Cogny: Building Autonomous Marketing Instead of Better Dashboards

For nearly two decades, digital marketing has been built around dashboards. Analytics tells you what happened. Search Console tells you what Google saw. Every platform answers one question while creating ten more, and marketers end up spending more time gathering information than acting on it. A Swedish startup called Cogny believes that era is ending.

Rather than building another analytics dashboard, Cogny is attempting something far more ambitious: creating AI agents capable of understanding business data, identifying opportunities, recommending actions, and eventually executing marketing work on behalf of companies.

It is an ambitious vision, and one that reflects the backgrounds of its founders, Tom Strom and Berner Setterwall, who have spent much of their careers solving increasingly complex marketing problems long before generative AI became mainstream.

Built by Marketers Who Became Engineers

Many AI startups today are founded by software engineers experimenting with language models. Cogny followed a different path.

Tom Strom spent more than twenty years working in advertising technology, optimization, and growth. Before Cogny he co-founded Campanja, an advertising technology company focused on automated media buying and optimization. Campanja eventually raised approximately $5 million in Series A funding before being acquired by 24/7.ai in 2015. Following that acquisition, Strom continued working with growth companies through GrowthHackers, helping organizations build repeatable marketing systems.

His co-founder, Berner Setterwall, approached the same problems from the technical side. As Campanja's CTO, Setterwall spent years building systems capable of processing enormous amounts of advertising data while keeping infrastructure reliable at scale. Rather than focusing solely on interfaces, his work centered on making complex systems dependable enough for enterprise marketing environments.

That combination, a growth strategist and a systems engineer, defines Cogny today.

The Marketing Problem Nobody Talks About

Marketing software has exploded. Businesses now have Google Ads, Meta Ads, LinkedIn Ads, CRM systems, analytics, SEO software, call tracking, email platforms, sales tools, and website analytics.

Ironically, having more data has often made decision making slower.

Instead of asking "What should we do today?" marketing teams ask "Where do we even find the answer?"

Cogny was designed around eliminating that problem. Instead of forcing humans to gather information manually, the platform attempts to answer business questions directly using connected data.

From SQL Copilot to AI Growth Platform

One of the most interesting stories shared by the founders came during their appearance on the Artificial Intelligence After Work podcast. They revealed that Cogny did not begin as the sophisticated AI platform seen today.

Its earliest version was essentially an AI assistant capable of generating SQL queries from natural language. Users could ask questions instead of writing database code.

While innovative, the founders quickly realized the product was too technical. As Berner Setterwall joked during the interview, the first version mostly had one user, himself.

The product evolved because the founders kept asking a different question: what work are marketing consultants actually doing manually that AI could perform?

That shift fundamentally changed the company.

Growth Tickets Instead of Reports

Traditional reporting ends with information. Cogny attempts to end with action.

Instead of producing a dashboard that says organic traffic declined 12%, the platform generates what it calls Growth Tickets, specific recommendations tied to measurable improvements. Examples include pages with declining click-through rates, missing metadata, technical SEO problems, paid advertising inefficiencies, audience opportunities, and landing page improvements.

Rather than asking marketers to discover problems, Cogny surfaces them automatically.

That philosophy appears repeatedly throughout the founders' discussions and published material. The goal isn't simply answering questions. The goal is reducing the time between insight and execution.

AI Doesn't Need Another Dashboard

One recurring theme during the podcast was that dashboards are becoming less important. For decades software expected humans to interpret charts. Modern AI changes that assumption.

Instead of showing impressions, conversions, CTR, bounce rate, and ROAS, the system can simply answer: what should I fix first?

That sounds deceptively simple. Behind the scenes it requires connecting multiple business systems while understanding context across all of them.

That challenge, not chatbot conversations, is where Cogny spends much of its engineering effort.

Why Hallucinations Matter More in Business

The founders also discussed one of AI's biggest challenges. Hallucinations.

When ChatGPT invents a paragraph, it's usually annoying. When AI invents financial analysis, companies can make expensive decisions.

Setterwall explained that early experiments with automated analysis occasionally produced inaccurate conclusions because large language models are probabilistic. The team responded by redesigning parts of the architecture around verification rather than blind generation.

That focus on trust appears repeatedly throughout Cogny's technical writings. Reliability matters more than creativity when businesses depend on numbers.

Marketing Is Becoming Engineering

Perhaps the most fascinating part of the founders' podcast wasn't about AI. It was about organizational change.

Tom Strom argued that marketers increasingly resemble engineers. Campaigns are no longer isolated creative projects. They're systems.

Every landing page, every CRM workflow, every API, every analytics event, and every conversion path must work together.

Success increasingly belongs to organizations capable of orchestrating those systems instead of managing isolated campaigns.

That perspective explains why Cogny isn't positioned merely as marketing software. It aims to become infrastructure.

Beyond SEO: GEO

Search itself is changing. People increasingly ask ChatGPT, Claude, Gemini, and Perplexity questions that once belonged exclusively to Google.

This introduces a new optimization challenge. Not simply Search Engine Optimization. Generative Engine Optimization.

The founders discussed how websites must evolve to become understandable not only by Google's crawler but by language models synthesizing information across thousands of sources.

Structured content. Entity relationships. Readable information. Trust signals. These become increasingly important in an AI-first internet.

AI Agents Instead of AI Assistants

The word "agent" is everywhere today. Unfortunately, it often means little more than a chatbot.

Cogny describes something different. Their vision involves AI systems capable of reading business data, detecting anomalies, creating action plans, coordinating multiple tasks, working across software platforms, and executing workflows.

Rather than answering a single question, agents become participants inside business operations.

This distinction may define the next generation of enterprise AI.

Lessons From Campanja

Experience matters. Long before generative AI exploded, the founders were building automated bidding systems for large advertisers.

Optimization wasn't theoretical. It involved millions of advertising decisions.

That experience appears to influence Cogny's philosophy. Automation should improve measurable outcomes, not simply produce impressive demonstrations.

A Different View of AI

Many companies ask how AI can help their employees. Cogny asks what work shouldn't require employees anymore.

That subtle difference changes product design. Instead of creating assistants waiting for instructions, the platform attempts to identify recurring work suitable for automation.

It's a significantly larger ambition.

The Future of Marketing

Whether Cogny ultimately succeeds remains to be seen. The AI industry changes almost weekly. New foundation models emerge constantly. Major technology companies continue integrating AI into existing products. Competition is intense.

Yet Cogny occupies an interesting position. Rather than competing to build the smartest chatbot, it focuses on connecting business data with business execution.

If that vision proves correct, tomorrow's marketing teams may spend less time building reports and considerably more time supervising autonomous systems capable of running large portions of digital growth.

Final Thoughts

Artificial intelligence is moving beyond content generation. The next frontier is operational intelligence. Cogny represents that transition.

Its founders aren't merely asking whether AI can write advertisements. They're asking whether AI can understand an entire business well enough to improve it continuously. That question may prove far more important.

As Tom Strom and Berner Setterwall suggested during their podcast appearance, the future of marketing isn't simply about faster campaigns. It's about creating organizations where intelligent agents become trusted collaborators, executing repetitive work, surfacing better decisions, and allowing people to focus on creativity, relationships, and strategy.

If that future arrives, companies like Cogny may end up being remembered not for building another AI tool, but for helping redefine what marketing work looks like in the age of autonomous intelligence.