No tool
Calendar checked in the morning, the rest from memory.
- Set-up effort
- Benefit
- Risk
Gaps are discovered in the meeting.
Care.Teach.Research.
The method for holding the three academic-hospital missions together with AI: anticipate deadlines, triage what comes in, produce faster, and stay in control.
3joint missions
French Education Code, art. L952-21
French Education Code, art. L952-21
14management tabs
problem, options, method
problem, options, method
39step-by-step guides
with a starter prompt
with a starter prompt
What you can do
Twelve things to try this month
The same kind of use cases people share everywhere, rebuilt for academics who teach, do research and care for patients.
Review a thesis in one pass
- Drop the .docx and the brief
- Claude reviews substance, structure, style
- Get a suffixed version with tracked changes
A literature watch that makes slides
- Pick 3 to 5 topics
- Claude queries PubMed every week
- Tick “make the slide”
Search your own mail and files
- Connect mail and Drive, read-only
- Ask in plain language
- Get the answer with its source
Prepare a congress talk
- Give topic, length, audience
- Claude proposes a sourced plan
- Approve, then build to your template
Rehearse a thesis defence
- Ask Claude to play the jury
- It asks and follows up
- Honest feedback on your answers
An online department rota
- Describe activities and team
- Claude codes and tests the tool
- Publish in four clicks
A career file always up to date
- One single data sheet
- Claude collects every month
- CV, reports and applications derived
Routines while you teach
- Write a run plan
- The routine runs morning and evening
- You approve when you wake up
An anticipation dashboard
- List your deadlines
- Claude builds the page and its database
- D-30 and D-14 alerts
MCQs and OSCEs from your lecture
- Drop your slides
- Claude drafts MCQs and stations
- You review and adjust
Anonymised clinical cases
- Anonymise the letter locally
- Claude builds 4 slides to template
- Added to your collection
Build your own skills
- Describe the recurring task
- Claude writes the skill
- Trigger it with one sentence
ClaudeProjectsSkillsArtifactsPubMedGmailCalendarDriveWordPowerPointScheduled tasksClaudeProjectsSkillsArtifactsPubMedGmailCalendarDriveWordPowerPointScheduled tasks
In action
The system at work, in two views
Real tools, filmed on fictitious data: what gets connected, and what you can query without ever seeing a name.
The method
From idea to tool, in five stages
Every skill, dashboard and routine on this site was born from an everyday irritant. Here is the path, with its pitfalls.
- 1Ideathe problem in one sentence
- 2Conceptform, deliverable, safeguards
- 3Buildversion by version with Claude
- 4Testtests, gaps, edge cases
- 5Go livepublish, announce, turn it into a skill
Anticipate with a single dashboard
Daily plan, one-month and six-month horizons, D-30 and D-14 alerts.
Triage your mail by action
Four columns, two passes a day, drafts ready to copy.
Review without overwriting
Three levels, tracked changes, every reference checked in PubMed.
Getting started5 days, 5 stepsStartA literature watch that makes slides
10 to 20 references a week, sorted and ready to present.
Safeguards first
Human approval, protected health data, verifiable sources.
Sourced presentations
A slide-by-slide plan approved first, then built to your template.
Automate without drifting
Four routines, a budget, a status page.
The tabs
Every problem, every option
The tabs mirror the layout of a dashboard in real daily use. For each: the problem, four ways to handle it compared honestly, and the chosen method.
The problem
Daily roadmap
Every morning, the day is pieced together from the calendar, emails and memory: where the room is, which slides are ready, who is waiting for an answer.
- Days fragmented between ward, faculty and lab
- Documents found at the last minute
- Logistics (travel, rooms, video calls) never consolidated
Management options
The problem
Work in progress
Manuscripts, applications, department projects: a dozen projects open in parallel, some of which lie dormant without anyone noticing.
- Projects forgotten for lack of reminders
- No overview of what is moving
- What is "done" disappears instead of being logged
Management options
The problem
Documents & files
Documents live in Downloads, several drives, the institutional Drive and email attachments. AI can do nothing with what it cannot see.
- Multiple versions of the same file
- Sensitive data mixed with everything else
- Deep folder trees nobody follows
Management options
The problem
Mail & calendar
Clinical care, faculty, research, learned societies, industry: emails come from everywhere, and the real question is not how important they are but what to do with them.
- Replies waiting with no reminder
- Dates buried in the body of emails
- Predatory solicitations hard to spot
Management options
The problem
Looking ahead
A congress in three weeks, a thesis defence in ten days, a report due: each deadline lives in a different tool and is discovered the day before.
