Agencies Weigh Building Proprietary AI Versus Using Big Tech Tools

As artificial intelligence becomes more central to marketing, agencies are deciding whether to build proprietary AI systems or rely on technology from large external providers.

Jellyfish has embedded several proprietary AI tools, such as Switchboard, Pencil, Now-Next-Soon, Share of Model, Jellyfish Cloud, and Social Agents, into its marketing workflows. According to the company, these tools have improved asset creation speed, content delivery, and campaign performance while reducing budgets and infrastructure costs. Jellyfish also reported that its AI agents contributed to faster campaign launches and a reduction in required headcount.

Monks has created Monks.Flow, a proprietary AI-powered operational backbone that integrates real-time AI agents across marketing processes. The company stated that Monks.Flow brings together capabilities from partners including AWS, Google, and NVIDIA to support creative ideation, production, and media performance. Media.Monks, part of Monks, was named Adweek’s inaugural “AI Agency of the Year.”.

Despite these examples, most mid-market marketers have not adopted custom-built AI tools. Data shows that only 19% of such marketers use employee-made AI, while a larger share use generative AI or features added to existing platforms. Partial AI adoption is more common than full integration among these companies.

What the numbers show

• Jellyfish reported a 55% increase in asset creation speed and 80% faster content delivery using proprietary AI
• Jellyfish stated that AI agents reduced infrastructure expenses by 22% and campaign launch time by 65%
• Only 35% of mid-market companies have fully embedded AI into daily operations, while 52% have partial adoption

Jellyfish also reported that its Pencil tool halved creative production time and scaled testing by five times, while Now-Next-Soon delivered weekly insights ten times faster than previous methods. The company stated that these advancements led to increased revenue in some cases and reduced overall marketing budgets.

In addition to developing proprietary tools, Jellyfish trained over 55,000 marketers worldwide in generative AI and prompt engineering. The company received recognition as “Most Innovative Use of AI by an Agency” at the 2025 Digiday AI Awards.

Organizations that choose to build their own AI agents can achieve differentiation, improved user experiences, and greater control over security and compliance, according to KPMG. However, this approach typically requires more time and resources compared to adopting or adapting existing solutions.

While some agencies are investing in proprietary AI, a larger portion of the market continues to use generative AI or features integrated into established platforms. The data indicates that full operational integration of AI remains limited among mid-market marketing firms.

* This article is based on publicly available information at the time of writing.