Running one's life admin: what it needs beyond tasks¶
The question: what does a person need to run their life admin (letters, money, offices, appointments, papers, the people who help) that a task manager does not give, and what should a calm admin app refuse? It draws on the research literature (executive function, prospective memory, administrative burden, autistic burnout, why tools are dropped), on public sources, and on the apps that compete in the neurodivergent and life-admin niche.
Researched on 3 October 2026. "Built", "partly", "next", "not built" and "refused" say where Sioul stood on 4 October 2026. Task design itself is in tasks.md; stopping for the day in done-for-the-day.md; the rules in short in research.md.
How evidence is graded: meta (meta-analysis or systematic review), RCT, study (smaller or observational research, its kind said), qualitative, grey (reports and surveys by organisations). For apps: vendor (the maker's own page), review (a third-party page), snippet (only a search-result summary, not opened: unverified). "(unverified)" marks what could not be confirmed at a primary source.
1. What the literature says matters¶
1.1 Remembering to act: prospective memory, reminders, cues¶
- Remembering at a time is the weak spot; remembering on a cue much less. In autism, a large impairment of time-based prospective memory and a small one of event-based. In adult ADHD, no difference on event-based, a large impairment on time-based. In adult ADHD the failures come mostly from planning, while recalling and carrying out the plan were largely intact.
- Landsiedel, Williams & Abbot-Smith (2017), A meta-analysis and critical review of prospective memory in autism spectrum disorder, JADD 47(3): https://pmc.ncbi.nlm.nih.gov/articles/PMC5352792 — meta.
- Altgassen, Kretschmer & Kliegel (2014), Task dissociation in prospective memory performance in individuals with ADHD, J Atten Disord 18(7): https://journals.sagepub.com/doi/full/10.1177/1087054713518236 — study, 25 + 25.
- Fuermaier et al. (2013), Complex prospective memory in adults with ADHD, PLoS ONE: https://pmc.ncbi.nlm.nih.gov/articles/PMC3590133 — study, 45 + 45.
- Prompting technology works when memory is the bottleneck: reminders improved memory performance against usual practice or a paper diary, d = 1.27 (acquired brain injury and degenerative disease, n = 147 in the pooled group studies).
- Jamieson, Cullen, McGee-Lennon, Brewster & Evans (2014), The efficacy of cognitive prosthetic technology for people with memory impairments, Neuropsychological Rehabilitation 24(3–4), doi:10.1080/09602011.2013.825632: https://eprints.gla.ac.uk/86137 — meta.
- Reminders raise attendance and follow-through. Text reminders increased attendance at healthcare appointments over no reminder or a postal one, as much as phone-call reminders (8 RCTs, 6,615 people, low to moderate quality). Reminders raised saving by 6 % and goal attainment by 3 points. Prompting people to write the date and time they plan to act raised flu vaccination by 4.2 points over 33.1 %.
- Gurol-Urganci, de Jongh, Vodopivec-Jamsek, Atun & Car (2013), Mobile phone messaging reminders for attendance at healthcare appointments, Cochrane CD007458: https://cochranelibrary.com/cdsr/doi/10.1002/14651858.CD007458.pub3/information — meta.
- Karlan, McConnell, Mullainathan & Zinman (2016), Getting to the top of mind: how reminders increase saving, Management Science 62(12): https://img.nber.org/papers/w16205 — field RCTs.
- Milkman, Beshears, Choi, Laibson & Madrian (2011), Using implementation intentions prompts to enhance influenza vaccination rates, PNAS: https://pmc.ncbi.nlm.nih.gov/articles/PMC3127912 — field RCT. Review: Rogers, Milkman, John & Norton (2015), Beyond good intentions, Behavioral Science & Policy: https://toddrogers.scholars.harvard.edu/publications/beyond-good-intentions-prompting-people-make-plans-improves-follow-through
- A reminder can move the cost elsewhere: credit-card payment reminders cut late fees by 14 % but raised overdraft fees by 9 %; people with a history of overdrafts paid 5 % more in total.
- Medina (2021), Side effects of nudging: evidence from a randomized intervention in the credit card market, Review of Financial Studies (doi:10.1093/rfs/hhaa108): https://www.socialscienceregistry.org/trials/3034 — field RCT, Brazil.
- Reminders help repetition, not habit; event cues build automaticity (Stawarz, Cox & Blandford 2015, CHI; in tasks.md). People offload an intention when offloading looks cheaper than remembering.
- Risko & Gilbert (2016), Cognitive offloading, Trends in Cognitive Sciences 20(9): https://discovery.ucl.ac.uk/id/eprint/1508770/ — review.
- Rule: remembering "on the 30th" is exactly what fails, so the software carries time: a reminder, and a date turned into a step; steps tied to events where possible; the planner does the planning. Noting a date must cost less than remembering it. A payment reminder reads the account's balance first.
- In Sioul: built. Reminders before events, dates asked, waits over and planned payments, once each, also with the window closed (reminders.md; not on Windows yet); a payment's reminder says whether the account will hold it (accounting.md, "The bank"); the plan does the ordering (tasks.md). Cues from events: next.
1.2 Administrative burden, and rights not claimed¶
- Burdens are built, and fall hardest on those with the fewest resources. A review of 119 papers finds them consequential, distributive and constructed, shaped by frontline delivery and government communication.
- Herd & Moynihan (2018), Administrative Burden: Policymaking by Other Means, Russell Sage Foundation: https://www.bsg.ox.ac.uk/events/administrative-burden-policymaking-other-means — book.
- Halling & Bækgaard (2024), Administrative burden in citizen–state interactions: a systematic literature review, JPART: https://pure.au.dk/ws/files/452909104/Administrative_Burden_in_Citizen_State_Interactions_Final_version_2024.pdf — meta (systematic review).
