Holak Scale v2.1p - private. Maturity of everyday AI use

This is the private track of the Holak Scale. The enterprise version, for teams and organizations, has its own article: Holak Scale v2.1e - enterprise.

The Holak Scale v2.1p measures the private maturity of using AI in everyday life: from simple chat, through personal instructions, home context, automations and agent workflows, all the way to a private agentic OS for daily life.

This isn’t about “living through AI”. It’s about AI helping where it genuinely reduces chaos, saves time, improves decisions and supports the user’s own goals.

UNESCO defines AI competency as the set of competencies needed to operate responsibly and creatively in a world with AI - including a human-centred approach, ethics, AI techniques and AI system design. That fits the private version well: AI should support a person, not take responsibility away from them.

Levels map v2.1p

Level	Name	In one sentence
0	I don’t use AI	No use, or a deliberate opt-out.
1	Basic chat	Ad-hoc questions: text, idea, quick answer.
2	Conscious questions	The user adds context, goal and answer format.
3	Private templates	Repeatable prompts for everyday tasks.
4	Personal instructions and token hygiene	AI knows the user’s preferences without being told every time.
5	Home context	Standing files and notes describe home, goals, preferences, constraints.
6	Private rules and boundaries	AI knows when to help, when to warn, and when not to pretend to be an expert.
7	Private skills and knowledge bases	AI has specialized areas: diet, budget, home, learning, pets, hobby.
8	Tools, automations and hooks	AI connects to calendar, notes, smart home, files, budget.
9	Daily agent workflows	AI runs whole processes: weekly plan, shopping, travel, documents.
10	Home orchestration	A few agents or roles cooperate on the household.
11	Private agentic OS	A personal operating system for home and life goals.

Level 0 - I don’t use AI

The user doesn’t use AI. They may not need it, may not trust the technology, or may not know where to start.

Barrier: anxiety, lack of need, no good first use case.
Success: a deliberate decision, not just one born of ignorance.
Trap: judging AI based on one bad example.

Level 1 - Basic chat

The user asks simple questions: “write a message”, “explain this”, “give me a dinner idea”, “summarize this text”.

Barrier: first contact.
Success: AI becomes a natural tool for small things.
Trap: treating the answer as truth without checking.

Level 2 - Conscious questions

The user starts adding context: who the answer is for, in what style, with which constraints, in what format.

Barrier: learning that AI doesn’t read minds.
Success: answers become more practical, shorter and better fitted.
Trap: asking too-general questions about important decisions.

Example. Instead of “Make a meal plan”, try: “Plan a vegetarian, lactose-free meal plan for 3 days, 1700 kcal a day, quick meals, working from home, one sweet snack a day, with a shopping list at the end.”

Level 3 - Private templates

The user has their own prompts for repeating things: shopping, weekly plan, bill analysis, drafting an email, comparing products, language learning, training plan.

Barrier: identifying the repeating cases.
Success: AI helps faster because the user doesn’t start from zero.
Trap: collecting prompts no one actually uses.

Level 4 - Personal instructions and token hygiene

AI knows the user’s standing preferences: language, style, dietary restrictions, translation style, level of detail, shopping preferences, planning style.

This is the level where private token and time saving begins. You don’t repeat every time: “write in Polish”, “don’t suggest expensive options”, “include the lactose-free diet”, “give me step-by-step”.

Barrier: setting and maintaining the profile.
Success: shorter questions produce better answers.
Trap: the profile gets too long and contradictory.

Level 5 - Home context

The user creates a standing context for household matters: appliance list, financial goals, food preferences, home layout, pet info, renovation projects, build notes, schedules, standard procedures.

Barrier: sorting out your own data.
Success: AI stops answering generically and starts helping in your situation.
Trap: dumping in too-sensitive data without thinking.

Private rule: not every document should go into AI. Medical, financial, children’s, ID and credentials data deserve extra care.

Level 6 - Private rules and boundaries

AI has described limits: when to ask, when to warn, when to suggest a specialist, when not to make decisions for the user.

Barrier: defining what is help and what is risk.
Success: AI supports decisions but doesn’t pretend to be a doctor, lawyer, financial advisor or licensed electrician.
Trap: over-trust in answers on high-risk topics.

Example rules:

“On health and medicine, always note that this doesn’t replace a medical consultation.”
“For electrical installation, list risks and indicate what an electrician should do.”
“On finance, show your assumptions and risks; don’t give false certainty.”

Level 7 - Private skills and knowledge bases

The user has specialized “AI areas”:

skill for meal planning,
skill for budget,
skill for build planning,
skill for smart home,
skill for pet care,
skill for learning,
skill for contract analysis,
skill for travel planning.
Barrier: separating topics and contexts.
Success: AI doesn’t mix things up - diet is diet, build is build, finance is finance.
Trap: private chaos of notes and instructions.

