This blog post explores how generative AI (Gen AI) can support sellers in conducting essential checks and verifications to engage with customers safely and responsibly. Customer due diligence is a critical process for evaluating the risks a potential customer may pose to a business, focusing on financial activity, credit history, sanctions, media reputation, and more.

 

Data Gathering

The foundation of due diligence is accurate, comprehensive data collection. For publicly listed companies, audited financial data can be sourced from platforms like Yahoo Finance, Dun & Bradstreet, and Alpha Vantage. Key metrics include total revenue, net income, EBIT, EPS, write-offs, important solvency ratios (debt-to-equity, debt-to-asset), and efficiency indicators (Days Sales Outstanding (DSO), and Days of Payables Outstanding (DPO).

Financial Analysis using Gen AI

After gathering the financial and other relevant data, large language models (LLMs) available via OCI Generative AI can analyze trends and flag financial risks.

Sample Prompt Template

templateAnalysis = “”” You are a CPA employed in Finance function ,tasked with assessing the customer risk.
                    Given the {context} information please perform the below task and provide the response.
                    Please respond only in English.
                    Task
                    —–
                    Analyse the provided financial data.
                    Compare the financial between the quarters and years ,noting any significant  movements  or trends.
                    Analyse the debt to equity , debt to asset, EBITDA, Days of Payables Outstanding(DPO) and Day Sales Outstanding (DSO) with derived        values.
                    Evaluate the implications of these movements for the company’s financial health and operation efficiency.
                    From the perspective of a company selling products or providing services to this company highlight key risks for the seller
                    based on what you have gathered from the financials.
                    Please respond only in English.

 

The LLM generates an initial financial risk summary that sellers can use for further checks.

Adverse Media and News Screening

Gen AI can be used to assess reputational risks by summarizing adverse news. Relevant search results are used as context to generate concise summaries.

Sample Prompt Template

 

templateAdverse = “””  Given the {context} information and based on prior knowledge ,summarize the latest adverse news about the {customerName}.
                         If {context} contains no adverse news for {customerName}, then just relay that No adverse news found.
                         Please respond only in English.
                        “””

 

Search for negative media and news about the specified customer by exploring the internet and online content. The gathered adverse news is used as context for the above prompt template to provide us with a concise response and valuable insight into customer behavior.

 

Sanctions Screening

Sanctions screening is essential to ensure compliance with global regulations. APIs like OpenSanctions can be queried for real-time data, which is analyzed using Gen AI.

Sample Prompt Template

templateSanctions = “””  Given the {context} information and not based on prior knowledge , please summarize the list of sanctions related records for the {customerName}.
                         Include the associated information and additional details as well.
                         If {context} contains No sanctions found for {customerName} just relay the same message.
                         Please respond only in English.
                        “””

 

Customer Risk Assessment

Well-informed companies can better assess the appropriate level of customer due diligence. This evaluation reveals potential risk factors, including the customer’s country of origin, work history, and financial activity trends.

Using prior analysis and search content, Gen AI can generate insights across multiple risk categories:

  • Financial
  • Operational
  • Compliance
  • Political
  • Reputational

 

Sample Prompt Template

 

template =     “””You are helpful AI assistant.Use the following pieces of context to answer the question at the end.
                If you do not know the answer just say you don’t know,DO NOT try to make up the answer.
                If the question is not related to the context ,politely respond that you are tuned to only answer questions related to the context.
                Given the {context} , {news_articles} , {search_content} data and based on prior knowledge, answer the following {question}. 
                Please align the response based on {customerName}.
                Generate a bulleted list that make the risks stand out. Each bullet point should start with key words for the risk, in bold, and then have a verbose description of the risk, in this format:
                **key word**:  detailed description
                Keep the response output format similar across the multiple queries related to customer risks assessment.
                Please respond only in English.
                “””

 

” Provide summary of financial risks including financial stability, creditworthiness, and potential bankruptcy risk   with customer ” + customerName +
” Please respond based on the latest news, articles and  information related to the said customer in last twelve months.”
” Please summarize and provide brief and concise response.”
” In scale of high, medium and low at the end include probability and severity of overall financial risks.”

 

Example Output:

  • Financial Stability: XXX Inc shows revenue growth and reduced debt, though a rising DPO raises concerns.
  • Creditworthiness: Improving metrics suggest better credit standing.
  • Bankruptcy Risk: Trends indicate low bankruptcy risk.

Overall Risk: Probability – Low to Medium | Severity – Low

Review Payment History

Past transactional behavior offers vital clues. Key indicators include:

  • Number of invoices issued
  • Total invoice value
  • Outstanding payments
  • On-time vs. delayed payments

This review helps assess reliability and operational risk, especially for existing customers.

Conclusion

Generative AI transforms customer due diligence by enhancing data collection, automating analysis, and uncovering hidden risks. With well-defined prompt templates and OCI’s enterprise-grade Gen AI capabilities, businesses can assess risks across financial, compliance, and reputational domains more efficiently.

While AI streamlines the process, final decisions should always include human oversight to ensure contextual relevance and accuracy.

For more information, see the following resources: