PlanlyAI
PlanlyAI builds fair staff rosters for hospitals, care homes and any shift-driven team. A classic planning engine generates a whole department's month in one click, while Flari, its beta AI assistant, handles individual edits and approvals conversationally.
What is PlanlyAI?
PlanlyAI is a web platform for building staff rosters, published by BeKa Software s.r.o. of Bratislava and aimed first at hospitals, clinics and social care facilities, then at any organisation running irregular shifts. It is built from two clearly separated parts, and the vendor insists on the distinction: a conventional planning algorithm lays out the shifts across the whole period, and Flari, an AI assistant in beta, handles individual actions on top of it. Flari does not create the schedule and, in the vendor's own words, does not even start the algorithm.
The planning engine takes in shift templates — morning, day, night, split or fully irregular — plus the department's own rules, and produces an optimised month for the whole unit in under a minute. It weighs qualifications, availability, time-off requests, hours already worked, recorded preferences and labour-law limits, and it deliberately ignores age, gender, family status, the reason for a sick leave, performance ratings, salary, ethnicity and religion. Before anything is published, the tool flags conflicts, broken rules and coverage gaps; afterwards, a mid-month rebuild is possible without discarding the rest of the period.
Flari sits beside that engine as a chat window. A team leader types a request in free text — move this person from Saturday to Sunday — or uses a slash command, and Flari prepares the edit, answers questions from the roster data, or approves a time-off request. Every output is a proposal: an approval screen stands between the assistant and the live schedule, and a human clicks through it each time.
Around the two sit the working parts of a workforce tool: teams, working units, expertise categories, mentoring pairs, live counts of hours, weekends and overtime, PDF and Excel exports, a payroll export, embedded PDF manuals and manual overrides at any point. The application runs in the browser as an installable web app, with its back end in Sweden and its database replicated between Sweden and Slovakia. Named references are Slovak: two university hospitals, a national paediatric hospital, a cardiology centre and one social care facility in active deployment. The product also takes part in TEF Health, the European facility for testing AI in real healthcare settings.
What it does
- Generate a full department roster in a single click
- Rebuild the schedule mid-month after sick leave without losing the rest of the period
- Flag conflicts, rule breaches and coverage gaps before a roster is published
- Apply the department's own rules, rest periods and fair-distribution constraints
- Ask Flari in plain language to move, swap or approve a specific shift
- Count hours worked, weekends and overtime as the roster changes
- Export the finished plan to PDF, Excel or payroll, and override anything by hand
When to use PlanlyAI / When not to
A quick filter to help you decide if PlanlyAI is the right fit.
When to use PlanlyAI
- Ward and nurse managers who rebuild a department roster every month and lose days to it
- Social care facilities running 24/7 cover with a small pool of carers and strict rotation rules
- Emergency services and security firms that must enforce rest periods across continuous shifts
- Manufacturing sites juggling multi-shift production with irregular patterns and precise hour counts
- Retail, restaurant and service managers scheduling part-time staff around fluctuating opening hours
When not to use PlanlyAI
- Buyers who need a public price list: no rate is published anywhere, and the vendor's pricing page sits behind a password
- Teams wanting to sign up and start alone: the only entry point is a booked 30-minute demo, with no self-service trial
- Developers looking to script rosters: there is no public API and no API documentation of any kind
- Staff who want to check their shifts from a phone app: the product ships as a web application only, with no App Store or Play listing
- Organisations needing a contract pack up front: no terms of service, privacy policy or DPA is published on either of the vendor's sites
How to use PlanlyAI
A typical end-to-end flow, from setup to results.
- Book the 30-minute demo through the form linked from the website; there is no self-service sign-up
- Agree the department rules with the vendor, who implements the specific policies the unit runs on
- Receive access to the web application and sign in; the tool installs as a browser app, with nothing to deploy
- Work through the built-in onboarding and the sectioned PDF manuals inside the product
- Set up the organisation: working units, teams, expertise categories and mentoring pairs
- Build the shift templates the department uses — morning, day, night, split or irregular
- Import or record staff availability, time-off requests and individual preferences
- Generate the period's roster in one click and read the alerts on conflicts, rule breaches and gaps
- Adjust by hand, or ask Flari in plain language to move, swap or approve a specific shift, then approve its proposal
- Publish, then export to PDF, Excel or payroll, and rebuild mid-month whenever an absence changes the plan
Pros & Cons
Pros
- EU data residency described down to the country: back end in Sweden, database replicated in Sweden and Slovakia, and the language model run in an EU region
- An explicit commitment that the vendor's LLM providers do not train on customer data
- An unusually detailed AI transparency notice: EU AI Act classification, articles cited, and a published list of what the algorithm does and does not take into account
- Human oversight built into the workflow — no AI proposal is written to the live roster without an approval click
- Named reference customers, including two university hospitals and one care facility in active deployment
- Practical output for real payroll work: PDF, Excel and payroll exports, with manual override kept available
- Participation in TEF Health, the European testing facility for healthcare AI
Cons
- No public pricing at all: the Slovak site's pricing page is protected by a password, and no amount appears anywhere else
- No terms of service and no privacy policy on either website, so every compliance fact rests on an in-app page the vendor can change at will
- No DPA offered and no named subprocessors, beyond a plural reference to the vendor's LLM providers
- No retention period is given, despite a section titled What is logged and for how long
- No security certification is displayed — neither ISO 27001 nor SOC 2
- No public API, no API documentation and no mobile app: a browser is the only way in
- Reference customers are almost entirely Slovak, and the AI assistant is still labelled beta with limited release
Pricing & Plans
No price is published. The vendor's Slovak website carries a pricing page, but it is protected by HTTP authentication and returns an unauthorised response to visitors; no amount, currency or billing unit appears on the English site or anywhere in the application either. There is no advertised free plan and no free trial. The only stated commercial entry point is a free 30-minute demonstration, described as involving no commitment and no installation, after which terms are agreed directly with the vendor.
Data, GDPR & hosting
A consolidated view of how PlanlyAI handles your data.
GDPR overview
There is no published privacy policy on either website. The GDPR material comes from an in-application transparency notice, which is nonetheless specific: processing is declared under Regulation (EU) 2016/679, with performance of a contract and the employer's legitimate interest as lawful bases under Article 6(1)(b) and (f). It names a right of access under Article 15, a right to object to automated decision-making under Article 22 (satisfied, the vendor argues, because a human approves every change), and a right to have the AI features switched off on request. The same page classifies the assistant as a limited-risk system under the EU AI Act while voluntarily applying high-risk safeguards. No DPA, no named subprocessors and no Article 27 representative are mentioned; none is required, the publisher being established in Slovakia.
Who owns the data?
No terms of service or privacy policy is published on either of the vendor's websites, so ownership is not stated in a contract. What the vendor does publish is an in-application transparency notice, and it positions the employing organisation as the party that controls the data: the stated lawful bases are performance of a contract and the legitimate interest of the employer, under Article 6(1)(b) and (f) GDPR. Staff are treated as data subjects rather than owners, with a right of access under Article 15 and a right to object to automated decision-making under Article 22. Requests about personal data go to the vendor's single published address. Because none of this sits in a contractual document, it can change without notice.
Reuse rights
The vendor states that data is retained exclusively for audit and operational purposes, and that its LLM providers do not use personal data to train models. It also publishes what the planning algorithm does and does not weigh: qualifications and certifications, availability and time-off requests, hours already worked, recorded preferences and labour-law limits are used; age, gender, family status, the reason for sick leave, performance ratings, salary, ethnicity, religion and health status are excluded, except where a qualification rule strictly requires it. Nothing published grants the end user a right to reuse or export the underlying data beyond the product's own PDF, Excel and payroll exports, and no reuse licence is described anywhere.
Data retention & training
Hosting summary
The vendor states that data does not leave the European Union. The application back end is hosted in Sweden, the database is replicated between Slovakia and Sweden — both EU locations — and the language model behind the Flari assistant runs in an EU region with a data residency guarantee. It adds that no personal data is transferred to the United States or to any other third country. Processing is declared under Regulation (EU) 2016/679, on the lawful bases of performance of a contract and the legitimate interest of the employer. The public marketing site itself is served from Bratislava by a Slovak hosting provider, not from a content delivery network. Two caveats matter: none of this appears in a contractual document, since the vendor publishes neither terms of service nor a privacy policy, and no subprocessor is named beyond a plural reference to the vendor's LLM providers.
Things to keep in mind
Risks and trade-offs to weigh before adopting PlanlyAI.
- Rosters decide who works nights, weekends and holidays: an optimiser that is trusted without review can entrench unfair patterns while looking neutral, and staff should be able to see why they were scheduled as they were
- The AI assistant is explicitly described as able to misread context and to produce convincing but inaccurate answers, so the approval step protects real people and should never be treated as a formality
- Every compliance commitment — EU hosting, no model training, logging, opt-out from AI features — lives on an in-application page rather than in a contract, so nothing here is enforceable as written
- No retention period is published even though the vendor documents keeping full transcripts of chats with the assistant, which in a hospital can carry sensitive detail about named staff
- Approving AI suggestions with a single click, month after month, is exactly the setting where human oversight quietly erodes into rubber-stamping
- With no published price, no terms and no DPA, a buyer's own procurement and data protection review has to reconstruct everything from scratch
- A separate social media scheduler also trades under the name Planly, including on the mobile app stores: check the publisher before installing or contracting anything
Setup & Integrations
Technical difficulty
Low for the customer, because the vendor absorbs the hard part. There is nothing to install: the product runs in a browser and can be added to a device as a web app. Onboarding screens and sectioned PDF manuals ship inside the application. The department's specific scheduling rules are implemented by the vendor rather than configured by the client, which removes the main source of difficulty but also makes the customer dependent on the vendor for changes. Expect the real effort to sit in preparing staff data, shift templates and unit structure, not in technical deployment.
Deployment
Supported languages
Behind PlanlyAI
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Does the AI actually build the schedule?
Can Flari change a published roster on its own?
Where is the data hosted?
Is my organisation's data used to train AI models?
What does the tool log about me?
How much does it cost?
Is there a mobile app?
Which languages does it support?
Can staff refuse to have AI involved in their schedule?
Who is behind the product?
Should you pick PlanlyAI?
PlanlyAI is a narrow, serious product rather than a general-purpose assistant. It solves one problem — laying out fair shift rosters for teams that work around the clock — and it solves it with a conventional optimisation engine, not with a language model. That honesty is the most striking thing about the tool: the vendor's own transparency notice states that the AI assistant does not build the schedule and does not even start the algorithm that does, at a moment when most competitors would claim the opposite. The same notice goes further than most, publishing which criteria the algorithm weighs and which it deliberately ignores, committing to EU-only hosting down to the country, and stating that no customer data trains any model.
The gap is commercial transparency. There is no published price, no terms of service, no privacy policy and no data processing agreement on either of the vendor's websites. Everything a compliance officer would want to read sits inside the application, on a page the vendor can rewrite at any time, and no retention period is given even where the heading promises one. No security certification is displayed. It reads less like bad faith than like a young company that invested in its product and its AI governance before its paperwork — but every contractual fact has to be obtained from the vendor directly.
Traction is real and concentrated: two university hospitals, a national paediatric hospital, a cardiology centre and one care facility in active deployment, all in Slovakia, plus a place in the European TEF Health testing programme. For a ward manager who rebuilds a roster every month, the proposition is concrete and the demo costs nothing. For anyone needing a signed contract pack and a published price first, the door is not open yet.
- Choosing a selection results in a full page refresh.
- Opens in a new window.