The Information Overload That Consumes 15 Hours of Strategic Team Time Every Week
For executive teams, competitive intelligence functions, and strategic marketing departments, staying informed about market developments, competitor activities, and industry trends is not optional — it's a core operational requirement. But the manual process of staying informed is itself a significant operational burden. Someone has to identify the relevant sources, find the specific articles, read them thoroughly enough to extract the key points, synthesise across multiple sources, and communicate the synthesised intelligence to the people who need it. At 10–15 hours weekly per team, this is a substantial and recurring drain on the most expensive talent in the organisation.
The quality problem compounds the time problem. Research quality varies dramatically based on who conducts it, how much time they have, and how familiar they are with the topic area. A thorough researcher with domain expertise produces different output than a generalist under time pressure — but both results look like "the research was done" in a status report. And without systematic knowledge retention, teams repeatedly research the same topics as people move between projects, take leave, or simply forget what was previously found — duplicating effort that could be eliminated entirely by an organisational memory system.
Building the Intelligence System: From Slack Query to Executive Summary in Minutes
GrowwStacks engineered a market intelligence automation built around the natural way executive teams actually request information — a message in Slack asking about a topic — and designed to make that natural request trigger a comprehensive research-and-synthesis workflow that delivers a polished intelligence brief back to the channel within minutes. The system uses Perplexity AI for web research because of its strength in finding and citing current news sources with temporal precision — returning the specific recent articles most relevant to the query rather than broad topic overviews. ChatGPT handles the synthesis layer, transforming Perplexity's research output into the executive summary format that makes intelligence immediately actionable.
The knowledge base layer — a ChatGPT assistant with vector file storage — is the feature that converts a research tool into an organisational intelligence system. Every topic the team researches is stored and retrievable, meaning a second query on a topic the team has previously covered returns an instant answer without triggering another web search. Over time, the knowledge base accumulates the organisation's complete market intelligence history — making the system faster and more efficient as usage grows rather than resetting to zero with each query.
From Slack Message to Intelligence Brief: The Complete Eight-Step Workflow
The system executes across eight automated steps that together cover every query type — text or voice, new topic or previously researched — without any manual steps between the request and the delivered summary. Here's the complete flow:
- Slack query capture: The Make.com watch module monitors the designated Slack channels for new messages in real-time. Both text and audio messages are captured — allowing team members to submit research requests in whichever format is most natural for their workflow. Channel-based organisation means different topics, departments, or projects can have their own intelligence channels with queries and deliveries kept separate and organised.
- Audio transcription — when needed: When a voice message is detected, the system retrieves the audio file from Slack and passes it to OpenAI Whisper for transcription. Whisper converts the voice query to text with high accuracy across different accents, speech rates, and audio quality levels — enabling team members to submit queries verbally during commutes or meetings, or when typing is inconvenient, without any reduction in the quality of the intelligence delivered.
- Knowledge base query: Before triggering any web research, the system queries the ChatGPT assistant's vector file knowledge base using the text of the query (whether originally typed or transcribed from audio). The assistant searches its accumulated research history for content matching the topic — looking for previous research on the same subject, related topics, or specific companies or trends the team has already investigated.
- Intelligent routing decision: The knowledge base query result determines the processing path. If the assistant returns relevant existing research — indicating the topic has been previously covered — the workflow routes to the instant delivery path, retrieving and formatting the stored analysis for immediate Slack delivery without any additional web search. If the assistant finds no relevant existing content, the workflow routes to the fresh research path.
- Perplexity AI fresh research — new topics: For queries not covered in the knowledge base, Perplexity AI receives the query with search parameters configured for recency and relevance. Perplexity searches the web for the latest news articles, market reports, analyst commentary, competitor announcements, and industry publications relevant to the topic — returning a comprehensive set of current sources with citations. The search parameters are tuned during implementation to prioritise high-credibility sources and recent publications over older or lower-quality content.
- ChatGPT executive summarisation: The Perplexity research output is passed to ChatGPT with an executive summary prompt engineered to extract the highest-value intelligence from the raw research findings. ChatGPT produces a structured brief containing: a 3–5 sentence high-level overview of the key development, the most important data points and statistics, identified implications for the business, and specific recommended actions or watch-items. The output is formatted for Slack — clear, scannable, and actionable without requiring the reader to follow up on the source articles.
- Slack intelligence delivery: The formatted summary is posted back to the channel where the original query was submitted — delivering the intelligence directly in the conversation context where the team is already engaged. Message formatting uses Slack's markdown to create clean, readable briefs with clear section headers for the overview, data points, and implications. Source citations from Perplexity's research are included for any team members who want to read the full articles.
- Knowledge base accumulation: The new research and summary are automatically added to the ChatGPT assistant's vector files — storing both the Perplexity research output and the synthesised summary under the relevant topic keywords. Future queries on the same or closely related topics will match against this stored content, delivering an instant response from the knowledge base rather than triggering another web search. The knowledge base grows with every new query, making the system progressively more responsive over time as the organisation's most frequently researched topics accumulate in the retrieval layer.
💡 The knowledge base: the feature that makes the system better with every use: Most AI research tools reset after every query — each request starts from scratch regardless of what the team has previously researched. The ChatGPT vector file knowledge base inverts this dynamic: the system becomes faster and more valuable as the team uses it more. After 30 days of active use, a significant proportion of queries return instant answers from accumulated research rather than triggering new web searches — and after 90 days, recurring monitoring topics (competitor activities, specific market segments, key trends) are almost always covered by the knowledge base, making the system feel like an institutional memory that grows alongside the team's information needs.