Level 8 - Tools, automations and hooks

AI starts acting with tools: calendar, task list, notes, smart home, files, spreadsheets, budget system, email, shopping app.

Hooks in the private version are automated rules - for example:

after adding a recipe, generate a shopping list,
when budget is exceeded, warn,
before a trip, check the packing list,
if the weekly plan is too packed, propose a simpler version,
if a smart-home automation touches security, ask for confirmation.
Barrier: integrations and privacy.
Success: AI starts saving real time.
Trap: automating everything, including things better left manual.

Level 9 - Daily agent workflows

The user states a goal and AI runs the whole process within a bounded scope.

Examples:

“Plan the week: meals, shopping, training, cleaning and free time.”
“Prepare a short trip plan with budget, weather, route and packing list.”
“Analyze the last few months of bills and show where money leaks.”
“Plan to cut chaos in household documents.”
Barrier: trust and control.
Success: AI delivers a ready plan; the user only approves and tweaks.
Trap: AI builds a plan no one actually uses.

Level 10 - Home orchestration

A few agents or roles cooperate on a single goal. They don’t technically have to be separate bots - they can be one model running in several roles.

Example: planning a household month.

Finance agent watches the budget.
Diet agent plans meals.
Logistics agent plans shopping and chores.
Smart-home agent checks automations.
Reviewer agent hunts risks and overload.
Barrier: coordination and simplicity.
Success: several life areas become more consistent.
Trap: domestic overengineering - more system than benefit.

Level 11 - Private agentic OS

This is a new level in v2.1p.

The user designs their own Life OS / Home OS - a private operating system for daily life, in which AI helps manage goals, context, rituals, automations and decisions.

It’s not about full autonomy. It’s about a coherent support system.

Elements of a private agentic OS:

goals - health, budget, home, learning, family, pets, energy, free time,
context - preferences, constraints, documents, plans, schedules,
memory - what is permanent, what changes, what needs periodic review,
automations - reminders, lists, smart home, summaries,
control - what AI can do on its own, what needs approval,
privacy - which data is restricted,
review - weekly or monthly check-in.
Barrier: designing a system that helps rather than tires.
Success: AI genuinely reduces daily chaos.
Trap: building a “life command center” that takes more time to maintain than it gives back.

Level 11 test: can you say, “I have a private system in which AI knows my goals, constraints and processes, helps me carry them out, but doesn’t take over decisions that should stay mine”?

Main differences from v2.1e

Area	v2.1e - enterprise	v2.1p - private
Goal	Business outcome, quality, speed, risk, cost	Better daily life, less chaos, time saved
Risk	Compliance, data, reputation, audit, permissions	Privacy, over-trust, automation chaos
Governance	Formal owner, policies, agent registry	Personal rules and limits
Level 8	Tools, MCP, hooks, identity, logging	Calendar, notes, smart home, budget, automations
Level 9	Agent runs a business process	Agent runs a daily workflow
Level 10	Multi-agent orchestration of processes	A few roles/agents coordinating life
Level 11	Agentic OS for a business process or domain	Life OS / Home OS for private goals

v2.1p anti-patterns

AI instead of a decision. Asking AI about everything, especially things you should decide yourself (health, finance, relationships).
Pasting everything into chat. Medical, financial and private documents without thinking about the tool’s privacy policy.
Automation chaos. Hooks and automations grow faster than the user understands. The lamp turns on at 7 because “it used to be cold” - and nobody knows why anymore.
Frankenstein profile. Custom instructions balloon into 2,000 words of contradictory preferences.
Life OS without a life. A system so elaborate that running the OS takes more time than living.

What’s next

The enterprise version of the scale has its own article: Holak Scale v2.1e - enterprise. That one measures AI maturity in organizations: governance, MCP, agentic OS for a business process.

Earlier version: AI maturity model (v1) as historical context. The v2.0 content has been folded into v2.1e and v2.1p.

Version history

v1.0 - AI adoption maturity model. Created in March 2026, published 15 April 2026.
v2.0 - Holak Scale. Created in April 2026, published 20 April 2026.
v2.1p (this article) - private. Splits private and enterprise tracks, places hooks at level 8, defines level 11 as a private agentic OS. 17 May 2026.
v2.1e - enterprise. Enterprise version published 18 May 2026. Read v2.1e →

The scale remains an iteratively evolving model. AI, tools and the way we work with agents change faster than most maturity models, which is why the Holak Scale will keep being updated as it gets used in organizations and everyday life.