- Fewer demands, less stress: reducing compliance demands on Danish benefit recipients raised their sense of autonomy in three studies and lowered stress in two.
- Bækgaard, Mikkelsen, Madsen & Christensen (2021), Reducing compliance demands in government benefit programs improves the psychological well-being of target group members, JPART 31(4), doi:10.1093/jopart/muab011: https://rucforsk.ruc.dk/ws/files/78389034/muab011.pdf — survey experiment + natural experiment.
- Complexity costs money; clarity pays. In an IRS field experiment with 35,050 people who had not claimed a credit, complex material cut take-up by 6 points and a longer worksheet by 4, while showing the amount at stake raised it by 8; a simple second mailing alone brought many claims.
- Bhargava & Manoli (2015), Psychological frictions and the incomplete take-up of social benefits: evidence from an IRS field experiment, American Economic Review 105(11): https://www.cmu.edu/dietrich/sds/docs/bhargava/bhargava-aer-2015.pdf — field RCT.
- Non-take-up in France is large: a third of households eligible for the RSA did not claim it in a given quarter in 2018 (34 %), one in five for three quarters running. About half for the minimum vieillesse (ASPA), per secondary sources (unverified in DREES). Reasons given in a DREES survey: lack of information 37 %, complex procedures 22 %, fear of consequences 17 % (reported by aide-sociale.fr; unverified in the DREES source).
- DREES (2022), Mesurer régulièrement le non-recours au RSA et à la prime d'activité, Les Dossiers de la DREES n° 92, 11 February 2022: https://drees.solidarites-sante.gouv.fr/publications-communique-de-presse/les-dossiers-de-la-drees/mesurer-regulierement-le-non-recours-au — grey. Secondary: https://www.aide-sociale.fr/non-recours/
- Admin anxiety is common, and the phone is a wall: 54 % of people with mental health problems have serious difficulty using the phone for essential admin, against 32 % of others; four in ten have severe "admin anxiety" (breathlessness, sweating, shaking); 23 % of all adults have admin anxiety severe enough to suggest a mild phobia.
- Money and Mental Health Policy Institute (2018), Access Essentials: https://www.moneyandmentalhealth.org/wp-content/uploads/2018/06/Money-and-Mental-Health-Access-Essentials-report.pdf ; press release with the figures: https://www.moneyandmentalhealth.org/press-release/access-essentials-pressrelease/ — grey (UK survey).
- Rule: explaining a letter in plain words and showing what is at stake (an amount, a date) is not cosmetic: it changes whether rights are claimed. Fewer demands; written channels before the phone.
- In Sioul: a paper letter is read into a card that says who wrote, what it is, how much it asks and by when, as a date: built (porch.md, "Paper letters"). The AI explaining a letter line by line, and showing what waiting costs on an aid request: not built.
1.3 Money: arrears, avoidance, ADHD¶
- Arrears grow with age in ADHD and precede suicide. Swedish register and credit data (11.5 million adults; credit data for 189,267): people with ADHD start adult life with normal default rates, then defaults grow steeply by middle age; among men with ADHD, unpaid debts rose in the three years before suicide.
- Beauchaine, Ben-David & Bos (2020), ADHD, financial distress, and suicide in adulthood: a population study, Science Advances: https://pmc.ncbi.nlm.nih.gov/articles/PMC7527218 — register study.
- Debt and mental health go together: odds ratio 3.24 for mental disorder, 2.77 for depression, 7.9 for completed suicide.
- Richardson, Elliott & Roberts (2013), The relationship between personal unsecured debt and mental and physical health, Clinical Psychology Review 33(8): https://pubmed.ncbi.nlm.nih.gov/24121465/ — meta, 65 papers.
- ADHD symptoms go with avoidant financial decision styles, impulsive buying and less retirement saving, though not independently of depression and personality.
- Bangma et al. (2020), Financial decision-making in a community sample of adults with and without current symptoms of ADHD, PLoS ONE: https://ideas.repec.org/a/plo/pone00/0239343.html — study, 1,292 adults, cross-sectional.
- The ostrich effect: people log into their accounts less when debt is high or balances low, and more on payday.
- Olafsson & Pagel (2017), The ostrich in us: selective attention to financial accounts, income, spending, and liquidity, NBER w23945: https://www.nber.org/papers/w23945 — study (panel of app logins).
- The "ADHD tax" (late fees, forgotten subscriptions, replacements) is a community term; no peer-reviewed measure of it was found (unverified as a quantity; e.g. https://www.simplypsychology.org/adhd-tax.html).
- Rule: missed bills are a health matter, not only a money one. Nobody looks when it is bad, so the software notices for them (a debit that did not happen, a balance that will not cover the rent) and says it calmly and early.
- In Sioul: built. The bank's own exports are read against the budgets' recurring payments: a payment missed, a payment changed, a balance that will fall short, the week's payments in one line, never in red (accounting.md, "The bank"). Contracts and subscriptions with their renewal and notice dates: built. Fetching from the bank itself: not built (exports are taken by hand).
1.4 Phone calls, written channels, appointments¶
- E-mail for strangers, the phone last: autistic adults ranked e-mail first for communicating with strangers and phone calls last, citing missing non-verbal cues, the unplanned, the fast pace and the cost of masking; writing gives time to process.
- Howard & Sedgewick (2021), "Anything but the phone!": communication mode preferences in the autism community, Autism, https://doi.org/10.1177/13623613211014995 — survey, 247 autistic adults.
- Booking by phone keeps people out of care: 62 % of autistic adults against 16 % of non-autistic adults had difficulty using the telephone to make appointments; 80 % had difficulty visiting a GP; one in three had been unable to access care for a potentially life-threatening condition.
