We built the intelligence infrastructure a founder didn’t have time to run — closing the content gap, aligning output with niche benchmarks, and transferring a reusable AI capability that extended beyond the original brief.
What Is IGGI?
IGGI is a content intelligence system developed by WalkerTrust. It automates the market research cycle by collecting, classifying, and structuring performance data from top creators in a client’s niche. For this engagement, data was sourced from Instagram via Apify — though the architecture supports any platform. Grounded outputs are delivered as strategic mandates via NotebookLM, enriched with proprietary context files that act as domain-specific instructions, and orchestrated end-to-end by n8n.
At a Glance
Challenge
An SMB founder without the bandwidth to run the research cycle required for quality content — leading to output that consistently missed niche trends.
Solution
Deployed IGGI, scraping Instagram’s top performers via Apify, structuring the data through n8n, and grounding AI outputs in NotebookLM — enriched with custom skills and context files.
Results
Content aligned with niche trends; intelligence cycle automated (10.5h/week, validated against our research benchmark); client independently operating NotebookLM for strategic use cases beyond the original brief.
The Challenge: When Content Expectations Exceed Available Capacity
The research cycle required to produce trend-aligned, niche-specific output demands more than 10 hours of focused weekly work. That capacity simply does not exist in the life of a founder.
To fill the gap, the client relied on generic AI tools. The output was consistent — consistently average. Without grounded data, those tools generated content from the statistical mean of the internet, not from the high-performance signals that drive competitive positioning.
How IGGI Works: Automated Market Intelligence in Three Layers
We began with a structured audit of the intelligence cycle, mapping four stages: Surveillance (2.5h/week), Intelligence (5h/week), Synthesis (1h/week), and Drafting (2h/week). Our research benchmark indicates that running this cycle fully requires 10.5 hours per week. The client was allocating roughly half that.
Layer 1 — Data Collection: Apify automated the collection of performance data from verified top-1% Instagram creators in the client’s niche. Instagram was selected based on where the target audience was most active — the same architecture can ingest from other platforms where the strategic fit differs. This replaced 2.5 hours of manual surveillance per week.
Layer 2 — Structuring and Classification: n8n orchestrated the pipeline, classifying and routing that data into a proprietary knowledge base. This eliminated 5 hours of weekly manual data entry.
Layer 3 — Grounded AI Reasoning: We integrated NotebookLM as the AI reasoning layer, grounding all outputs in the curated Instagram dataset. Beyond the scraped performance data, we structured a set of costume reference files — covering brand voice, audience profiles, and niche-specific frameworks — that function as persistent domain context for the AI. These files operate similarly to system instructions in other AI tools: they ensure every output is anchored in the client’s specific positioning, not in generic patterns. This combination makes it structurally difficult for the model to default to average outputs. We also trained the client to operate NotebookLM independently, extending the capability to strategic use cases outside the original content brief.
Results: Validated Against a 10.5-Hour Weekly Research Benchmark
- Content output aligned with niche trend benchmarks following stabilisation of the implementation
- Full intelligence cycle automated, consistent with our 10.5h/week research-validated framework
- Client independently operating NotebookLM for strategic use cases outside the original content scope
- AI outputs grounded in top-1% Instagram performer data and proprietary context files, replacing generic AI generation
- Proprietary knowledge base compounds over time — each data cycle deepens niche intelligence without additional manual input
Before and After:
| Dimension | Before IGGI | After IGGI |
|---|---|---|
| Research source | Manual browsing + generic AI | Automated Instagram top-1% data |
| AI grounding | None — internet average | Performance data + proprietary context files |
| Weekly founder time required | 5–10 hours | Near-zero ongoing input |
| AI output quality | Internet-average content | Grounded, niche-specific mandates |
| Strategic AI capability | Limited to content production | Applied across multiple business functions |
What Comes Next: Closing the Loop With Conversion Data
With the intelligence cycle automated, the next logical layer is measurement. IGGI tells you what content your niche’s top performers are producing. What it does not yet tell you is how your specific audience responds — which formats convert, which topics drive enquiries, and which platforms generate commercial traction.
We are currently scoping the next phase: automated conversion tracking across the client’s website and active platforms. This involves building data pipelines that aggregate performance signals — website events, platform analytics, lead activity — into a unified view the client can interrogate without manual reporting.
This distinction matters. AI is well-suited to accelerating research and structuring intelligence. It is less suited to replacing the human judgement required to interpret conversion data, adjust positioning, and make strategic decisions. The next phase is designed to support that judgement — not to remove it.
Frequently Asked Questions
What tools does IGGI use?
IGGI is built on three integrated layers: Apify for automated Instagram data collection, n8n for workflow orchestration and classification, and NotebookLM for grounded AI reasoning — enriched with proprietary context files. Once configured, the system operates without ongoing manual input.
Is IGGI exclusive to Instagram and Apify?
No. Instagram was selected for this engagement because it matched where the client’s niche was most active. The architecture — built on Apify and n8n — supports data ingestion from a wide range of sources, including TikTok, YouTube, newsletters, and custom datasets. The platform selection is determined during the intelligence audit based on where the relevant signal actually lives.
What are the proprietary context files?
These are structured reference documents — covering brand voice, audience profiles, competitive positioning, and niche frameworks — that are loaded into NotebookLM alongside the scraped performance data. They function like system instructions in other AI tools: they constrain the AI’s outputs to the client’s specific context rather than letting it default to generalised patterns. The combination of live performance data and static context files is what makes the system produce mandates rather than generic content.
Is IGGI a content generation tool?
No. IGGI automates the research and intelligence cycle that precedes content creation. It structures and curates the data that informs strategic decisions. Content generation remains human-guided — IGGI ensures it is grounded in validated, niche-specific intelligence rather than generic AI outputs.
How long does it take to see results?
Implementation follows three phases: audit, deployment, and stabilisation. The intelligence cycle becomes operational within the first phase. Post-stabilisation, results align with our research-validated 10.5h/week benchmark across the full cycle.
Can the system be applied beyond content?
Yes. The NotebookLM capability — particularly the proprietary context files — transfers across strategic functions. Our client independently extended the system to competitive analysis and strategic positioning outside the original content brief.
Who is this engagement designed for?
IGGI is designed for SMB founders and operators who need a structured intelligence system but cannot dedicate 10+ hours per week to manual research. Teams that already have a functioning research operation may be better served by our analytics and automation services.
Work With Us
If your content operation is producing output without clear visibility into what is converting, that is the gap worth closing next. We work with founders and operators who want structured, data-driven answers to that question — not more automation for its own sake.
To discuss where your current setup stands, contact us at pedrorcosta@walkertrust.com or connect on LinkedIn.
Pedro Ribeiro da Costa is a Partner at WalkerTrust, specialising in AI-powered process automation and digital transformation. He has led automation and intelligence initiatives across Supply Chain and Business Intelligence in environments up to €1.2B in revenue.