1/

// AI-NATIVE ORGANIZATIONS SPRINT

shared workspace → product agent → growth laws

from second brain for 30 people to AI-first value chain creation

Seva Ustinov

×

Alexander Povaliaev

dialog session · 10.04.2026 · 18:00 CET · 2h · Plurio AI × AI Mindset

2/

// AGENDA

Session Structure

block 1

Second Brain for Organization

~30 min

block 2

Human Side of Transition

~40 min

block 3

AI-first Value Chain

~20 min

block 4

Product as Agent — Sierra Model

~15 min

block 5

7 Growth Laws of AI Companies

~15 min

format: Seva presents + shows · Alexander guides with questions · last guest session before Demo Day (13.04)

3/

01

Second Brain
for Organization

single entry point for 30 people — humans and agents in one tree

~30 min

4/

// BLOCK 1 — SECOND BRAIN

Root → Hub-Router → Departments

every department has local rules, but everyone sees one structure — navigation works without skills, just folder instructions

5/

// BLOCK 1 — SECOND BRAIN

Tool-Agnostic Principles

SecondBrain DB

Code repositories & architecture
Client data & CRM
Knowledge base & docs
Meeting transcripts & decisions
Owner-folders with read-access for all

Portability

Principles designed before Claude Code existed
Cursor → Claude Code — "ran on same folder, worked out of the box"
Sync via GitHub + Dropbox backup
Context lives in the tree, not the tool

6/

// BLOCK 1 — SECOND BRAIN

The Anti-Pattern

silos

One department did something
Rest of the company doesn't know
Knowledge trapped in people's heads
Onboarding = "ask someone"
Context lost between projects

        second brain
        Single tree — everyone sees everything
Agents read the same context as humans
Knowledge accumulates in the system
Onboarding = "read the tree"
Decisions are traceable

      

"I've seen a million companies — one department does something, the rest has no idea."

— Seva Ustinov

7/

02

Human Side
of Transition

concrete timeline and cases — where the audience finds their own pain

~40 min

8/

// BLOCK 2 — HUMAN SIDE

Transition Timeline

9/

// BLOCK 2 — HUMAN SIDE

Kill Old Tools

"We closed Confluence access. No gradual migration. If you need information — it's in the new system. Period."

— Seva Ustinov, Plurio AI

Close access to legacy tools — no dual systems
Migration cost is real, but dual maintenance is worse
People adapt when there's no alternative
Head of Ops called every person individually — 2h each

10/

// BLOCK 2 — HUMAN SIDE

The Peak Moment

6 hours

Head of Ops, Friday evening
built a full project status analysis system from scratch

before

2 hours conversations with a manager to understand client project status

after

1 query tasks, blockers, messages, status — all in one view

11/

// BLOCK 2 — HUMAN SIDE · LIVE INSIGHT 10.04

It Always Starts With The Founder

"В каждом успешном кейсе, который я видел, это начиналось с founder'а. Истории 'найму архитектора, он сделает transition' — не работают."

— Seva Ustinov

manager swap story: I replaced a team lead with his AI-forward junior — both happy, no conflict

11/

// BLOCK 2 — HUMAN SIDE

Role Evolution & Metrics

People

→

People Managers

→

Agent Managers

0

assistants
(had them, removed)

-10%

FTE reduction

40%

roles transformed

4%

of payroll
spent on tokens

~3 months — full agent manager onboarding program

12/

03

AI-first
Value Chain

from individual skills to main agent orchestrator

~20 min

13/

// BLOCK 3 — VALUE CHAIN

AI-first Value Chain Architecture

14/

// BLOCK 3 — VALUE CHAIN

Evolution: Skills → Orchestrator

01

Employee

Individual skills, manual work, context in head

02

Co-pilot

AI assists human. Employee responsible. Skill to adopt.

03

Autonomous Agent

Agent executes full process. Company responsible. Skill for auto-tests.

04

AI/SaaS Product

Agent IS the product. Team supports and improves it.

knowledge stops living in people's heads — accumulates in agent, skills, and tools

15/

// BLOCK 3 — VALUE CHAIN

Client Integration: Weeks → Days

"Knowledge accumulates in the agent, not in people's heads. When someone leaves — the knowledge stays."

