Baizhan Kanafin, Founder & CEO of Documentolog and d8n.ai
FOUNDER-LED · BUILD IN PUBLIC

Baizhan Kanafin

Baizhan Kanafin is the founder and CEO of Documentolog. He writes about rebuilding the company around AI agents: processes, mistakes, management decisions, and measurable outcomes.

PERSONAL CONTENT TRACK

Latest articles

Only long-form articles and expanded pieces are published here. Short posts stay on social media.

AI agents

One Chat Instead of Ten Tabs: 185 Employees and 90% Retention in Two Months

How a digital assistant became the whole company's daily tool: 10,500 dialogues, calls up 2.3x with zero mandates — and why the long tail of routine decides.

AI transformation

Who Actually Implements AI in a Company (Spoiler: Not the IT Department)

Why the fate of an AI project is decided not by the model or budget but by the AI champion — the person who owns the process. Our experience and our clients'.

AI agents

The Minutes Are Born from the Audio: 400 Meetings Later

Two months of AI-run meetings: 400+ recordings, minutes in minutes, and the rule 'an action item = one owner + one deadline.' An honest breakdown.

AI agents

Half a Minute Instead of Half an Hour: How Our Agent Processed 220 Incoming Documents in a Month

One month of the d8n Registry Clerk on our own registry: a 33-second median per document, 97% accuracy on precedents, and what stays with humans.

AI economics

Saved hours are a weak metric if the process doesn't change

Saved hours are easy to count and just as easy to discount. The metrics that show the real effect of AI agents when the process itself changes.

AI economics

Where the effect of AI actually lives: time, errors, SLA, control, decision speed

Five levers where AI agents actually deliver: process time, errors, SLA, cycle time, and manageability — and how to build a CFO case without invented numbers.

AI economics

A business case for the CFO: how to sell an AI project to someone who doesn't buy revolutions

Six blocks after which a CFO calmly signs off on an AI project — and three things you must never put in a business case.

AI agents

Why the knowledge base matters more than the prompt

A whole industry has grown around prompts. But in enterprise AI what decides the outcome is not how you asked — it is what the system knows about your company.

security

AI got cheaper for attackers and more expensive for everyone else: what that means for your documents

IBM's new report: one in four malicious breaches is AI-enabled, at $6M each. Why regulated business cannot simply move documents into cloud AI.

AI transformation

How to Tell Whether Your AI Agent Actually Works

How to measure an AI agent before and after deployment: baselines, process metrics, DATM, and practical examples from Documentolog and Klarna.

Build in public

The Six Months When AI Agents Rewrote Our Company

Half-year results at Documentolog: eight AI agents inside real processes, every client moved to d8n.ai, and why our people became co-owners of the outcome.

AI transformation

I Want to Adopt AI: 7 Steps to a Measurable Result — and the «Records» AI Agent in Practice

Why 95% of AI pilots deliver no impact, the seven steps to a real result, and a hands-on breakdown of our «Records» AI agent with the metrics to track.

Build in public

How We Lost 14 Leads — and What We Rebuilt After

How Documentolog rebuilt Lead Engine after an AI outage hid 14 new leads—and why AI must accelerate critical processes, not hold them up.

CEO DIGEST

Read new materials first

Longreads when published and one weekly digest. Email confirmation is required.