- Doherty, Neilson, O'Sullivan, Carravallah, Johnson, Cullen & Shaw (2022), Barriers to healthcare and self-reported adverse outcomes for autistic adults: a cross-sectional study, BMJ Open 12:e056904: https://bmjopen.bmj.com/content/12/2/e056904 ; figures as summarised by the Brighton and Sussex Medical School: https://www.bsms.ac.uk/about/news/2022/02-23-one-in-three-autistic-adults-unable-to-access-healthcare-for-potentially-life-threatening-conditions.aspx — cross-sectional, 507 + 157.
- "Respect the way I need to communicate with you": barriers include providers' willingness to allow written communication and to receive personal information before the appointment. The AASPIRE toolkit's Autism Healthcare Accommodations Tool makes a personalised accommodations report for the doctor; over 94 % of 259 autistic adults found it easy, important and useful.
- Nicolaidis et al. (2015), "Respect the way I need to communicate with you": healthcare experiences of adults on the autism spectrum, Autism: https://pubmed.ncbi.nlm.nih.gov/25882392/ — qualitative, 39 autistic adults + 16 supporters.
- Nicolaidis et al. (2016), The development and evaluation of an online healthcare toolkit for autistic adults and their primary care providers, J Gen Intern Med: https://pmc.ncbi.nlm.nih.gov/articles/PMC5023610 ; the toolkit: https://researchautism.org/healthcaretoolkit — single-arm pre/post.
- Speech can fail under load: speaking autistic adults may have intermittent or unreliable speech and use typed communication part-time.
- Zisk & Dalton (2019), Augmentative and alternative communication for speaking autistic adults, Autism in Adulthood: https://pmc.ncbi.nlm.nih.gov/articles/PMC8992808 — overview.
- Rule: a written channel first. A one-page accommodations sheet is an evidence-based object: keep it, attach it when writing to a care service or an office for the first time. For the calls that cannot be avoided: a script, a typed fallback, and the outcome noted.
- In Sioul: partly. A "call" kind of task, offices' opening hours, "Call" on a contact (tasks.md). The accommodations sheet, call cards and a "please answer me in writing" draft: not built.
1.5 Burnout, demands and autonomy¶
- Autistic burnout is exhaustion, withdrawal, loss of executive function and lower tolerance of stimuli, distinct from depression. Accounts of recovery stress reducing demands, rest, practical support and autonomy; routines help.
- Raymaker et al. (2020), "Having all of your internal resources exhausted beyond measure…": defining autistic burnout, Autism in Adulthood 2(2): https://pmc.ncbi.nlm.nih.gov/articles/PMC7313636 — qualitative.
- Higgins, Arnold, Weise, Pellicano & Trollor (2021), Defining autistic burnout through experts by lived experience, Autism: https://researchers.westernsydney.edu.au/en/publications/defining-autistic-burnout-through-experts-by-lived-experience-gro/ — Delphi, 23 autistic experts.
- Mantzalas, Richdale, Adikari, Lowe & Dissanayake (2022), What is autistic burnout? A thematic analysis of posts on two online platforms, Autism in Adulthood 4(1), doi:10.1089/aut.2021.0021: https://pmc.ncbi.nlm.nih.gov/articles/PMC8992925 — qualitative.
- Clarey, Abel, Ireland & Brownlow (2025), Autistic burnout on Reddit: a Sisyphean struggle with daily tasks, JADD, doi:10.1007/s10803-025-06765-4: https://inclusivefutures.griffith.edu.au/autistic-burnout-on-reddit-a-sisyphean-struggle-with-daily-tasks — qualitative, 249 Reddit threads.
- Demand avoidance in adults rests on anxiety and intolerance of uncertainty; adults with PDA ask for autonomy and flexibility, and name trust and safety.
- Johnson & Saunderson (2023), Examining the relationship between anxiety and pathological demand avoidance in adults: a mixed methods approach, Frontiers in Education: https://www.frontiersin.org/journals/education/articles/10.3389/feduc.2023.1179015/full — mixed methods.
- Kenny & Doyle (2024), A phenomenological exploration of the lived experience of adults experiencing pathological demand avoidance, DCU: https://doras.dcu.ie/30963/ — qualitative.
- Energy accounting and "spoons" are clinical and community tools without trials (Toudal & Attwood, JKP 2025; no peer-reviewed evaluation found). Burnout research supports the principle (lower demands), not a points system.
- Rule: every demand Sioul relays arrives with its uncertainty already reduced, as a choice, with a way to hand it over. Lower demands, not points.
- In Sioul: the day's weather and what a task takes, without points: built (tasks.md). Handing a case to a helper: not built (section 3).
1.6 Company, senses, shame¶
- Body doubling was found by neurodivergent people themselves (Eagle & Ringland 2024); in virtual reality, 12 adults with ADHD finished faster with a human or an AI double than alone, with divergent preferences (Ara et al. 2025, arXiv:2509.12153, https://arxiv.org/abs/2509.12153 — study). The evidence is thin: offer it, do not build around it.
- Animation costs autistic users more: irrelevant animations distracted autistic participants more, raised mental demand and made them rush.
- Uitdenbogerd, Spichkova & Alzahrani (2022), Web-based search: how do animated user interface elements affect autistic and non-autistic users?, ENASE: https://arxiv.org/abs/2211.11993 — study.
- Shame and criticism: finding 23 of research.md, and tasks.md (Beaton, Sirois & Milne 2022).
- In Sioul: company: next (tasks.md). Reduced motion respected, no badges, no counts: built (design.md, "Keyboard and senses").