— Seva Ustinov

16/

// BLOCK 3 — VALUE CHAIN · LIVE INSIGHT 10.04

The Master Skill — 200 Lines That Merged Two Teams

6 wk

before — manual

1-2 day

first 80% automated

3-10x

clients per same team

5 wk ago

shipped end Feb 2026

"Running projects manually is now forbidden. We go by the skill, stumble, fix, improve. Product team + integration team merge around one agent."

— Seva, live 10.04

16/

04

Product as Agent
Sierra Model

from copilot to service provider — taking responsibility for outcomes

~15 min

18/

// BLOCK 4 — PRODUCT AS AGENT · LIVE INSIGHT 10.04

The Sandwich: Manager → Agents → People

before

This operational pattern was only accessible to huge companies with multi-layer management — like call centers with ops teams, group heads, automation pipelines.

now

1 manager + 3 reports can run this. Process iteration cycle: weeks → minutes. "There are no managers left except product managers."

17/

// BLOCK 4 — SIERRA MODEL

Sierra: The Benchmark

4 wk

PoC timeline

3% → 80%

cases by AI

50-70%

gross margin

6/6

design partners converted

20/

// BLOCK 4 — PRODUCT AS AGENT · LIVE INSIGHT 10.04

5 Stages of AI Product Adoption

how Plurio's clients actually adopt their ad-optimization agent — stage by stage

Stage 4 is the hardest push · Stage 5 is where Plurio's defensibility lives — agents improving agents

18/

// BLOCK 4 — SIERRA MODEL

From Copilot to Service Provider

"We're merging product and integration teams around one agent. The agent is the product — everything else supports it."

— Seva Ustinov, Plurio AI (next 1-2 months)

19/

// BLOCK 4 — SIERRA MODEL

Plurio AI: Applying the Model

Company Profile

30+ people, education industry
AI agent for marketing analytics
Clients: €100M+ ad budgets
Tripleten, Finom, InDrive

Transition in Progress

From copilot → service provider
Takes responsibility for ad optimization outcomes
Product + integration teams merging
Timeline: next 1-2 months

rebrand: Elly Analytics → Plurio AI (March 2026, funding round announced)

20/

05

7 Growth Laws
of AI Companies

research of 16 fastest-growing sales-led AI companies

~15 min

21/

// BLOCK 5 — GROWTH LAWS

Research Base: 16 Companies

16

companies analyzed

7

growth laws (0–6)

63-100%

prevalence range

Sierra

customer support

~$200M ARR / 24mo

Harvey

legal AI

~$200M ARR

Glean

enterprise search

40+ paid POCs

Hebbia

PE/finance analysis

9/10 top PE funds

Abridge

clinical documentation

2wk sales via Epic

Moveworks

IT helpdesk

250M+ tickets

Gong

revenue intelligence

NRR 140%

Writer

enterprise AI

NRR 209%

22/

// BLOCK 5 — GROWTH LAWS

Law 0: The Prerequisite

00

Value Magnitude as Prerequisite

the value gap must be so large it's undeniable

100%

Sierra

$13/contact → <$1/contact
13x cost reduction — impossible to argue against

Harvey

$250-400K/yr associate → $1,200/mo
per lawyer — the math sells itself

without massive value magnitude, none of the other laws matter

23/

// BLOCK 5 — GROWTH LAWS

Laws 1–3: Focus & GTM

01

One Workflow. Huge Obvious Value. Proven in 4 Weeks.

Harvey = M&A doc review · Sierra = support containment · 15/16 started with one wedge

94%

02

Win the Buyer the Market Follows

Harvey: Allen & Overy (Magic Circle) · Hebbia: 9/10 largest US PE funds in ~12 months

63%

03

Domain-Expert GTM Outperforms Generic Sales

hire practicing lawyers/bankers — they sell to colleagues · Harvey: 3-6 months at $500K+

75%

24/

// BLOCK 5 — GROWTH LAWS

Laws 4–6: Prove & Expand

04

Build Proof That Can't Be Argued With

Sierra: 6 design partners, 100% conversion · Gong: 11/12 alpha converted · Glean: 40+ paid POCs

