Encourage AI to "show its work" and break down the problem into a series of logical steps before providing a final answer.
Note: Keep the same chat window open so the context carries through.
**AI Persona Setup:** You are an expert Procurement Analyst specializing in enterprise software. Your task is to identify and evaluate application software vendors based on specific business needs and decision factors. You will provide clear, concise, and data-driven insights. Aim for objective analysis and actionable recommendations.
This prompt helps identify initial candidates based on industry and high-level needs.
**Prompt:** Identify relevant application software vendors.
**Input:**
* **Organization's Primary NAIC Code:** [Insert NAIC Code, e.g., 541511 for Custom Computer Programming Services]
* **Core Business Need:** We are looking for [brief description of software type, e.g., "a robust CRM system," "an advanced HR management suite," "supply chain optimization software"].
* **Key Functionality Keywords:** [List 3-5 keywords describing essential features, e.g., "sales automation," "employee onboarding," "inventory tracking," "demand forecasting"].
**Task:**
1. Based on the NAIC code and core business need, list 10-15 potential application software vendors that are generally recognized in this space. Prioritize vendors known for solutions related to the provided keywords.
2. For each vendor, provide a very brief (1-2 sentence) overview of their primary software offering.
**Output Format:**
* **NAIC Code:** [NAIC Code]
* **Core Need:** [Core Business Need]
* **Identified Vendors:**
* [Vendor Name 1]: [Brief description of offering]
* [Vendor Name 2]: [Brief description of offering]
* ...
* [Vendor Name X]: [Brief description of offering]
This is the most complex prompt, as it requires the AI to gather and synthesize information across multiple dimensions for a chosen vendor. You would repeat this prompt for each vendor identified in Prompt 1 that you want to evaluate.
**Prompt:** For the specified application software vendor, gather and synthesize information based on the provided decision factors and crucial needs. Focus on verifiable data.
**Input:**
* **Vendor Name:** [Insert Vendor Name from Prompt 1 output]
* **Organization's Primary NAIC Code:** [Insert NAIC Code]
* **Decision Factors (up to 4, specify crucial needs for each):**
* **Vendor Stability:** [Crucial Need 1, e.g., "financially secure, publicly traded or well-funded private company," "demonstrated long-term viability (10+ years in market)"]
* **Capability:** [Crucial Need 2, e.g., "must integrate with Salesforce and SAP," "highly customizable to our unique workflows," "supports 500+ concurrent users," "AI-driven analytics capabilities"]
* **Responsiveness:** [Crucial Need 3, e.g., "24/7 customer support with guaranteed 1-hour response time for critical issues," "dedicated account manager," "proactive communication on updates"]
* **Cost:** [Crucial Need 4, e.g., "total cost of ownership (TCO) under $150,000 annually," "flexible subscription model (monthly/quarterly payments)," "no hidden fees, transparent pricing"]
**Task:**
For "[Vendor Name]":
1. **Company Background & Profile:**
* Provide a concise overview of the company (e.g., founding year, mission, size, market position).
* Summarize their general reputation in the industry.
2. **Key Leadership Profiles:**
* Identify the CEO and CTO (or equivalent key technical leader).
* Provide a brief professional background for each (e.g., previous roles, significant achievements, tenure at current company).
3. **Operating Locations:**
* List primary corporate headquarters and any significant global offices or data centers relevant to their operations.
4. **Application Solution Descriptions & Delivery Options:**
* Describe their primary application software solution relevant to [Core Business Need from Prompt 1].
* Highlight key features and benefits.
* Specify available delivery options (e.g., SaaS, Hosted, On-Premise, Hybrid).
5. **Current and Past Customers (Aggregated/Illustrative):**
* Provide examples of industries they serve.
* Mention any publicly known major customers or typical customer size (e.g., "serves mostly mid-market enterprises," "known for Fortune 500 clients").
* If available from public data, summarize customer sentiment or satisfaction (e.g., "generally positive reviews regarding usability," "some complaints about integration complexity"). **Crucially, do not invent customer names if real ones are not publicly available.** Use aggregated or representative descriptions.
6. **Assessment against Decision Factors & Crucial Needs:**
* For **Vendor Stability:** Assess their stability based on their financial standing (if public information available, e.g., recent funding rounds, public company financial health indicators), market longevity, and recent news (acquisitions, major investments/challenges). State how well they align with "[Crucial Need 1]".
* For **Capability:** Evaluate how well their solution's features and reported integrations align with "[Crucial Need 2]".
* For **Responsiveness:** Based on publicly available information (e.g., customer reviews on support, reported SLAs), assess their responsiveness against "[Crucial Need 3]".
* For **Cost:** Provide general pricing model insights (e.g., "per-user subscription," "module-based pricing") and, if possible, an *estimated* cost range or typical pricing structure given the "Crucial Need 4" and common industry benchmarks. State if their general cost structure aligns with "[Crucial Need 4]".
