Everything FitRank does today
44 features in 19 capabilities, all built and working, shown here on real screens with demo data. Each one keeps the final decision with a person and records why it was made.
Resource managers
Staff a task in minutes and see exactly why each person was suggested.
Plain-language task intakeType the role you need in plain words. FitRank turns it into skills, level, dates and location, and you confirm every field before matching runs.
Explained matching and shortlistsOne click ranks everyone who can take a task. Each person shows the rules they passed, the facts, the model's probabilities and the band they landed in.
Decisions, outcomes and learningManagers accept or reject with a reason, and check in six weeks later. FitRank turns those decisions into training data, but never switches itself on.
All 10 resource managers features
- Describe the task in plain words
- Type the task the way you'd say it. FitRank turns it into fixed fields (skills, level, start date, location) and you confirm them before anything runs.
- One-click matching
- A match run checks availability and hard rules, works out the facts for each person, and ranks everyone who qualifies. It runs in the background while you carry on.
- Shortlist, Review and Hidden bands
- Every person gets a probability, and your thresholds turn it into a band. Strong, agreed matches are shortlisted; anything uncertain or contradictory goes to Review.
- A reason for every suggestion
- Each person shows the five answers behind their score, a numbered list of facts, and an explanation that cites those facts. Nothing is written that the facts don't support.
- See who was left out, and why
- The excluded list names the rule each person failed: not free in time, on leave, wrong location, outside the cost band or missing clearance.
- Accept or reject with a reason
- Your decision is recorded with its reason (skill gap, level, domain, timing). Those decisions are what FitRank learns from next.
- Resume-backed skills
- The shortlist marks which must-have skills a person's resume actually supports, so you can tell claimed skills from evidenced ones.
- Ask HR when nobody fits
- If no one inside qualifies, send HR a hiring request for one to ten candidates, then mark each candidate they send as a fit or not.
- Six-week outcome check-ins
- For every placement, FitRank asks how it went: working well, struggling or released early. Real outcomes, not just clicks, show whether the picks were good.
- Business-unit context
- Each person shows their business unit and its head, so you know who to talk to. It informs you; it never filters anyone out.
Planning and delivery heads
See next quarter's staffing problems while there's still time to fix them.
Staffing dashboardOne screen for delivery leads: model agreement, profile readiness, open tasks with a shortlist, tasks by month and the next 12 months of bench idle cost.
Bench and roll-offsSee who is free now or within 30, 60 or 90 days, the weekly idle cost, and the open tasks each person could take next, with roll-off warnings.
Skill gap analysisFitRank marks every must-have skill across upcoming tasks as shortage, tight or covered, and shows who is closest, so you can train before you hire.
Staff several tasks at oncePick the tasks of one project and FitRank proposes one person per task from finished matching runs, with alternatives and the margin of each plan.
Hire or moveFor every task, FitRank compares the best internal person with the best outside candidate by fit, start date and margin, with the reasons for each side.
What-if staffing for pre-salesDescribe the roles and start date of a prospective project and see who could staff it from inside, with billing and margin per week. No task is created.
Employee value and client profilesEstimated revenue, clients, ratings and margin per person from your rate card, plus a printable client profile with no costs or ratings on it.
All 8 planning and delivery heads features
- Staffing dashboard
- One screen for how staffing is going: how often managers agree with the model, how many profiles are ready, which open tasks have a shortlist, tasks opened and filled each month, and the bench's idle cost for the next 12 months with the part open tasks could recover.
- Bench view
- Who is free now, or within 30, 60 or 90 days, what that idle time costs each week, and up to three open tasks each person could take.
- Skill gaps
- Every must-have skill across upcoming tasks is marked shortage, tight or covered, with the people closest to it, so you can train before you hire.
- Staff several tasks at once
- FitRank proposes one person per task across all open tasks without double-booking anyone, and shows the margin of each plan.
- Hire or move
- For each task, compare the best internal person with the best external candidate, side by side, with the reasons for each.
- Roll-off warnings
- See who is coming off a project soon and act early: hold them for a known task or plan upskilling.
- What-if staffing for pre-sales
- Check whether you could staff a deal before you win it. It uses rules only and creates no task.
- Employee value and client profiles
- Estimated revenue per person from your rate card (never salary), and a printable client-ready profile with no cost or ratings on it.
HR and talent teams
Keep people data clean and bring in outside candidates when the bench runs out.
HR home and hiring requestsManagers send HR a hiring request from a task with no internal fit. HR sees every request and candidate on one page and tracks each one to the end.
Privacy-first resume readingUpload a PDF or Word resume. Code removes names, contacts and personal attributes before AI reads it, and HR checks every field before it is saved.
Two-way candidate matchingScore an outside candidate against every open task, or a task against every candidate, with the same rules as employees and blockers in plain words.
