finds any point that she wishes to read further, then she can go back to the PDF transcript file and search for the keyword and read more about it.

In this manner, an investor can analyse a lot of transcripts and other investment documents in quick succession without losing interest in the long-drawn process of fundamental investment analysis.

AI/LLM tools like ChatGPT can be a really good help for any investor to process a large amount of information quickly and improve her productivity.

Nevertheless, while using such tools, an investor should keep the following things in mind:

1. Their output is never the same on two occasions. Using the same prompt again will produce a slightly different output. So, the answer will always be different and unique.
2. Many times, ChatGPT/LLMs hallucinate, i.e. they may tell you things that are not even there in the main document. For an example of hallucination while analysing Gensol Engineering, click here. So, please be cautious while relying solely on the output of ChatGPT. However, despite a low probability of such hallucination, such tools remain a useful resource.
3. AI and LLMs are a very rapidly evolving field with breakthroughs happening almost every day. Therefore, even if currently, you find that they do not meet your expectations, still, you should become familiar with them and learn how to use them because if not today, in the near future, the technology will overcome its current shortcomings and will improve its outputs that you might find more useful.

Like you, we are also in the process of learning more about AI/LLMs/ChatGPT and on our journey, plan to share our learnings with other investors.

You may freely use the above prompt in ChatGPT or other LLMs to analyse conference call transcripts of any company (Google Docs link). Feel free to make any tweaks, improvise it and use it in any manner you find useful.

Let us know whether you use AI/LLM tools like ChatGPT in your analysis. If yes, then in the comments section below, please share your learnings, your prompts or anything that you believe can be useful to us and to other investors.

Let us all grow together.

All the best for your investing journey!

Regards,
Dr Vijay Malik

P.S.

Disclaimer

Registration status with SEBI:

I am registered with SEBI as a research analyst.

Details of financial interest in the Subject Company:

I do not own stocks of the companies mentioned above in my portfolio at the date of writing this article.