What This System Does That Manual Research Workflows Can't
Perplexity AI Web Research
Conducts comprehensive web searches finding the latest news articles, market intelligence, competitor announcements, and industry insights with source citations — automatically and within seconds of a query. Eliminates manual source hunting entirely, providing teams access to current information without research expertise or time investment required.
ChatGPT Knowledge Base
Stores previously researched topics in vector files — building a searchable organisational intelligence repository that grows with every query. Delivers instant answers for recurring topics without redundant web searches, eliminating 85% of duplicate research effort as the knowledge base accumulates the team's complete market monitoring history.
Executive Summary Generation
ChatGPT transforms Perplexity's raw research output into structured executive briefs with key takeaways, critical data points, business implications, and recommended actions — formatted for immediate strategic use without requiring the reader to process the source articles. Intelligence arrives ready to act on, not ready to read.
Voice Query Support
OpenAI Whisper transcribes audio message queries with high accuracy — allowing team members to submit research requests verbally without typing. Particularly valuable for executives on mobile, in transit, or during brief windows between meetings where a voice message is faster and more natural than typing a detailed text query.
Intelligent Knowledge Routing
Every query checks the knowledge base before triggering web research — delivering instant responses from stored intelligence when topics have been previously researched, and routing to fresh Perplexity research only when genuinely new topics are requested. Optimises both response speed and research cost across the full query volume.
Slack Channel Distribution
Delivers formatted intelligence summaries directly to the Slack channel where the query was submitted — keeping market intelligence within the team's existing communication workflow rather than requiring a separate research tool or email distribution chain. Different channels serve different departments or topics with organised, channel-specific intelligence delivery.
The System in Action
.png?updatedAt=1768976911020)
Before vs. After: What Changes When Market Intelligence Delivers Itself
Before: Executive teams and strategists spent 10–15 hours weekly manually identifying news sources, finding relevant articles, reading through lengthy pieces to extract key points, synthesising across multiple sources, and sharing findings with the team through email or Slack — with quality varying significantly based on who did the research and how much time they had. Critical market intelligence was delayed by the manual research process, causing decisions to be made on stale information. The same topics were repeatedly researched as team members forgot or were unaware of previous work. There was no organisational memory that accumulated across individual research efforts.
After: Any team member types or speaks a research query into Slack and receives a comprehensive, well-structured executive brief within minutes — covering the latest developments, key data points, business implications, and recommended actions. Previously researched topics return instant answers from the knowledge base. The organisation builds a growing intelligence repository that makes every subsequent research session faster. Decision-making is no longer bottlenecked by the time required to gather and synthesise information — the intelligence is available before the meeting where it needs to be used.
Implementation: Live in 8 Weeks
- Slack channel configuration: Designated research channels are created or identified — typically organised by topic area (market trends, competitor intelligence, industry insights) or by department (marketing, strategy, business development). Webhook integrations connecting Slack to Make.com are configured and tested. Team query format guidelines are established so members know how to submit requests and what level of specificity produces the best results.
- Audio processing setup: The OpenAI Whisper integration is configured for audio file retrieval from Slack and transcription. Transcription accuracy is tested across different voice types, accents, and audio quality levels. A fallback path is established for transcription failures — defaulting to a "please retype your query" message rather than silently failing. Edge cases (background noise, very short queries, mixed language) are tested and handled appropriately.
- Knowledge base development: The ChatGPT assistant is created with vector file storage configured for the organisation's topic structure. An initial knowledge base is populated with core topics relevant to the team's regular monitoring needs — seeding the system with intelligence on the most frequently researched subjects so it delivers value from day one rather than starting empty. Knowledge retrieval logic is configured to match query intent rather than just exact keywords, enabling related topic matches that extend the knowledge base's utility.
- Perplexity AI integration: The Perplexity AI API is connected to Make.com with search parameters optimised for the team's research profile — source recency filters, domain quality preferences, and topic-specific search prompt templates. Search quality is tested across the range of query types the team typically submits, and prompts are refined until the returned sources are consistently high-quality and directly relevant to the queries.
- Summarisation, routing, and deployment: The ChatGPT summarisation prompts are engineered to produce executive briefs at the right length, depth, and structure for the team's decision-making context. The intelligent routing logic is built and tested with a range of both new and previously-researched topic queries. The knowledge base accumulation step is validated to confirm new research is stored correctly and retrievable on follow-up queries. The complete workflow is tested end-to-end with the team before production deployment, and team training documentation is provided covering query best practices and channel protocols.
The Right Fit — and When It Isn't
This solution delivers maximum value for executive teams monitoring markets, competitive intelligence departments, marketing teams tracking industry trends, strategic planning functions, business development groups, and any organisation where senior staff are currently spending meaningful hours each week on manual news research and synthesis. It's particularly powerful for teams that monitor the same topics repeatedly — competitor activities, regulatory changes, specific market segments — where the knowledge base accumulation effect delivers compounding efficiency gains over time.
One practical note: the system delivers its highest value on publicly available information — news articles, analyst reports, company announcements, industry publications, and web-accessible market data. For intelligence needs that require proprietary data sources, paywalled research, or primary market research (surveys, interviews), this system handles the publicly-available layer while those sources require separate access. Additionally, the executive summary format is optimised for strategic overview and decision support — teams that need granular data analysis or quantitative modelling from research outputs benefit from adding a data extraction layer tailored to their specific analytical requirements, which we scope as an extension during discovery.