1.7 Why tools are dropped¶
- Most people stop within two weeks: across 93 popular mental-health apps, median retention was 3.9 % at 15 days and 3.3 % at 30 days.
- Baumel, Muench, Edan & Kane (2019), Objective user engagement with mental health apps, JMIR 21(9):e14567: https://jmir.org/2019/9/e14567/ — usage panel.
- People stop tracking because it costs effort, because the data feel useless, or out of guilt; financial tracking included. Self-tracking makes "substantial interpretive and emotional demands" on neurodivergent people (Rudberg Selin, Unéus & Knudsen 2026, CHI; finding 11 of research.md).
- Epstein, Caraway, Johnston, Ping, Fogarty & Munson (2016), Beyond abandonment to next steps, CHI: https://pmc.ncbi.nlm.nih.gov/articles/PMC5428074/ — 193 surveyed + 12 interviewed.
- Adults with ADHD describe planning anxiety and spending more time organising the tool than doing the task, and reject tools that seem to watch them (Chen, Meng & Nie 2026, CSCW; in tasks.md).
- Rule: every new feature arrives filled from mail, a scan or a bank file, never as a form to keep up.
- In Sioul: built for papers (kept from a mail's attachment, kind guessed from its name), contracts (found from recurring payments and mail), bank lines (the bank's exports) and letters (scans): "the form asks only what cannot be guessed" (papers.md, accounting.md).
1.8 Crisis plans¶
- Safety-planning-type interventions lowered suicidal behaviour (relative risk 0.57, NNT 16; 6 studies, 3,536 people) but not ideation.
- Nuij et al. (2021), Safety planning-type interventions for suicide prevention: meta-analysis, British Journal of Psychiatry, doi:10.1192/bjp.2021.50: https://www.doi.org/10.1192/bjp.2021.50 — meta.
- Rule: a plan written by the person and held by them has evidence behind it; in an admin app it fits only as a private note the person opens themselves.
- In Sioul: not built.
2. What competing apps offer¶
Checked on 3 October 2026. The task and focus apps (Tiimo, Goblin Tools, Llama Life, Structured, Routinery, Amazing Marvin, Sunsama, Lunatask, Inflow, Focusmate, Akiflow, Motion, Saner.AI, Todoist, Things, Super Productivity, Finch, Habitica) are in tasks.md; this table adds the paperwork, money and phone apps.
| Feature | Apps that offer it | In Sioul |
|---|---|---|
| Phone app, widgets, watch, push reminders | Tiimo (iOS, Android, watch, web; widgets and Live Activities); Structured (free tier: notifications, widgets, Live Activities, watch); TickTick ("nag" reminders, review); Todoist, Things, Lunatask | Universal in the niche. Sioul is a desktop app; Android is on the roadmap (roadmap.md, slice 12). |
| Reminders before dates (events, deadlines, payments) | Tiimo, Structured, TickTick, Todoist; fileee ("reminds you of upcoming deadlines, such as payment terms"); BriefAI ("deadline reminders before dates pass"); Snoop (a Sunday "look ahead" of next week's payments); Ohai (reminders suggested from dates in e-mails and PDFs) | The most common feature. Built: once each, quietly, also with the window closed (reminders.md). |
| Paper letters scanned: OCR, sender and kind recognised, deadline found, filed | fileee; Paperless-ngx (open source, GPL-3.0: Tesseract OCR, full-text search, learned matching of correspondent, type and tags, mail ingestion, workflows); BriefAI, LetterSnap, Clario, Briefed, Bureaucracy Buddy (phone scans of official letters); Digiposte (collects documents automatically from 300+ French bodies) | The core of the paperwork niche, absent from task apps. Built: scans dropped in a folder are read on the computer and wait as cards (porch.md, "Paper letters"); an intake address for a helper: not built. |
| An official letter explained: what it is, what to do, by when, the risk | BriefAI (US, DE, AT, CH, FR: "what you must do, by when", "a real date, not 'within 30 days'", risk level, amounts and references, scam "legitimacy hints", a drafted reply; $4.99/month); Mika (German tax letters); Bureaucracy Buddy (a three-sentence summary, urgency); Clario (18 languages, scam checks); Goblin Tools Professor | Partly: the letter card by rules (who, what, how much, by when as a date, registered or not). The companion explaining line by line and the "is this genuine?" checks: not built (ai.md). |
| A reply drafted, a real letter sent, registered if needed | LetterSnap ("send a real letter directly from the app", snippet); Merci Facteur (prints, envelopes and posts; registered with acknowledgment, about €7.16 for one page in 2026, review); La Poste's online registered letter (a PDF up to 5 pages; the acknowledgment €1.45 in 2026); AR24 (eIDAS-qualified electronic registered letter, La Poste group, about €6.49, review) | Not built: Sioul writes e-mail; a contract's cancellation letter is drafted for the person to sign and send. |