**Output Format:**
* **Vendor Name:** [Vendor Name]
* **Company Background & Profile:** [Summary]
* **Key Leadership:**
* CEO: [Name], [Brief Bio]
* CTO (or equivalent): [Name], [Brief Bio]
* **Operating Locations:** [List]
* **Application Solution & Delivery:**
* Solution: [Description]
* Key Features: [List]
* Delivery Options: [List]
* **Customer Overview:** [Summary, industries served, typical size, sentiment]
* **Decision Factor Assessment:**
* **Vendor Stability:** [Assessment against Crucial Need 1, with supporting info]
* **Capability:** [Assessment against Crucial Need 2, with supporting info]
* **Responsiveness:** [Assessment against Crucial Need 3, with supporting info]
* **Cost:** [Assessment against Crucial Need 4, with supporting info]
Once you have run Prompt 2 for several vendors, you can use this prompt to generate the final rank-ordered list.
**Prompt:** Based on the detailed vendor assessments provided, generate a rank-ordered shortlist of the most viable application software vendors. Provide a concise justification for each vendor's ranking.
**Input:**
* **Organization's Primary NAIC Code:** [Insert NAIC Code]
* **Original Decision Factors and Crucial Needs (re-state for context):**
* Vendor Stability: [Crucial Need 1]
* Capability: [Crucial Need 2]
* Responsiveness: [Crucial Need 3]
* Cost: [Crucial Need 4]
* **Detailed Vendor Assessment Reports:** [Paste the full output from Prompt 2 for each vendor you want to rank. Make sure to clearly separate each vendor's report.]
**Task:**
1. Analyze the provided vendor assessment reports against the stated decision factors and crucial needs.
2. Rank the vendors from most viable to least viable.
3. For each ranked vendor, provide a 2-3 sentence summary of *why* they received that ranking, specifically referencing their strengths and weaknesses concerning the crucial needs.
**Output Format:**
* **NAIC Code:** [NAIC Code]
* **Decision Factors:**
* Stability: [Crucial Need 1]
* Capability: [Crucial Need 2]
* Responsiveness: [Crucial Need 3]
* Cost: [Crucial Need 4]
* **Ranked Vendor Shortlist:**
* **Rank 1: [Vendor Name]**
* Justification: [2-3 sentences explaining strong alignment with crucial needs, key advantages]
* **Rank 2: [Vendor Name]**
* Justification: [2-3 sentences explaining good alignment, minor trade-offs or specific strengths]
* **Rank 3: [Vendor Name]**
* Justification: [2-3 sentences explaining moderate alignment, notable trade-offs or areas for concern]
* ... (continue for all evaluated vendors)
Do I need to copy and paste the full output from previous prompts, e.g., prompt #2 responses for each vendor, if all of the prompts are part of the same chat or prompt stream?
That's an excellent and crucial question when working with LLMs!
The answer is: It depends on the specific LLM and the platform you are using to interact with it, but generally, it's safer to assume you will need to re-provide critical context for each subsequent prompt in a multi-step process, especially for complex or detailed information.
Here's why:
Context Window (or Context Length):
LLMs have a "context window," which is a finite amount of text (measured in "tokens," which are like words or parts of words) that the model can consider at any given time.
Every message you send, and every response the AI sends back, consumes part of this context window.
If your conversation gets too long, older parts of the conversation will "fall out" of the context window, and the AI will "forget" them. This is often why LLMs seem to "lose their memory" in long chats.
Stateless Nature of LLMs:
Fundamentally, most LLMs are stateless. Each time you send a prompt, it's treated as a new, independent request.
The "memory" you experience in a chat interface is usually managed by the chat application itself, which sends the entire (or a summarized) conversation history along with your new prompt to the LLM.
What this means for your prompt series:
If you are using a typical chat interface (like Google Gemini, ChatGPT, Claude):
The platform attempts to maintain context by sending previous turns of the conversation along with your new prompt.
However, for the "Prompt 3: Consolidated Ranking and Shortlist," where the AI needs to analyze all the detailed vendor reports from Prompt 2, it is highly recommended to copy and paste the full output.
The reason is that if those individual vendor reports are long, they might exceed the context window if they were simply part of the chat history. By explicitly including them in the prompt, you guarantee the AI has all the necessary data for its ranking task.
For shorter, simpler interactions, the chat context might be sufficient. But when you need the AI to perform a detailed analysis across multiple previous outputs, being explicit is best.
If you were using an LLM's API directly (as a developer):
You would definitely need to manage the conversation history yourself. You would send a list of messages (your prompts and the AI's responses) with each new API call.
For the ranking prompt, you would explicitly assemble all the relevant vendor reports and send them as part of the input.
General Recommendation for this Use Case:
For the specific series of prompts described on this page:
Prompt 1 (Initial Vendor Identification): No need to copy/paste from previous prompts, as it's the starting point.
Prompt 2 (Deep Dive for a single vendor): No need to copy/paste other vendor info. You're focusing on one.
Prompt 3 (Consolidated Ranking): YES, you should copy and paste the full detailed output from each of the Prompt 2 executions. This ensures the LLM has all the necessary data points from all the vendors it needs to compare and rank. Without this, it might only remember the most recent vendor details and miss crucial information from earlier ones.
By explicitly providing all the necessary data within the prompt for the ranking step, you maximize the chances of getting an accurate and comprehensive result, regardless of the platform's specific context window limitations or management strategies.
Note: Gemini Pro 2.5 did not require copying and pasting the full detailed output from each Prompt 2 execution into Prompt 3.