Notion, Make, and GPT-4: automate competitive intelligence in 2 hours per week

Competitive intelligence is often the first thing cut in a small business. Without time, method, or budget, 68% of French SMEs have no structured process for monitoring their competitors (Bpifrance, 2024). Yet companies that conduct regular competitive monitoring are 2.3 times more likely to identify a market opportunity before their rivals (Gartner, 2023).

This tutorial shows you how to fill that gap with three accessible tools, one afternoon of configuration, and less than 2 hours of work per week going forward.

Marc's story: an independent accountant in Lyon

Marc runs a small accounting firm in Lyon with two partners. He was spending 3 to 4 hours every week manually checking competitor websites, LinkedIn pages of other firms in the region, and tax news feeds. The result: incomplete intelligence, often stale information, and mounting frustration at the time wasted.

In March 2026, he implemented the Notion + Make + GPT-4 workflow described below. Results after 30 days:

• Monitoring time reduced to 1.5 hours per week
• A Notion dashboard updated every Monday morning at 8am
• One key strategic insight found in the very first week: a competitor launching a "fast-track payroll audit" offer targeting his own SME clients

Total system cost: 42 EUR per month.

Why these three tools?

Make (formerly Integromat) is the orchestrator. It automates information gathering from multiple sources and sends the data to GPT-4 for processing. Its visual, no-code interface makes it accessible without programming knowledge.

GPT-4 analyzes, summarizes, and ranks information by relevance. Where a human would read 50 articles to extract 5 useful ones, GPT-4 does it in seconds with a calibrated prompt.

Notion stores results in a structured dashboard, sends alerts, and maintains a searchable history.

The 5-step workflow

Step 1: Identify your sources

List 5 to 10 sources relevant to your business: the blogs or news pages of your 3 main competitors (via their RSS feeds), 1 or 2 LinkedIn profiles of competitors or visible sector influencers, and 2 or 3 industry newsletters. For a physiotherapist in Bordeaux, this might include competing clinic websites, LinkedIn accounts of visible local practitioners, and the newsletter of their professional federation.

Step 2: Set up Make to aggregate feeds

In Make, create a new scenario with a "Schedule" trigger (weekly, Monday at 7am). Add an "RSS - Watch items" module for each RSS source. For sources without RSS, use the "HTTP - Make a Request" module to retrieve page HTML, then a "Text parser" module to extract titles and summaries.

Step 3: Send data to GPT-4

Connect your collection modules to the "OpenAI - Create a Chat Completion" module. Your prompt should include three elements: your professional context, your target competitors, and what you want to detect.

Example prompt used by Marc:

"You are a competitive analyst specializing in accounting. Here is a selection of articles and posts published this week. Identify the 3 most relevant pieces of information for an accounting firm in Lyon targeting small businesses with fewer than 10 employees. For each item, provide: a 2-sentence summary, the source, relevance level (High / Medium / Low), and the recommended action."

Step 4: Store summaries in Notion

Use the "Notion - Create a database item" module to send each structured summary to your Notion database. Configure these properties: Date, Source, Relevance level, Information type (commercial offer, recruitment, communication, pricing), and Recommended action.

Step 5: Enable alerts

In Notion, create a filtered view for "High relevance" items and activate an email or Slack notification (via native integration). Every Monday morning, Marc receives a competitive summary without having opened a single extra tab.

Concrete ROI after 30 days

| Metric | Before | After | |---|---|---| | Weekly monitoring time | 3.5 hours | 1.5 hours | | Monthly cost | 0 EUR (but 14h of work) | 42 EUR | | Missed information | Frequent | Near zero | | Structured history | No | Yes (Notion) |

According to McKinsey (2023), companies that integrate AI into their competitive intelligence process reduce analysis time by an average of 60%. Marc falls within that range, saving an estimated 8 hours per month on low-value tasks.

Going further with the IMPACT method

This workflow covers the "Monitoring" dimension of the IMPACT methodology: maintaining a clear view of your competitive environment to make better operational decisions. Once this first automation is in place, the next step is connecting your intelligence to your commercial strategy and acquisition processes.

If you want a full diagnosis of your processes and a personalized AI roadmap, the TransformAudit is designed for you: 2 days of audit, a map of your quick wins, and a 90-day action plan for 1,490 EUR. Get in touch to learn more.