AI Productivity OS Builder

How the generator decides

The builder is a set of plain rules, not a model that studies you. It never reads your calendar, inbox or files; every output is derived from the thirteen answers you typed, which stay in your browser.

• Deep-work estimate. The focus blocks in your suggested day plan, times five working days. The plan first reserves planning, a break, lunch, a shutdown slot, your meetings and an admin batch whose size depends on role (about 4 h/week for engineers and students, up to 10 h for operations roles) and grows with team size; what is left becomes focus time. It is a rough starting number, not a measurement.
• Daily workflow. Fixed rituals first (planning, breaks, shutdown), then a meeting window sized from your weekly average, then an admin batch. Whatever is left becomes one or two deep-work blocks whose maximum length follows your work style: makers get up to 150 minutes, structured planners 120, flexible workers 90 and managers 75 plus a touchpoint slot. Times start at 09:00 only because a start time is needed to draw the day; shift it freely.
• Weekly rhythm. Low meeting load (under 5 h) gives more maker days, high load (over 15 h) batches 1:1s on Tuesday, and teams of 21+ get a mid-week priorities check.
• Priority system. Teams of six or more led by a manager, executive, founder or operations role get Kanban with WIP limits. Leaders fighting unclear priorities or interruptions get Eisenhower triage. People with many inputs, context switching or a knowledge goal (and some system already) get GTD-style capture and review. Everyone else gets the lightest option, a Weekly Big 3 with a daily highlight.
• App stack. Tools you already use are kept. A $0 budget shows free plans only. "Self-hosted / no cloud AI" removes cloud assistants and meeting recorders and favours self-hostable, offline and open-source tools; the EU option favours tools the catalog marks as EU-based; a regulated industry adds a policy warning. The rest is scoring by role, team size, goals and obstacles, one tool per category, six at most (four if you named tool sprawl).
• Automations and AI. Recipes are triggered by obstacles and role and are assigned to a tool already in your stack, or to built-in email and calendar features if no automation tool fits. AI suggestions scale with your stated experience and are switched to local-only models when you chose self-hosting.

The evidence behind the rules

Planning rituals and time blocking. The daily planning ritual asks you to decide when and where a task happens, which is the core of an implementation intention. A meta-analysis of 94 studies found a medium-to-large effect of implementation intentions on goal attainment (d = 0.65), one of the better-replicated findings in behavioural science.^[1] A Scheduling a highlight into a concrete block is a direct application; the exact block lengths in the tool are conventions, not findings.

Task switching. Switching between tasks carries measurable costs: slower responses and more errors right after a switch, and a residual cost that practice reduces but does not remove.^[2] B That is why the plan batches shallow work and meetings instead of scattering them, and why the automation recipes aim to reduce the number of places you have to check.

Interruptions. In a controlled study, interrupted participants finished their tasks in less total time but reported significantly higher stress, frustration, time pressure and effort.^[3] B The often-quoted figure that it takes "23 minutes" to return to a task after an interruption comes from an interview with the lead author, not from this paper, so the tool does not use it. The design lesson is still sound: protect a block, and let the inbox wait for its batch.

AI delegation. In a field experiment with 758 consultants, those using an AI assistant completed more tasks, faster and at higher rated quality for tasks inside the model's competence, but were less likely to be correct on a task designed to sit just outside it, the "jagged frontier."^[4] B A preregistered online experiment with professional writing tasks found large time savings and quality gains, with the biggest gains for weaker writers, while raising questions about skill development.^[5] B Together they justify the tool's two AI rules: delegate first drafts and summaries early, and keep a verification habit because the failure mode is quiet.

Limitations

• The plan is a starting template built from self-reported answers and common evidence-informed practices. It is not an assessment of you, your performance or your team.
• The deep-work number is arithmetic on your inputs, not a measurement, and the admin estimate is a guess by role.
• Block lengths, weekly themes and the choice of priority system are conventions that the evidence supports in general terms; no study tested these exact combinations.
• The catalog is small. Links go straight to the vendors and we are not in any affiliate programme yet; our methodology explains how rankings would stay independent of commissions if that changes.
• No productivity gain is promised. Try one change for two weeks, keep what helps, and re-run the builder when your situation changes.

Frequently asked questions

Does the AI Productivity OS Builder read my calendar or data?

No. It uses only the thirteen answers you give, keeps them in your browser's local storage and makes no network calls. The share link encodes your answers and nothing else.

Why does it recommend tools I did not ask for?

Each recommendation follows rules from your role, team size, budget and privacy answers, and the one-line note on every card names the rule. Tools you already use are kept so you are not pushed to switch.

Can I use it without any AI tools?

Yes. Choose "None" for AI experience or "Self-hosted / no cloud AI" for privacy, and the AI section shrinks to small experiments or local-only options.

How long until the system works?

The checklist spreads setup over four weeks; the estimate assumes about 90 minutes per new tool and 45 minutes per automation. Expect to adjust the plan after the first week.

Is the premium automation plan available?

Not yet. The button only records your interest in the newsletter announcement; nothing is charged and no launch date is implied.

Sources and evidence

1. Implementation intentions and goal achievement: A meta-analysis of effects and processes — Gollwitzer, P. M.; Sheeran, P. (2006), meta-analysis, Advances in Experimental Social Psychology 38 A doi.org/10.1016/S0065-2601(06)38002-1 (retrieved 2026-11-09)
2. Task switching — Monsell, S. (2003), review, Trends in Cognitive Sciences A doi.org/10.1016/S1364-6613(03)00028-7 (retrieved 2026-11-09)
3. The cost of interrupted work: more speed and stress — Mark, G.; Gudith, D.; Klocke, U. (2008), controlled experiment, CHI 2008 B doi.org/10.1145/1357054.1357072 (retrieved 2026-11-09)
4. Navigating the Jagged Technological Frontier: Field experimental evidence of the effects of AI on knowledge worker productivity and quality — Dell'Acqua, F.; McFowland, E.; Mollick, E. R.; Lifshitz-Assaf, H.; Kellogg, K.; Rajendran, S.; Krayer, L.; Candelon, F.; Lakhani, K. R. (2023), field experiment, HBS working paper 24-013 (SSRN) B doi.org/10.2139/ssrn.4573321 (retrieved 2026-11-09)
5. Experimental evidence on the productivity effects of generative artificial intelligence — Noy, S.; Zhang, W. (2023), preregistered online experiment, Science 381(6654) A doi.org/10.1126/science.adh2586 (retrieved 2026-11-09)