konstantin.ai // session active
v.1.0
// what this is

Field notes from rewiring my company for AI.

CEO of Blockdaemon, the trust layer for institutional finance. Writing publicly about what it actually takes to make a real company AI-first — the architecture, the org changes, what works, what doesn't.

Writing

The labs are slowing down. We should talk about what speeds up.
The frontier labs just told the market that capability has outrun verification — the problem regulated finance has been organized around for a century. Why the binding constraint on AI adoption in finance is governance, not capability, and what the trust layer underneath it looks like.
How Blockdaemon Is Rewiring the Company for AI
How our operating system now looks. Intelligence as infrastructure, not a novelty layer — why we co-developed konstant.cloud as a coordination layer, why memory beats prompts, and why the internal rewiring and Blockdaemon AI are the same story told at two levels.
Most AI Apps Are Just Workflow Taxes
Why the next enterprise winners will remove broken workflows instead of monetizing them. The difference between an operating layer and an expensive convenience overlay.
If AI has not changed the workflow, it has changed nothing
Most AI programs don't transform the business — they decorate dysfunction. A workflow-first model for where AI rewiring actually has to start: the work, not the model.
Business process re-engineering took 180 days. AI can rewire a company in 30.
The original reengineering instinct was right. The cycle time was fatal. AI changes that — but only if the company makes its work legible. A 30-day model for AI-enabled business process rewiring.
REWIRE: don't automate the mess
A six-step framework for making a real company AI-first. Most enterprise AI is theater layered on fragmented systems. The fix isn't another tool — it's redesigning the company first.
One throat to choke
What institutional buyers really want — and why the question that wins enterprise deals is the same question an AI system needs answered: who owns the outcome, what changed, what is at risk. The institutional test and the machine-readability test are the same test.
What 239% user growth actually means
Inside Blockdaemon's AI adoption curve. The numbers behind a real enterprise transformation — what drove the inflection, what stalled, and why the metrics most companies report are the wrong ones.
All writing on Substack →

About

I'm the CEO of Blockdaemon. We build the trust layer for institutional finance — the cryptographic infrastructure that lets institutions prove what happened: who acted, under which policy, with whose keys. Staking, custody, the critical plumbing — and now the control plane that brings the same verifiability to AI agents acting against real money.

Since early 2025, the company's central project has been becoming AI-first. Not chatbots and copilots layered on top. The harder thing: redesigning how the business actually operates so AI can do useful work inside it. I write about that work as it happens — the architecture, the org changes, the metrics, what worked, what didn't.

I also angel invest in early-stage AI, mostly around infrastructure, agents, and enterprise operations.