A new hire reaches independence in days, not months.
Onboarding used to mean shadowing a veteran for a season and hoping enough rubbed off. The veterans are retiring, there's no one left to shadow, and the labour market can't supply replacements fast enough. Dimbo gives every new hire the expert colleague that never leaves — a role-scoped assistant that answers "how do I do this" from your live company knowledge, filtered to exactly what their role should see.
The old way to onboard was to stand next to someone who'd done it for thirty years. That person is walking out the door.
Time-to-productivity has always depended on a senior with the patience to be interrupted. But the mid-market backbone has the oldest workforce in Europe, led by the oldest owners in Europe — and the ones who used to carry a new hire through the first six months are retiring on a fixed timetable. Onboarding can't lean on a resource that's disappearing.
Stop routing every question through the one person who's busiest. Route it through the graph.
A new hire's real onboarding cost isn't the training day — it's the hundred small interruptions afterward, each one pulling a senior off the line. Dimbo turns that flow into an assistant grounded in your live knowledge graph: the same procedures, how-tos, and captured judgment your best people already put there, served back on demand, in the new hire's language, scoped to their role. The senior's time is freed; the newcomer never waits.
The colleague who never clocks out
Night shift, no supervisor in the building — the assistant answers at 3 a.m. as well as it does at noon.
Answers from your knowledge, not the open web
Every answer is retrieved from your own KnowledgeDoc store and company graph — the procedures and how-tos you actually run, never a generic chatbot's best guess.
It speaks the way the floor speaks
Language is enforced per client from the evidence in your own traffic — an Italian operator asks and is answered in Italian.
Every seat gets its own expert — and sees only what it should.
The assistant isn't one chat box bolted on the side. It's grounded in the live company graph, filtered at the moment of retrieval to the role of the person asking, and it gets smarter every time someone uses it.
The expert colleague, always available
A new hire asks in plain language — "how do I set up this fixture," "what's the release check for this batch," "who do I call when the line stops" — and the assistant answers from live company knowledge: the how-tos, the captured voice notes, the runbooks, the graph that ties a machine to its part to its procedure. No senior interrupted, no answer invented.
Operator knowledge, never the cap table
Access is filtered where it matters — at the moment knowledge is retrieved, not after. An operator sees operator knowledge; the cap table, the board minutes and the payroll simply never enter the answer.
The screen fits the job
Every level of the hierarchy gets its own home — from a line operator's tablet work-order board to a plant manager's command view — capability-gated so each person sees the surface built for their role.
Every question a new hire asks also enriches the company
When a newcomer asks something the graph can't fully answer, that gap doesn't vanish — the hungry assistant (knowledge_hunter) registers it and asks a senior to fill it, in their language, before the next hire needs it. Onboarding stops being a drain on knowledge and becomes the thing that builds it. The more the assistant is used, the better it gets.
The new hire asks. The company answers.
A line operator three days into the job needs a setup they've never done. Instead of hunting for a supervisor, they ask — and the answer comes back grounded in a how-to and a veteran's own captured note. Representative company, real mechanism.
An answer, not a wait
- Retrieved from your own knowledge — never the open web.
- Filtered to operatore_linea at retrieval, not after.
- Answered in the language the floor actually speaks.
The same platform, a different first day for every role.
A new hire lands on the home built for their job — nothing to configure, nothing they shouldn't see. Each surface is capability-gated, so onboarding is the same story at every level of the hierarchy: here's your screen, here's your assistant, ask it anything.
Line operator
A tablet work-order board — today's jobs, the how-to for each, and the assistant one tap away. The whole shift on one glanceable screen.
- tablet board
- how-tos
- ask the company
Shift lead
The line and its alerts together — what's running, what's slipping, what needs a call — so a new lead sees the department the way a veteran would.
- line status
- alerts
- work orders
Plant manager
A command view of the plant — state, decisions and exceptions in one place, so a new manager knows where to look on day one.
- plant state
- exceptions
- decisions
Quality manager
Release checks, audit trail and the procedures behind them — the quality picture, with the evidence one query away.
- release checks
- audit trail
- procedures
Owner · GM
The command view — state, decisions and the exceptions that need an owner — the whole company legible from one screen.
- state
- exceptions
- the ask
Governed by construction
Every surface — and every answer the assistant gives — is filtered to the role at the moment of retrieval. The wrong knowledge never reaches the wrong seat.
- capability-gated
- per role
- at retrieval
The assistant carries what there's no longer time to shadow.
Onboarding stops being a one-way cost. Each question the new hire asks either has an answer already — or becomes the answer the next hire inherits.
A new hire asks
Plain language, in their own tongue — "how do I do this?"
The graph responds
Retrieved from live company knowledge, RBAC-scoped to their role.
Or the gap is caught
No answer yet? knowledge_hunter asks a senior to fill it.
The next hire inherits it
Answered once, kept forever — the base of knowledge only grows.
Half the modelled annual know-how loss is recovered when capture and the role-scoped assistant work together — ≈ €24k/yr on the reference firm, before counting a faster, cheaper first six months for every new hire. (Dimbo value model §4)
The Silent Retirement
The full analysis behind this page: ageing, succession, and the vanishing know-how of Italian industry — why "hire and train" can't keep pace, and why capturing knowledge into a role-scoped assistant is what makes a new hire productive before the veterans are gone. With the sources, in full.
Read the white paper →Business continuity & succession
The other half of the story — capturing the veteran's knowledge before they walk out the door.
Keep the know-how →Grounded, sovereign, governed
The assistant answers from your own knowledge, inside your own walls, filtered to each role. No open-web guessing, no data leaving the building.
Give your next hire a colleague who never leaves.
The free Deadline Audit lands value in 48 hours — and it's your first look at the platform that turns every new hire's first months from a drain into a head start. On your data, on your infrastructure.