Entrepreneur Ira Bodnar has revealed that she woke up to find her startup “killed” by a new feature from AI tool Claude. Here’s what she shared on Twitter (now
Her company built an AI agent that automatically managed clients’ Google and Meta ads. In just two months, she gained several hundred paying customers.
She had a 70% close rate. It means she was converting 7 out of every 10 potential customers into paying clients. But, after Claude and Manus launched new connectors for Meta Ads, her close rate dropped sharply to 20%.
Right now, Claude can only analyze Meta Ads and cannot directly make changes or access Google Ads. However, Ira believes it is only a matter of time before it gains those capabilities.
She feels this could make her entire product category outdated. Marketers may soon rely on AI tools to launch and manage complete ad campaigns with simple instructions.
Looking at the bigger go-to-market ecosystem, she believes CRMs and lead databases will remain safe. It’s because AI cannot fully own or collect such large customer data sets.
However, outreach automation tools and related systems may soon decline. Creative work for small businesses may also be directly posted to Meta and Google platforms. Niche tools and complex enterprise systems for large brands, she says, are likely to stay strong.
Future of marketing
The AI entrepreneur thinks companies need to rethink the entire go-to-market, or GTM, stack. According to her, some areas will remain strong.
Customer data platforms like CRMs are safe because AI tools like Claude cannot store a company’s private customer data. Lead database companies may also survive, since AI cannot easily compile massive, verified contact lists.
However, she believes outreach automation tools may soon become useless. In the future, AI agents could directly handle cold emails and domain setup without separate software.
Creative production tools may also struggle. Platforms like Meta and Google could generate ads directly inside their systems.
Niche tools such as server-side tracking may survive. Large enterprise systems handling complex workflows will also remain strong.
According to Bodnar, AI has made building products easier, but distribution has become harder. When everyone can build, getting attention becomes the real challenge. She believes more focus and money will now shift towards distribution.
Future of AI marketing
The entrepreneur has shared bold predictions about the future of marketing in the AI era. She thinks Model Context Protocol (MCP) may become the new App Store. AI tools like Claude choose services directly. Users may not even see other options as AI will decide for them.
Bodnar believes ads inside large language models will become a trillion-dollar opportunity, similar to how Google built ads around websites.
She also predicts agentic commerce, where AI agents compare products only on value, not brand. This could push prices down across industries.
Bodnar says AI-to-AI sales will change enterprise buying: “AI agents will read docs, compare specs, analyze pricing. They don’t watch demos or get charmed.”
She warns that social media may soon be flooded with AI-generated and bot-driven content. However, she feels brand, trust and storytelling will become more important than ever. While AI makes content cheap, real marketing skill still matters.
