Epiminds Raises $6.6 Million For AI Marketing Agents

Stockholm-based startup Epiminds has raised $6.6 million in seed funding led by Lightspeed Venture Partners, with participation from EWOR, Entourage, and several angel investors including the former CMO of Booking.com. Founded earlier this year by Elias Malm, formerly with Google, and Mo Elkhidir, an engineer who led machine learning teams at Spotify and Kry, the company is emerging from stealth with a working product and early revenue from major Nordic and US agencies.

The founders say their platform automates much of what marketing agencies do each day, using multi-agent artificial intelligence to analyze performance data, generate reports, and even execute campaign optimizations across Google, Meta, TikTok, and other ad networks. They describe it as an AI team for marketers. The system’s visible front end is “Lucy,” an AI marketing manager who orchestrates more than twenty specialized agents trained on live campaign data, agency playbooks, and industry benchmarks.

Clients can be onboarded in less than a minute, immediately gaining access to Lucy’s virtual staff of data analysts, creative optimizers, and media buyers. Each agent connects directly to the agency’s clients accounts, pulling and processing live performance metrics such as conversions, impressions, budgets, and audience trends. These agents analyze results and execute approved changes inside client platforms, from adjusting bids to pausing underperforming keywords.

Malm says this system reduces the manual reporting and cross-platform management that dominate agency work. The average performance marketer juggles thousands of data points across dozens of campaigns. The firm’s multi-agent framework processes more than 60 million marketing-specific data points each day. “Humans exhibit biases when making data-driven decisions,” Malm said. “We built Epiminds to take that burden off people so they can focus on creative strategy instead of spreadsheets.”

The product is built on commercial large language models, including those from OpenAI, Anthropic, and Google, with proprietary layers that specialize them for marketing. The company’s long-term plan is to build its own LLM model specialised for marketing. Each client can embed its internal best practices, templates, and checklists into the system, training the agents to behave like in-house staff rather than generic tools.

Epiminds currently works with seventeen agencies managing over 250 brands. Most early customers are midsized agencies firms, but conversations with larger global agencies are underway. Malm says the company’s ideal customer profile is an agency with 200 to 1,000 employees—large enough to need automation, agile enough to move quick and stay ahead of the biggest change marketing has seen in a decade

”We’ve integrated Epiminds into our advertising workflows.” Said John Axelsson, Founder & CEO of Swedish Ad Agency BBO. “It has really started transforming how our specialists work by automating analysis and surfacing deep actionable insights, while we remain in control of strategic decisions. The result is faster optimization, more effective collaboration between human expertise and AI, and ultimately better outcomes for our clients.” Malm and Elkhidir’s business plan was accepted by incubator EWOR (a European Y Combinator) and developed their prototype, signed their first paying agencies, and secured funding in twelve weeks. In addition to mentorship, EWOR provides early-stage founders with up to €500,000 in capital. That combination of speed and discipline caught the attention of Lightspeed, and led the $6.6 million seed round.

Other companies are mining this fertile territory as well. Triple Whale aggregates data across ad channels and surfaces insights through dashboards and AI-driven alerts. Other firms, including Northbeam and Rockerbox, specialize in attribution modeling, while Omneky uses AI to automate creative generation and campaign optimization. Epiminds say none combines insight, orchestration, and execution through autonomous agents that collaborate across reporting, bidding, pacing, and creative tasks in real time.

Lightspeed partner Paul Murphy calls the company’s approach “a direct attack on the structural inefficiency of digital marketing.” He likens the shift to the early SaaS wave, when automation platforms replaced manual reporting and fragmented spreadsheets.

Elkhidir describes the company’s architecture as its core advantage. “Each agent knows what the others are doing,” he said. “They fetch data, analyze results, recommend changes, and execute them as a coordinated team. It’s a living system, not a dashboard.”

Malm frames it more simply: “We’re building the operating system for marketing. Lucy is just the interface, the product you see. The real value is the multi-agent network behind her that learns, adapts, and improves with every campaign.”