Employee records and CSV importEdit employees, skills, leave and clients, see whose profile is missing or stale, and import from CSV or an HRMS export with a preview first.
All 10 hr and talent teams features
- HR home
- Open tasks, tasks with no internal fit, the candidate pipeline and candidates due for deletion, on one page.
- Employee and client records
- Add and edit employees, their skills, leave and business unit, and your client list. Every change is audited, and a data-health view shows what's missing.
- CSV import with mapping
- Import clients, employees and skills from a spreadsheet or an HRMS export. Preview first, map your own level, practice and skill names, and only valid rows are saved.
- Resume reading with privacy first
- Upload a PDF or Word resume. Names, contact details and personal attributes are removed in code before any AI reads it; the AI fills professional fields only and HR reviews before saving.
- Nothing guessed silently
- When a resume doesn't state notice period, location or level, the field is marked unconfirmed rather than filled in with a guess.
- Candidate pipeline
- Track every outside candidate from new to screened, contacted, interviewing and hired or not taken.
- Two-way candidate matching
- Score a candidate against every open task, or a task against every candidate, with plain blockers such as “can join 6 Nov; task needs 16 Oct”.
- Suggest a candidate to a manager
- HR can check a resume against a specific task and put a candidate forward without waiting to be asked.
- Skill suggestions from resumes
- Upload an employee's resume and FitRank suggests skills it found evidence for. HR accepts or dismisses each one.
- Consent and automatic deletion
- Candidate consent is recorded, and a candidate's file is deleted automatically one year after upload unless they're hired.
Admins and leadership
Control the model, prove it's fair, and keep a full audit trail.
Thresholds and model versionsSet your own shortlist and review cut-offs, compare model versions on unseen tasks, and switch or roll back in one click. Every change is versioned.
Health and fairness dashboardsSee how often managers agree with the shortlist, override rates by week and model, real placement outcomes, and recommendation rates tested for fairness.
Roles, sign-in and auditAdmin, resource manager, HR and viewer roles, hashed passwords and short-lived tokens, row-level security per company, and an audit CSV of every decision.
All 9 admins and leadership features
- Users and roles
- Admin, resource manager, HR and read-only viewer. Temporary passwords must be changed at first sign-in, and the last admin can't be removed.
- Your own thresholds
- Set the shortlist and review cut-offs for your company, with a reason. Every change is versioned and audited.
- Model versions with a safety gate
- Switch to a newer model, or roll back, in one click. A model that ranks worse than the current one is refused unless you override it on purpose.
- Health dashboard
- How often managers agree with the shortlist, override rates by week and by model, reject reasons and the outcomes of top picks.
- Fairness checks
- Recommendation rates by location and practice among qualified people, tested with the four-fifths rule and a significance test.
- What managers value
- The final scorer shown as plain factors, so you can see what your managers' decisions have taught it.
- Learning loop you approve
- Feedback is grouped into themes in a weekly report with names removed. Admins accept or dismiss each suggestion; nothing changes on its own.
- Rate card and margin target
- Rates per role and a margin target (30% by default) drive the value and planning views.
- Audit export
- A CSV of every recommendation and decision: the model and versions used, and what the manager did.
Platform and integrations
Built to run inside your company, on your data, at almost no running cost.
In-house decision modelMatching runs on a small transformer distilled from an open-weight LLM. It runs on an ordinary CPU, costs nothing per run and is calibrated per company.- Decision APIYour own services ask FitRank typed questions over an API and get a probability for every option, with per-project keys, rate limits and budgets.
All 7 platform and integrations features
- An in-house decision model
- A small transformer model, trained from an open-weight LLM, answers the matching questions. It runs on an ordinary CPU, so scoring costs nothing per run.
- Calibrated per company
- Each company gets its own final scorer, calibration and cut-offs, learned from its own managers' decisions.
- Nightly retraining that waits for you
- The scorer retrains on new decisions overnight and tells admins what changed. It never switches itself on.
- Multi-company by design
- Every record belongs to one company and the database enforces it with row-level security, so one company can never see another's data.
- Decision API
- Other internal tools can ask FitRank the same typed questions over an API, with per-project keys, rate limits, monthly budgets and a Python client.
- Secure sign-in
- Hashed passwords, short-lived tokens with a secure refresh cookie, lockout after failed attempts and a role check on every request.
- Runs in containers
- Docker images for the API, worker and web app, with a Postgres-only job queue. No extra infrastructure to run.
On the roadmap
Not built yet. Early-access customers help decide the order.
- Microsoft Entra ID single sign-on
- HRMS connectors (Keka first, then Darwinbox, Workday and SuccessFactors)
- Email and Microsoft Teams notifications
- Bulk resume upload
- Hosting in your own Azure tenant with Azure OpenAI
Be one of the first teams to staff with FitRank
We're taking on a small number of pilot companies. Pre-register for early access, or talk to us about an annual licence.