88%

05

Price Against Labor Cost, Not Software

Harvey: $14,400/yr vs $250-400K · Decagon: 3.2x ROI, 65-95% cost reduction

75%

06

Build Expansion Into the Product Logic

11/16 companies >120% NRR · Writer: 209% NRR · Hebbia: inferred >200%

69%

% = prevalence among 16 benchmark companies

25/

// BLOCK 5 — GROWTH LAWS

Prevalence Overview

L0

Value Magnitude

100%

L1

One Workflow, Proven Fast

94%

L4

Proof Before Scale

88%

L3

Domain-Expert GTM

75%

L5

Labor-Budget Framing

75%

L6

Expansion Flywheel

69%

L2

Prestige First

63%

sorted by prevalence · research: posts, speeches, interviews of key people across 16 companies

26/

// BLOCK 5 — GROWTH LAWS

Secondary Patterns

S1

High-Touch Implementation as Moat

Glean: 80% adoption within 90 days · Harvey: 98% gross retention

S2

Trust Architecture as GTM Accelerant

Harvey: first AI startup with SOC 2 Type II + ISO 27001 — shortened procurement by 3-6 months

S3

Three-Phase Product Arc: Wedge → Platform → Agents

Harvey, Glean, Abridge, Gong, Decagon — same trajectory

S4

Rare Advantages Compress, Don't Replace the Playbook

Companies without rare advantages arrived 6-12 months later at same destinations

S5

ICP Discipline Before Scale

Decagon: 100+ discovery interviews · real ICP gave $150K+/yr WTP vs $1K/mo from non-ICP

S6

Non-Black-Box AI as Enterprise Prerequisite

11/13 companies built non-black-box from day one

27/

// BLOCK 5 — GROWTH LAWS

Counter-Intuitive Patterns

!

Inverted Onboarding

First onboard the largest clients, then small ones. Opposite of typical startup logic.

!

Domain Sellers

Hire practicing lawyers to sell to lawyers. Train them on AI, not the other way around.

!

15/16 Single Wedge

Almost all started with one narrow workflow. Platform ambition came after Series B/C.

Wedge

→

Platform

→

Agents

Three-Phase Product Arc — the universal trajectory

28/

// TAKEAWAYS

shared workspace →
human transition →
agent as product

the path is not optional — it's a sequence

01

Build the tree

Single context for humans & agents — tool-agnostic

02

Move people

Kill old tools, 3mo onboarding, evolve roles to agent managers

03

Ship the agent

Product = agent, outcome pricing, growth laws apply

29/

// DELIVERABLES

What You Take Away

01

AI-first Organization Principles

Second Brain architecture, tool-agnostic approach, owner-folder model

02

Transition Timeline

3 months onboarding + 2 months automation — step-by-step checklist

03

AI First Workspace Template

GitHub template to fork and customize for your team

04

7 Growth Laws Framework

Research data + link to full analysis at etc.sevaustinov.me

05

Miro Board: AI-first Transformation

Visual framework for value chain creation

30/

// AI-NATIVE ORGANIZATIONS SPRINT

Q&A

Seva Ustinov — Plurio AI
Alexander Povaliaev — AI Mindset

seva ustinov

×

alexander povaliaev

10.04.2026 · ai-native.aimindset.org

shared workspace → product agent → growth laws

Session Structure

Root → Hub-Router → Departments

Tool-Agnostic Principles

SecondBrain DB

Portability

The Anti-Pattern

silos

second brain

Transition Timeline

Kill Old Tools

The Peak Moment

before

after

It Always Starts With The Founder

Role Evolution & Metrics

AI-first Value Chain Architecture

Evolution: Skills → Orchestrator

Employee

Co-pilot

Autonomous Agent

AI/SaaS Product

Client Integration: Weeks → Days

The Master Skill — 200 Lines That Merged Two Teams

The Sandwich: Manager → Agents → People

before

now

Sierra: The Benchmark

5 Stages of AI Product Adoption

From Copilot to Service Provider

Plurio AI: Applying the Model

Company Profile

Transition in Progress

Research Base: 16 Companies

Law 0: The Prerequisite

Sierra

Harvey

Laws 1–3: Focus & GTM

Laws 4–6: Prove & Expand

Prevalence Overview

Secondary Patterns

Counter-Intuitive Patterns

Inverted Onboarding

Domain Sellers

15/16 Single Wedge

shared workspace →human transition →agent as product

Build the tree

Move people

Ship the agent

What You Take Away

AI-first Organization Principles

Transition Timeline

AI First Workspace Template

7 Growth Laws Framework

Miro Board: AI-first Transformation

Q&A

shared workspace →
human transition →
agent as product