| Contracts and subscriptions found, notice periods, cancellation | Finanzguru (contracts recognised from bank lines, notice periods, cancellation letters filled in, a reminder for deadlines); Rocket Money (every recurring charge listed, cancellation concierge, bill negotiation for 35–60 % of a year's savings); Snoop (renewal dates, alerts 3–4 weeks before); Bankin', Linxo | Built: renewal and notice dates, the notice usually asked in France, one reminder before the last day a notice can leave, the cancel page or a drafted letter; never an offer (accounting.md, "Contracts and subscriptions"). Online cancellation "in three clicks" has been mandatory in France since 1 June 2023 (loi 2022-1158). |
| Bank data: balance forecast, overdraft warning, missed or changed debits | Linxo (30-day forecast, alerts "to anticipate bank fees or overdraft risk", Premium); Bankin' (end-of-month balance); Snoop (alerts on big swings in regular bills); Rocket Money, Finanzguru; open source: Kresus with Woob (GPL), Actual Budget and Firefly III (GoCardless bank sync, closed to new sign-ups since July 2025) | Built from the bank's exports: missed, changed, short (accounting.md, "The bank"). Fetching from the bank: not built. |
| A papers wallet with expiry reminders | Folio (details extracted, reminders before expiry, end-to-end encrypted); Expiry Guard, RenewalKit, Expiro (snippets); Digiposte (ID, payslips, attestations, sharing with third parties); DossierFacile (a state service: a verified tenant file shared by link; 1.8 million users) | Built: typed papers with validity, renewal reminded once and planned as a task, attached from the wallet in a mail (papers.md). |
| Procedures and rights explained, simulators | mesdroitssociaux.gouv.fr (eligibility to 58 national and local aids); service-public.fr (practical sheets with steps, forms and online services, published daily as open data in XML under the Licence Ouverte); "Je change de coordonnées" (one declaration of a new address to many offices, from 3 months before to 3 months after a move) | Not built. A paper's renewal task carries the steps known for its kind (papers.md). |
| Self-employed admin: declarations, reminders, receipts register | Abby (an official Urssaf partner: declare and pay from the app, revenue filled from invoices, deadline reminders, receipts and purchases register, bank sync); Indy, Shine (unverified) | Invoices with French legal mentions: built (projects.md). Declarations as tasks and the receipts register: later (accounting.md). |
| AI breaks a task into steps; brain dump or voice into tasks | Goblin Tools; Tiimo's AI co-planner; Todoist Ramble (speak, tasks appear live, 40+ languages, beta); Saner.AI; Structured AI | Next: splitting by AI. Capture is one typed line; voice memos are recorded as audio only (tasks.md). |
| Tone help when writing and reading | Goblin Tools Formalizer (rewrite in a chosen tone) and Judge ("am I misreading the tone?") | Reading: the shield reads hostile tone on a public address (porch.md). Writing: not built. |
| An energy budget per task or day | Structured (Energy Monitor); Lunatask (energy and stress trackers); Bearable | Built without points: the day's weather and what a task takes (tasks.md). |
| A visual day; routines played step by step | Tiimo, Structured (timeline), Routinery | Built: Tasks ▸ The day, and Routines (tasks.md). |
| Scheduling into the calendar by itself | Motion; Reclaim (focus time, habits, buffer and "decompression" time after meetings); Akiflow; Structured AI | Refused: the plan proposes days and writes nothing into tasks or calendars. |
| Mood, symptom and energy journals | Tiimo, Daylio ("Year in Pixels", achievements), Bearable (correlations, summaries for appointments), Lunatask, Finch | Refused: tracking is work (finding 11 of research.md). |
| Habits, streaks, trophies | TickTick, Lunatask, Tiimo, Habitica, Finch, Daylio | Refused. |
| Company while working | Focusmate (booked 25/50/75-minute video sessions); Inflow (co-working rooms) | Next. |
| Focus and rest sounds | Brain.fm (vendor), Endel (soundscapes adapting to heart rate, weather and time, vendor) | Built: noise and nature made by Sioul, and your own recordings (sounds.md). |
| Someone, or an AI, does the admin for you | Duckbill (human assistants with AI; $99/month for about 4–6 tasks); Carly (an AI with its own e-mail address that replies, books and follows up on its own, vendor); Ohai; Google "Ask for Me" (AI calls businesses, US) | The AI drafts and never sends (ai.md). Handing a case to a person: designed ("hand to my helper", design.md), not built. |
| Phone calls made easier | Apple Hold Assist and Call Screening (iOS 26); Rogervoice (French live captions of calls, type-to-speak, for deaf and hard-of-hearing users); Google Ask for Me | A "call" kind, offices' hours, "Call" on a contact: built. |
| Mail turned into tasks, one inbox | Akiflow, Saner.AI, Ohai, Carly | Built: tasks, events and notes from a message; the Porch; projects. |
| Time needed against time available | Shovel ("the Cushion"); Sunsama (workload threshold); Llama Life (projected finishing time) | Built: a time budget before a date asked (tasks.md). |
What recurs and Sioul refuses on purpose: streaks, trophies, mood journals, scheduling that moves things, AI that acts alone (section 4).
