The Challenge

Influencer marketing has become a cornerstone strategy for brands aiming to reach their target audiences effectively. Europe, for instance, witnesses hundreds of thousands of sponsored social media posts daily, spanning from long-form videos to 24-hour disappearing Instagram stories. Ensuring that each influencer abides by advertising standards is almost impossible. This leads to overlooked systemic violations and the continuous broadcast of irresponsible advertising online.

Our Solution

Our platform helps SROs, brands, branch organisations and governmental bodies to proactively monitor online ads and detect possible rule violations by leveraging computer vision and natural language processing.

Our solution has three distinct components, working in unison:

Data collection agents that reply on AI-powered processing pipelines for the incoming content such as audio to text, OCR, and domain specific image recognition. Leveraging Go and a producer-consumer design architecture ensures the agents are easy to maintain and adapt to changes in web structures. 
Data storage relying on PostgreSQL, an open-source DBMS known for its active community and regular updates. Its full-text search feature is essential for our application, and extensions like pgvector allow us to explore AI-powered solutions effectively. 
Client-facing interface built on React that enables our users to:

a. Easily find and add influencers
b. Search through influencers’ content via keywords
c. Check whether content complies with advertising rules
d. Detect whether images and videos show children

In addition, our cloud-enabled platform is horizontally scalable and standardised across environments, utilising a custom Terraform setup. Automated deployments for both infrastructure and code are made possible through GitHub Actions. And, our deployment pipelines incorporate code quality assurance measures such as linting, formatting, and testing steps.

The Impact

Proactive Compliance Monitoring

Based on a robust combination of computer vision and natural language processing models that enable SROs to monitor online advertisements and quickly detect potential violations of advertising rules.

Strengthened Brand Reputation

Enabling brands to safeguard and protect brand reputation against unethical practices and regulatory penalties.

Strengthened Regulatory Oversight

Providing insights into evolving advertising trends and allowing governmental bodies to act promptly against instances of non-compliance. Doing so safeguards consumer interests and preserves market integrity.

Spotlight

Working together with our launching partner, Stichting Reclame Code, our solution resulted in significant time savings in content scanning and and increased awareness of influencer marketing practices.

The tool has been used by the following organisations:

SROs we worked with:

// Reklamombudsmannen // ÖRT // IAP // ASA // SRC // RAC // ARPP // ASAI // auto-regulacao publicitaria

Industry Associations we worked with:

// FNLI // STIVA // Neproffarm // Keuringsraad // ZAW //Other WUR

Regulatory bodies:

// KOAG KAG // ACM // Commisariaat van de Media// AFM // Ministry of Finance

"Thanks to DEUS' AI expertise, Stichting Reclame Code is making influencer marketing trustworthy and transparent at unprecedented pace. We now detect violations of advertising rules on a daily basis. Alone, we would have never been able to deploy computer vision, speech, and text recognition, or large language models for our social mission, but close partnership has made Influencer Monitor possible."

Tjomme Dooper

SRC

“With the Influencer Monitor, we found a post that violated the Dutch Advertising Code for Alcoholic Beverages. After notifying the advertiser, the post was removed within 24 hours”

Rogier van Oosterhout

STIVA

conclusion

Leveraging a combination of computer vision and natural language processing models, the Influencer Monitor enables SROs, brands, branch organisations and governmental bodies to easily achieve their core objectives: monitor online advertisements and detect potential violations of advertising rules. As such, it serves as an example of a highly effective and practical application of AI.

Influencer Monitor: From reactive to proactive advertising monitoring, using computer vision and NLP