- No one-month or six-month view
- Preparation started too late
- Recurring deadlines (reports, bonuses, activity reviews) forgotten
Management options
The problem
Department & staff meetings
Rotas, staff meetings, laboratory, unit reports: running the department is invisible, shared work, and prone to errors.
- Conflicts between clinical sessions and academic duties
- Staff-meeting talks redone with no memory of past years
- Annual reports in imposed templates
Management options
The problem
Faculty & diplomas
Lectures, MCQs, OSCEs, assessment design, exam boards, university diplomas: requests arrive by email with short deadlines, often at the same time as a congress.
- Slides and MCQs made in a rush
- Student lists rebuilt every year
- Delegated duties without follow-up
Management options
The problem
Papers & theses
Manuscripts in several crossed versions, theses to review, contradictory co-author feedback, reference lists never checked.
- Substantive review buried under formatting
- Versions overwritten, history lost
- Inaccurate or untraceable references
Management options
The problem
Studies & cohorts
A closed cohort to write up, another in data collection, a multicentre study to set up, a congress deadline approaching: each project has its own timeline.
- Data collection that will miss the deadline
- Analysis plans not versioned
- Sensitive data mixed with working data
Management options
The problem
Congresses & talks
An invitation to speak in six weeks: you start from a blank page although similar topics have been presented ten times already. Then comes the logistics.
- Old decks lost or outdated
- Unsourced figures and numbers
- Tickets, programme and talk scattered
Management options
The problem
Requests & peer reviews
Congresses, podcasts, interviews, invitations to review or publish: you say yes to everything, dates overlap, and predatory publishers slip into the pile.
- Date conflicts settled too late
- Reviews at the "final reminder" stage
- Predatory solicitations hard to spot
Management options
The problem
Career
When an application, a bonus or a national sub-section file comes up, years of teaching, supervision and publications are rebuilt in a rush.
- CV never up to date
- Supervision and responsibilities forgotten
- Sub-section criteria poorly known
Management options
The problem
Skills & routines
The temptation is to create one scheduled task per need. Two months later: twenty-five tasks, six notifications a day, costs spiralling and silent failures.
- Instructions repeated in every conversation
- Scheduled tasks stepping on each other
- Failures nobody notices
Management options
The problem
Quality & self-learning
AI produces quickly, but an error that goes unnoticed repeats itself: a reference that cannot be found, a misspelt name, a date off by a day, a broken link. Human proofreading wears thin, and nobody knows any more what has been checked.
- Errors that recur from one deliverable to the next
- Long and uneven human proofreading
- No record of what was checked
Management options
The triptych
All guides, by mission
Pick your career stage: only the guides useful at that stage stay visible.
I am
Care & department
The hospital department: schedules, staff meetings, laboratory, quality, clinical database. AI prepares, you validate.
Organizing the department's staff meetingsList recurring staff meetings, keep the index of past talks, schedule the year's slots and spot topics never covered.Department schedule and organizationSpecify before building, cross-check assignments with the schedule without ever writing to it, spot conflicts every week.Preparing a staff meeting talkStart from the bank of talks already given, pick a topic never covered, and produce a sourced update in a few slides.Building a clinical caseFrom an anonymised discharge letter, a standard block of four slides: presentation, investigations, course, take-home message.Imaging: from export to teaching materialInventory an anonymized echocardiography or MRI export, sort the sequences, produce animated slides and quizzes.Annual activity report for a unitFill in the official template for an annual report (laboratory, unit, technical platform) from a log kept throughout the year.Putting together a department projectNew position, task delegation, equipment, team seminar: a presentation for hospital management and a medico-economic section, prepared in one session.Organizing a day or an eventPatient day, public event, training day: agreement, contacts, day plan, press and logistics in a single file.Laboratory protocols and proceduresReading protocol, report validation and distribution workflow, the team's typical week: write, version and distribute the department's rules.Requests for expert opinion and patient pathwaysRequests for opinion, referrals, multidisciplinary meetings: a clear circuit, response templates and follow-up, without ever circulating identifying data.Building the department's scheduling toolA shared schedule on a single web page: template week, one tab per activity, month view, sync for the whole team, published online. Built with Claude, step by step.Absences, freed-up rooms and stand-insWhen someone is absent, the room they were booked into becomes free: spot it automatically, suggest an available stand-in, and link absences to emails in both directions.Delivering, checking and publishing the scheduleLocal preview, automated triple test, double check of half-days and names, then publication in four clicks and online verification.No guide for this stage yet: see “Show all”.
Teaching
Lectures, MCQs, university diplomas, supervision: AI structures, schedules and produces the materials.