Sources for the table: https://www.tiimoapp.com/product · https://coolcuration.com/structured-app-review-uk (review) · https://help.structured.app/en/articles/1897986 · https://2sync.com/blog/ticktick-vs-todoist (review) · https://www.lunatask.app · https://snapcraft.io/lunatask · https://en.fileee.com/features · https://apps.apple.com/us/app/-/id6757674090 (BriefAI) · https://moneytothemasses.com/banking/snoop-app-review (review) · https://www.ohai.ai/features/ai-reminders · https://github.com/paperless-ngx/paperless-ngx · https://kx.cloudingenium.com/en/paperless-ngx-self-hosted-document-management-ocr-tagging/ (review) · https://www.laposte.fr/digiposte/famille · https://hunted.space/product/clario-4 (snippet) · https://getmika.de/en/blog/no-more-bureaucratic-german-mikas-bureaucracy-translator-explains-any-official-letter · https://www.appgoblin.info/apps/6756278057 (snippet) · https://goblin.tools/About · https://apps.apple.com/us/app/-/id6746690443 (LetterSnap, snippet; the page returned 404 when opened) · https://www.totalbug.com/?p=14340 (review) · https://aide.laposte.fr/contenu/qu-est-ce-que-la-lettre-recommandee-en-ligne · https://www.laposte.fr/tarifs-postaux · https://www.lafabriquedunet.fr/logiciel/ar24 (review) · https://finanzwissen.de/anbieter/finanzguru/test/ (review) · https://www.neuebanken.de/finanzguru/ (review) · https://www.fincomparelab.com/reviews/rocket-money-review/ (review) · https://www.inc-conso.fr/content/tout-savoir-sur-la-resiliation-en-3-clics-avec-la-cnafc · https://selectra.info/finance/guides/compte-bancaire/comparatif-agregateurs · https://www.boursedescredits.com/guide-linxo-fonctionnalites-avis-agregateur-comptes-6422.php · https://woob.tech/applications/bank · https://www.lafabriquedunet.fr/logiciel/kresus · https://actualbudget.org/docs/advanced/bank-sync/gocardless/ · https://folio.id/id-wallet-app/ · https://mwm.ai/fr/apps/expiryguard-document-reminder/6777413276 (snippet) · https://partenaire.dossierfacile.logement.gouv.fr · https://www.economie.gouv.fr/particuliers/simulateur-droits-prestations-sociales · https://www.data.gouv.fr/datasets/fiches-pratiques-et-ressources-de-service-public-gouv-fr-particuliers · https://www.service-public.gouv.fr/particuliers/vosdroits/R11193 · https://abby.fr/auto-entreprise · https://www.todoist.com/help/articles/from-voice-to-tasks-ramble-july-1 · https://www.producttalk.org/building-todoist-ramble-how-doist-turned-voice-braindumps-into-real-time-task-capture/ · https://www.saner.ai/ · https://reclaim.ai/buffer-time · https://pipeline.zoominfo.com/sales/reclaim-ai-features (review) · https://appshunter.io/ios/app/daylio-journal/id1194023242 · https://apps.appfollow.io/ios/bearable-symptom-tracker/1482581097?country=ca · https://www.focusmate.com/how-it-works/ · https://www.brain.fm/blog/best-focus-music-apps-2026 (vendor) · https://endel.io/adhd · https://lp.getduckbill.com/blog/introducing-duckbill-the-ultimate-to-do-tackler · https://www.usecarly.com/blog/best-ai-apps-for-life-admin/ (vendor) · https://macrumors.com/2025/01/30/google-search-ask-for-me · https://www.att.com/device-support/article/142470/apple/iphone-11/ · https://www.carenews.com/fr/news/l-appli-rogervoice-retranscrit-les-conversations-telephoniques-pour-les-personnes · https://apps.apple.com/us/app/-/id1467742357 (Shovel) · https://dormway.app/blog/dormway-vs-shovel (review)
3. What the evidence asks of Sioul, and where it stands¶
| Feature | Why | In Sioul |
|---|---|---|
| A money watch from bank data: the bank's statements (CSV, OFX, CAMT.053) imported; a debit or standing order that did not happen said; a warning before the balance falls short of the next debits; the week's payments in the Porch's summary | Payment reminders that ignore the balance push people into overdraft (Medina 2021); people stop looking when accounts go bad (Olafsson & Pagel 2017); arrears grow with age in ADHD and rise before suicide (Beauchaine, Ben-David & Bos 2020); debt and depression, OR 2.77 (Richardson et al. 2013) | built (accounting.md, "The bank"); fetching from the bank: not built |
| Paper letters into the Porch: a scan, a phone photo or a PDF dropped in a watched folder → OCR → sender, amount and dates read → a card in the next admin window; the image filed | Opening post is part of admin anxiety (Money and Mental Health 2018); people avoid information they expect to hurt (finding 4 of research.md) | built (porch.md); a mail intake address for a helper: not built |
| The letter card: what it is, what it asks, by when, as a date (the French legal mentions: "délais et voies de recours", "à compter de la notification", "mise en demeure", registered or not), the amount, what happens if nothing is done, and "is this genuine?" (the office's domain, IBAN, references); the date proposed as a task | Time-based remembering fails, cue-based much less (Landsiedel et al. 2017; Altgassen et al. 2014); uncertainty fuels anxiety (finding 5 of research.md); clearer material and the amount shown raised take-up by 6–8 points (Bhargava & Manoli 2015) | partly: who, what, how much, by when (a delay counted from the notification), registered or not, and "A task for that date": built; what happens if nothing is done, and the genuineness checks: not built |
| Quiet reminders that reach the person: with the window closed, one silent notification per date asked, appointment or answer due, never repeated; tied to an event where possible | Prompting technology d = 1.27 (Jamieson et al. 2014); text reminders raise attendance (Gurol-Urganci et al. 2013); a date and time written raises follow-through (Milkman et al. 2011); event cues beat clock reminders for habits (Stawarz et al. 2015) | built (reminders.md); not on Windows yet; event cues: next |
| Hand a case to a helper: a one-page dossier per case (what it is, where it stands, what the helper is asked to do, the dates, the documents and drafts) as PDF, print or e-mail; a fixed "I'm stuck on X" message | Autistic inertia eased by "external intervention" (Buckle et al. 2021); supporters are part of good care (Nicolaidis et al. 2015); shared planning helps some and intrudes on others, so it is per case (Chen, Meng & Nie 2026) | not built ("hand to my helper" is designed, design.md) |
