Bear in mind when AI was nearly chatbots and picture era? These days are behind us. We’re coming into the period of agentic AI – the place AI methods don’t simply reply, they take initiative. And for the companies trade, this can be a game-changer.
I’ve written about Agentic AI on VC Cafe in Could 2024 and particularly about AI brokers because the evolution of RPA (robotic course of automation) in September 2024. The speedy developments within the AI agent house have meant that there’s higher infrastructure enabling founders to construct AI-native startups by incorporating current foundational fashions and open-source applied sciences, and the speedy tempo of funding into this house has additionally supplied entrepreneurs extra nuance as to the place are the massive alternatives for startups within the agentic AI sector.
On this submit, I collated among the newest analysis printed on Agentic AI, with the purpose to share information, primarily with founders, about what could be constructed on this house. However first, a refresher on what’s Agentic AI, the way it’s totally different and why VCs are unexpectedly so enthusiastic about it.
What’s Totally different About Agentic AI?
Agentic AI refers to AI methods that may autonomously handle and execute complicated workflows, involving a number of steps and clever decision-making. Not like conventional AI functions, agentic AI methods can assess conditions, make selections, and take actions to attain targets with out fixed human intervention.
Consider conventional AI as a strong however passive device, like a calculator. Now think about a proactive assistant that not solely crunches numbers however manages your total monetary planning course of. That’s agentic AI – methods that may independently execute complicated workflows and make selections.
Actual-world instance: Sierra’s AI brokers can now deal with total customer support conversations, making selections about refunds, escalations, and follow-ups with out fixed human oversight. This isn’t simply automation; it’s autonomous problem-solving.
VCs see Agentic AI as $4 Trillion Alternative
“A multi-trillion-dollar alternative to reimagine software program, companies, and work”
Perception Companions
A brand new report by Battery Ventures is asking it the Coming AI Market “Supercycle”. The concept is that AI has the capability to disrupt software program, companies, and labor markets throughout nearly each trade. If you happen to consider the $2 trillion greenback spend on companies, $1.5 trillion greenback spend on labor, $300 billion spend on SaaS and $380 billion spend yearly on infrastructure, AI is probably growing the TAM of software program to $4 Trillion.
The AI Brokers market is projected to develop from $5.1 billion in 2024 to $47.1 billion by 2030, representing a CAGR of 44.8%. Because of this, VCs aren’t simply excited – they’re recalibrating their total funding thesis round agentic AI. The pie for software program companies has simply grown dramatically, and sensible cash is betting that agentic AI will create extra unicorns than cell or cloud computing did.
However there’s a catch – the rise of agentic AI will probably result in elevated startup formation and enterprise capital exercise, making a extra complicated and aggressive market. The returns will probably be extra concentrated. The stakes are increased, and the technical limitations to entry are substantial.
Right here’s the place the cash is flowing in Agentic AI
October 2024 broke a document in international enterprise capital deployed, $32 billion in a single month, with AI attracting $12.2B, or 38% of the overall funding. Right here is the place the cash is flowing for AI Agent startups:
- AI-First Purposes/ Vertical AI: As a substitute of bolting AI onto current software program, startups are constructing intelligence-first options. Take Harvey AI in authorized tech – it’s not simply looking paperwork; it’s drafting contracts, suggesting negotiations methods, and managing total case workflows.
- Human-AI Collaboration Platforms: Probably the most promising startups aren’t making an attempt to exchange people – they’re remodeling how we work. GitHub’s Copilot doesn’t simply write code; it pairs with builders to deal with complicated engineering challenges, studying and adapting to every group’s coding type.
- Programs of Brokers: Platforms that permit a number of AI brokers to collaborate, focus on duties, and constantly be taught from one another, mirroring human groups. For instance, Crew AI supplies a framework for managing and deploying multi-agent AI methods, enabling brokers to coordinate, work together, and automate duties effectively.
- AI Brokers for Skilled Companies: AI brokers that automate high-value duties historically finished by professionals in fields like gross sales, customer support, cybersecurity, accounting, and healthcare. A superb instance is Benchmark-backed 11x, which is constructing AI digital staff, or in Insurance coverage is Sixfold, which creates an AI copilot for insurance coverage underwriters, ingesting knowledge from varied sources to make optimum suggestions.
- AI-Powered Design Instruments: Options leveraging AI to streamline and improve design processes in varied disciplines, akin to:
- Video: AI-powered video era, modifying, and avatar creation instruments like Hour One (disclosure: a Remagine Ventures portfolio firm)
- 3D Design: Platforms that simplify 3D asset creation, provide collaborative workflows, and allow no-code prototyping. This additionally contains corporations like Echo3D, a DAM for 3D belongings that permits viewing of heavy 3D fashions within the browser (disclosure: a Remagine Ventures portfolio firm)
- Net Design: AI-powered web site builders with superior customization choices and business-in-the-box performance. Firms akin to Israeli startup Weavy.ai, which caters to skilled designers.
Sapphire Ventures created a devoted panorama for AI powered design instruments. There’s a complete vary of startups on this house, together with open supply ComfyUI, which affords builders a strong design companion.
- Unstructured Knowledge Infrastructure: Applied sciences for extracting, processing, managing, and storing unstructured knowledge, which is crucial for coaching and fine-tuning agentic AI fashions
What This Means for Founders
For startup founders, this implies a brand new alternative to reinvest SaaS by constructing intelligence-first functions, adopting new design patterns that leverage agentic AI’s capabilities.
The playbook for constructing profitable AI corporations is being rewritten. The winners will likely be those that:
- Personal priceless knowledge sources (like Bloomberg does for monetary knowledge)
- Resolve actual labor shortages (significantly in healthcare {and professional} companies)
- Create solely new service classes with AI on the core (suppose AI-powered preventive authorized counsel)
Wanting Forward
We’re nonetheless within the early innings of the agentic AI revolution. The following 18-24 months will likely be vital as early adopters show out use instances and enterprise fashions. For founders and buyers keen to guess massive on this shift, the chance is unprecedented.
However keep in mind: with nice energy comes nice accountability. The businesses that win gained’t simply have the most effective expertise – they’ll be those that tackle moral considerations head-on and construct belief with each customers and regulators.
At Remagine Ventures we’re energetic deploying on this house. Two of our newest investments in 2024 (nonetheless in stealth) are the Agentic AI house: a product supervisor co-pilot (beginning with gaming and client apps) in addition to an agent-to-agent service for the transactional net. We’re excited to share extra about these corporations quickly. Are you constructing within the Agentic AI house? I’d love to listen to about your expertise within the feedback beneath, or attain out to me instantly for suggestions and assist.
I am a former common companion at google ventures, head of Google for Entrepreneurs in Europe and founding head of Campus London, Google’s first bodily hub for startups.
I am additionally the founding father of Techbikers, a non-profit bringing collectively the startup ecosystem on biking challenges in assist of Room to Learn. Since inception in 2012 we have constructed 11 faculties and 50 libraries within the creating world.