Organising the start of a DIU academic yearRebuild the mailing list, settle the calendar, prepare the emails in the right order: the start of year of a multi-site diploma in one session.Lectures, MCQs and examsThe chain of lecture, MCQs, exam-question review (docimology), exam and marking board, driven by D-30 and D-15 deadlines rather than by urgency.Preparing a thesis defenceDetect the defence at D-14, prepare the speaking notes for the supervisor or the jury president, and the report.Supervising a thesis or dissertationOne tracking sheet per student, dated milestones, versioned reviews: supervise several projects without losing track of any.Watching continuing education platformsNew webinars, modules and replays from continuing medical education platforms, collected every week and added to the dashboard.Delegating and tracking teaching assignmentsAssessment design, clinical semiology, tutoring, ECOS (OSCE): share assignments among junior faculty, appoint leads, follow the annual cycle without chasing people from memory.No guide for this stage yet: see “Show all”.
Research
Literature watch, theses, manuscripts, conferences, databases: AI speeds up every step without ever signing in your place.
Weekly literature watchA few thematic PubMed feeds, deduplicated and sorted every week: 10 to 20 references, and a "make the slide" checkbox.Literature review on a questionExhaustive PubMed search, tick-box sorting, retrieval of open-access full texts, a Word synthesis and one slide per reference.Preparing a conference presentationScoping, literature, inventory of your old decks, approved slide-by-slide outline, then building to your template and review.Reviewing a thesis or a manuscriptThree-level review in Word tracked changes, verified bibliography, arbitration of co-author feedback, always on a new version.Organising a research databaseFrom raw file to versioned database: disease-specific core datasets, a pseudonymised index, quality checks, and a query page that shows only aggregates.Organising a conference tripTickets, programme and talks gathered into an hour-by-hour itinerary, with the points to watch.Tracking your manuscripts through to submissionA single manuscript portfolio: role, current version, next step, pre-submission checks. No more papers stuck at V12.Peer reviews for journals and solicitationsAccept few, review well, decline fast: sort review invitations, spot predatory publishers, meet your deadlines.Managing your studies and cohortsOngoing data collection, site initiations, analysis plans, submission deadlines: one table per study and milestones planned backwards.Working on a single slideCheck a slide's reference, add one, bring it back to the template, extract a figure from a PDF: the unit task that feeds every presentation.No guide for this stage yet: see “Show all”.
The common foundation
Organizing your files for AIA single root folder, flat work folders, a typology: the condition for AI to find, file and reuse your documents.Setting safeguardsSix rules written once and for all, applied across all your uses: never overwrite, never delete, never send, never write without validation.Sorting your email by actionTwo email passes a day, sorted not by importance but by action: to reply, to do or to put in the calendar, to file, for information.The Reactor, an anticipation dashboardA single entry point that answers the question: what is coming up in the next few days, and what is ready?Scheduled routines without runaway costsFour orchestrators (morning, evening, weekly, monthly) driven by plans written in files, rather than twenty scattered tasks.Organising your skills into hats (roles)One orchestrator per role that routes to specialised building blocks: your know-how becomes reusable, by you and by others.CV, reports and career fileA single data sheet, updated as you go, from which the CV, the activity history and the annual reports are derived.Prioritising your invitationsConferences, podcasts, interviews, working groups: a calendar of invitations, date clashes settled early, and "no"s you can stand behind.Cross-checked inventory of your activityEvery two months, cross-check skills, dashboard, files and conversations to spot what is dormant, what diverges and what is missing.The quality loop: check, assess, learnA living checklist per deliverable, a double test with evidence, and a self-learning loop: every error found becomes a rule that Claude rereads before producing.
The career path
What AI changes at each stage
From residency to full professor, in medicine, pharmacy and dentistry, with the official texts.
1Resident / PhD studentOrganise, read, searchRoot folder, literature watch, literature review.2Fellow (CCA, AHU)Keep the paceStaff meetings, clinical cases, first lectures, mail triage.3Associate prof. (MCU-PH)StructureSupervision, manuscripts, career file, anticipation.4Full prof. (PU-PH)LeadDiplomas, department, skills and routines, monitoring.
Mailing list
Join the initiative
This site is a starting point. Choose one or more ways to take part.
- Receive new guidesA short letter whenever a guide or career path is published.
- Join the pilot groupApply the method for 3 months and share your feedback.
- Request a workshopFor a residency programme, a diploma, a doctoral school or a department.
- Contribute a guidePropose a method from your own discipline.
Contact · AI Campus
Get in touch with AI Campus
A question, feedback from your own practice, a workshop or partnership request: write to us, or follow updates on X.