| Procedure kits: a case made from a kit with its steps, waits, documents to gather, offices and drafts, for common French procedures; steps, forms and online services read from service-public.fr's open data (XML, daily, Licence Ouverte) | Learning costs are one of the three costs of administrative burden (Herd & Moynihan 2018); a third of eligible households do not claim the RSA (DREES 2022), lack of information and complexity being the main reasons given (unverified figures) | not built; sioul tasks import turns a TOML plan into tasks (tasks.md, "Importing") |
| A postal outbox: a formal French letter laid out as PDF (sender, address placed for a window envelope, "Objet", "Pièces jointes"), printed or sent online (La Poste's online registered letter, Merci Facteur, AR24); the proof of posting and the acknowledgment filed in the case | Phone and post are where admin stalls (Money and Mental Health 2018) | not built |
| Waiting, with legal delays and a follow-up: "waiting since …, an answer is due by …" from known delays (one month for a GDPR access request; eight days for medical records, two months if over five years old; two months of an administration's silence as an implicit decision); a polite follow-up drafted when the delay passes | Uncertainty (finding 5); a waiting board that says what can be done meanwhile (Choiceworks, in tasks.md) | partly: waits after a step ("wait 14 days for the answer") and their reminders: built; legal delays and drafted follow-ups: not built |
| A papers wallet with validity: typed papers with their validity or a "less than three months old" rule; attached from the wallet; renewal planned early (passport and ID take weeks; the CSS must be asked again 2–4 months before it ends and does not renew by itself) | Time-based remembering (Landsiedel et al. 2017) | built (papers.md) |
| A written-channel kit: an accommodations sheet kept and offered when writing to a care service or an office for the first time; a "please answer me in writing" draft; for unavoidable calls, a call card (what to say first, reference numbers, questions, a typed log) whose outcome becomes tasks and a confirmation e-mail | E-mail first, the phone last, for strangers (Howard & Sedgewick 2021); 62 % against 16 % struggle to book by phone (Doherty et al. 2022); 54 % against 32 % for phone admin (Money and Mental Health 2018); the AASPIRE report found useful by over 94 % (Nicolaidis et al. 2016); speech can fail under load (Zisk & Dalton 2019) | not built; the "call" kind, offices' hours and "Call" on a contact are |
| The companion in the app: a letter explained line by line, a task split at a chosen granularity, a reply drafted into the outbox, the tone checked before sending; per-case consent, masking, a log | Autistic workers liked an LLM for communication advice, some of it doubtful (Jang et al. 2024); used for breaking tasks down and summaries, with worries about accuracy (Carik et al. 2025) | designed (ai.md); only the shield's reading of a public address is built |
| Contracts and subscriptions: what one is bound to, with renewal and notice dates, what each insurance covers, how to cancel, a cancellation letter drafted; linked to presets and bank lines | Forgotten subscriptions are part of the "ADHD tax" (community term, unverified as a quantity) | built (accounting.md) |
| Rights, and the cost of waiting: a link to mesdroitssociaux.gouv.fr and the CAF's simulator from a case; on an aid request, one line on what each month of waiting costs when the aid is never backdated | The amount shown raised take-up by 8 points (Bhargava & Manoli 2015); non-take-up is large in France (DREES 2022) | not built |
| Appointments from mail: confirmations and letters with a date become events with the place, the questions to ask, the accommodations sheet and a reminder | Reminders raise attendance (Gurol-Urganci et al. 2013); personal information sent before the visit (Nicolaidis et al. 2015) | partly: .ics invitations and a paper letter's appointment ("Into the agenda"): built; dates in plain-text mail: not built (roadmap.md, slice 5) |
| A small phone companion: capture a line, photograph a letter, receive the quiet reminders, see the one next step; no Porch, no mail list, no feeds | Every competitor is phone-first; checking mail less often lowers stress (Kushlev & Dunn 2015) | not built (roadmap.md, slice 12) |
| What a task takes, and the day's room: light, normal, heavy; a day that refuses to overfill and says so in words; rest as a planned deposit | Burnout recovery rests on fewer demands (Raymaker et al. 2020; Higgins et al. 2021); energy accounting is clinical, untested (Toudal & Attwood) | built (tasks.md) |
| A running time cushion before a date asked | Time made visible (finding 18 of research.md) | built (tasks.md) |
| Voice memo to tasks: a memo transcribed on this computer and proposed as task lines | The cost of capture decides whether thoughts are kept (Todoist's reviewers, in tasks.md) | partly: memos are recorded as audio only |
| A crisis plan as a private note: a template the person fills and holds, on this computer only, never proposed, never read by an AI, sent nowhere | Safety-planning interventions lowered suicidal behaviour, RR 0.57 (Nuij et al. 2021) | not built |
| Watch data as a hint: the watch's readings shown beside the day's weather, never changing the plan by itself | Tools that seem to watch are rejected (Chen, Meng & Nie 2026); tracking is work (Rudberg Selin et al. 2026) | built: the watch's readings on the Health page, and a morning line on the Tasks page that offers a lighter day; the plan never changes by itself (health.md, "Your watch"; wearables.md) |
4. What to refuse, and why¶
| Refused | Seen in | Evidence or reason |
|---|---|---|
| Streaks, trophies, points, achievements | Tiimo, Daylio, TickTick, Lunatask, Habitica, Finch | Broken streaks lower later engagement, more with self-blame (Silverman & Barasch 2023, J Consumer Research); expected tangible rewards undermine intrinsic motivation (Deci, Koestner & Ryan 1999, Psychological Bulletin, meta); Habitica's damage spirals (tasks.md). |
