Build agents visually
Assemble agents from reusable blocks, without writing code.
All work
AI agents, without writing code. A concept I led, down to the whitepaper.
Agentic Tower was a concept for a no-code platform where people build AI agents, connect them into workflows, and publish or trade them in a marketplace. I spearheaded it at INFINITE8, wrote the product whitepaper and shaped its consumer and business direction.
The problem.
AI agents were becoming powerful, but building them still required engineering expertise. Agentic Tower explored a way for domain teams to assemble and configure useful agents themselves, while giving those agents access to the tools and data needed to do real work.
It also explored how those agents could be packaged, shared and distributed, instead of staying isolated internal tools.
What people could do.
One platform for building, running and trading AI agents, for individuals and for businesses.
Configure capabilities
Give agents instructions, tools and access to relevant data.
Orchestrate multi-agent workflows
Connect specialised agents so work moves between them as part of a larger workflow.
Share and trade
Publish agents to a marketplace, with consumer and business directions.
Architecture.
Agentic Tower was designed around modular AI agents: assembled from reusable capabilities, given access to tools and data, and orchestrated to complete multi-step workflows.
Agent builderNo-code assembly
Agent marketplaceShare and trade agents
ToolsActions an agent can take
DataWhat an agent can read
InstructionsWhat an agent is for
Agent runtimeRuns each agent
OrchestratorChains agents into multi-agent workflows
What I owned.
The concept from first idea to written strategy.
- Product concept
- Whitepaper
- Consumer and business direction
- Roadmap and priorities
Why it mattered.
Agentic Tower was an early exploration of what an AI-native product could look like beyond a chatbot. It made me think about where agents should make decisions, where people should stay in control, and how individual agents become reusable parts of larger workflows.
It became an important part of my move from building traditional digital products to treating AI as product architecture.