| Counts of what was missed, red "overdue", postponement counters | Todoist, Marvin, Sunsama, most apps | Self-forgiveness lowers later procrastination (Wohl, Pychyl & Bennett 2010); shame tracks depression more than guilt (Kim, Thibodeau & Jorgensen 2011, meta); adults with ADHD meet more criticism, and low self-compassion goes with poorer mental health (Beaton, Sirois & Milne 2022). For money too: no red balances, no "late payment" badges. |
| Repeating "nag" reminders, many notifications | TickTick, reminder apps | A notification costs attention even unanswered (Stothart, Mitchum & Yehnert 2015); batching beats constant alerts (Fitz et al. 2019); a reminder can push a cost elsewhere (Medina 2021). One quiet reminder per date, inside working hours; insistence only per item and by choice. |
| Mood, symptom and energy journals, correlations, "insights" | Bearable, Daylio, Tiimo, Lunatask | Self-tracking makes "substantial interpretive and emotional demands" on neurodivergent people (Rudberg Selin, Unéus & Knudsen 2026); people stop tracking from effort and guilt (Epstein et al. 2016); median 30-day retention of mental-health apps is 3.3 % (Baumel et al. 2019). |
| Inferring capacity or mood from behaviour or a watch, and changing the plan by itself | CSCW concepts; Endel's heart-rate adaptation | "It's watching me": adults with ADHD rejected nudges based on patterns and liked a "brain weather" they set themselves (Chen, Meng & Nie 2026). The day's weather stays the person's word; watch data, when shown, only as a hint (wearables.md). |
| Scheduling that moves things without asking | Motion, Reclaim, Structured AI | "If losing control triggers anxiety, avoid fully automated systems" (Morgen's ADHD guide); Motion users found days "overscheduled" (tasks.md); predictability is Sioul's first rule (findings 1 and 5 of research.md). |
| AI or services that act alone: send, book, pay, cancel, call | Carly, Google Ask for Me; the AI side of Duckbill | Trauma-informed computing puts safety, trust and agency first (Chen et al. 2022, CHI); language-model advice to autistic workers was sometimes questionable (Jang et al. 2024); 75–80 % of ADHD and dyslexic users worried about accuracy (Carik et al. 2025). The AI drafts; the person sends. A human helper acting with the person's consent is another matter. |
| Commercial nudges: "better deals", switching offers, bill negotiation for a share of savings, upsells | Snoop, Finanzguru, Rocket Money | Each offer is a new decision on a tired day; choice overload appears when options are hard to compare and effort is to be saved (Chernev, Böckenholt & Goodman 2015, meta); a sales voice does not belong in a calm space. Facts about one's contracts, yes; offers, no (accounting.md). |
| Reminders to keep in touch with people | Lunatask (relationship tracking), family assistants | A reminder to contact someone is a demand; demand avoidance rests on anxiety and intolerance of uncertainty, and adults with PDA ask for autonomy (White et al. 2022; Kenny & Doyle 2024); such nudges can also push contact a person has decided against. Never by default. |
| A phone app that brings the Porch, the mail list or feeds to the phone | most apps are phone-first | Checking mail less often lowers stress (Kushlev & Dunn 2015). If a phone app comes, it captures, scans and reminds, nothing else. |
| Animation, confetti, colour bursts | Llama Life, many apps | Irrelevant animations distracted autistic users more and raised mental demand (Uitdenbogerd, Spichkova & Alzahrani 2022); W3C COGA (2021). |
| Forms to keep up by hand | Finanzguru and wallets done by hand, Marvin's setup | Tools are dropped when they need upkeep (Epstein et al. 2016; Baumel et al. 2019); adults with ADHD spend "more time organizing the to-do app than doing the task" (Chen, Meng & Nie 2026). Every new place fills itself from mail, scans or bank files and asks only for confirmation. |
| Courses, tips feeds, daily lessons | Inflow | Median use was one module finished in seven weeks (Inflow's feasibility study, in tasks.md); one sentence where it is needed instead. |
| A crisis feature that detects, alerts or contacts anyone | safety apps | Only a plan the person writes and holds; Sioul infers nothing from behaviour (tasks.md, rule 7) and sends nothing without the person (design.md). |
| Public sharing of statistics | Daylio ("share statistics with your friends") | Privacy and trust (trauma-informed computing, Chen et al. 2022); nothing in Sioul is shown to anyone unless the person sends it. |
5. French rules used¶
- The free health cover (CSS) does not renew by itself; it is asked again 2–4 months before it ends: https://forum-assures.ameli.fr/questions/2906870-renouvellement-complementaire-sante-solidaire
- Medical records within eight days, two months if over five years old (Code de la santé publique, L1111-7): https://affairesjuridiques.aphp.fr/textes/acces-au-dossier-medical/
- An administration's two-month silence, acceptance by default, and its many exceptions: https://www.justice.fr/fiche/regle-silence-vaut-accord-sva-demandes-concernees
- GDPR access requests answered within one month (art. 12(3)): https://eur-lex.europa.eu/eli/reg/2016/679/oj
- Online cancellation "in three clicks" since 1 June 2023 (loi 2022-1158): https://www.inc-conso.fr/content/tout-savoir-sur-la-resiliation-en-3-clics-avec-la-cnafc
- "Je change de coordonnées": https://www.service-public.gouv.fr/particuliers/vosdroits/R11193
- Service-public.fr's practical sheets as open data (XML, daily, Licence Ouverte 2.0): https://www.data.gouv.fr/datasets/fiches-pratiques-et-ressources-de-service-public-gouv-fr-particuliers
- mesdroitssociaux.gouv.fr (58 aids): https://www.economie.gouv.fr/particuliers/simulateur-droits-prestations-sociales
- DossierFacile: https://partenaire.dossierfacile.logement.gouv.fr
- La Poste's online registered letter and its 2026 prices: https://aide.laposte.fr/contenu/qu-est-ce-que-la-lettre-recommandee-en-ligne · https://www.laposte.fr/tarifs-postaux
6. Not confirmed¶
- The DREES survey's reasons for non-take-up (37 % information, 22 % complexity, 17 % fear) and the ASPA's 50 % non-take-up: secondary sources only.
- LetterSnap, Clario, Briefed, Bureaucracy Buddy, Expiry Guard: search summaries, pages not opened.
- Indy's and Shine's Urssaf features; whether Merci Facteur or La Poste offer an API for sending from another application.
- The "ADHD tax" as a measured quantity.