Updated: August 2026 · By Álvaro ArrescurrenagaCEO of Voicit
Starting at a new company is both exciting and overwhelming. The first few days determine much of what comes next: whether the person understands their role, feels supported, and becomes productive quickly. And yet, it's one of the things we do worst.
Artificial intelligence can be a great help here, but with one important caveat: It's not for automating the welcome process, but for freeing up time so you can dedicate it to people.In this guide you will see what is worth automating (and what is not), a 30-60-90 day plan, what tools exist, how to measure it and the most common mistakes.
What is AI-powered onboarding?
In practice, AI can personalize itineraries based on job roles, send documentation at the right time, instantly answer frequently asked questions, prepare access logs, and summarize follow-up meetings. All of that time currently spent on administrative tasks could be transformed into productive conversations.
Why onboarding fails (and what the data says)
In short: Onboarding fails primarily due to a lack of structure and support, not a lack of tools. According to Gallup, just one 12% Most employees believe their company does a good job of bringing in new people.
The patterns that repeat when onboarding doesn't work:
- The entire first day is paperwork: The person signs documents and doesn't know anyone.
- Information avalanche: Everything is handed over at once, when there is still no context to understand it.
- The manager doesn't have time: The responsibility for welcoming them is delegated to whoever is available.
- Nobody follows up: After two weeks, it is considered finished.
- It is not measured: Without data, the process never improves.
The cost is high: a failed integration results in lower productivity, a worse work environment, and often an early departure that forces the entire selection process to be repeated.
What to automate with AI and what not to
In short: Automate administrative and repetitive tasks (documentation, access, reminders, frequently asked questions, meeting summaries). Don't automate the relationship: welcome, one-on-one interactions, feedback, and evaluation must remain human.
| Yes, automate | Keep human |
| Sending and signing of documentation | The welcome on the first day |
| Registration and access to tools | The presentation to the team |
| Itinerary and reminders | The 1:1 with the manager |
| Answers to frequently asked questions | Feedback and evaluation |
| Summaries and minutes of the sessions | The difficult conversations |
| Surveys and data collection | The decision about the person |
The rule is simple: If the task doesn't require judgment or empathy, automate it.If it's necessary, protect it—and use the time you've gained to make it better. We explore the ethical dimension of all this in Ethics and technology in HR.
30-60-90 day onboarding plan
Onboarding isn't just two days: it's a three-month process. This structure works for most teams and is easy to maintain.
Preboarding (before the first day)
Signed documentation, prepared equipment, access credentials created, and a welcome message from the manager. It's the easiest thing to automate and what prevents the most anxiety.
First 30 days: understanding
Objective: context and relationships. Introductions, basic training, a mentor, and weekly one-on-one meetings. By the end of the month, the person should know what is expected of them.
Days 30-60: contribute
Objective: Initial deliveries with increasing autonomy. Frequent and specific feedback, and adjustment of expectations if necessary.
Days 60-90: Consolidate
Objective: autonomy and personal goals. It concludes with a review discussion: what has worked, what is missing, and where their development is headed.
| Moment | Focus | A sign that things are going well |
| Preboarding | Logistics and welcome | The first day arrives with no pending paperwork. |
| 0-30 days | Context and relationships | She knows who to ask and what to expect. |
| 30-60 days | First installments | Contributes with light supervision |
| 60-90 days | Autonomy | He has his own goals and asks for feedback. |
Remote and hybrid onboarding
In short: In a remote setting, you need to make explicit what happens naturally in the office: introductions, informal context, and opportunities to ask questions. More one-on-one meetings, more written documentation, and a dedicated colleague assigned from day one.
What you learn by osmosis in an office—how to speak here, who to ask, what's considered urgent—doesn't happen automatically in a remote setting. It's helpful to:
- Schedule 1:1 presentations with key people, not just a group meeting.
- To document in writing what in the office would be a hallway conversation.
- Record the training sessions so you can review them at your own pace.
- Increase the frequency of the 1:1 meetings the first few weeks.
More context about these environments can be found in our guide. remote and hybrid work.
AI-powered onboarding tools
| Tool | What is it for? |
| HiBob | Automated onboarding workflows and people management |
| Talmundo | Personalized onboarding experiences by profile |
| Leena AI | Assistant who answers frequently asked questions from new employees |
| Internal chatbots | Support and answering of doubts during the first few weeks |
| Voicit | Record, transcribe, and summarize onboarding sessions and 1:1 meetings, with minutes and agreements ready. |
Not all of them are necessary: the important thing is to cover three things — logistics (documents and access), the doubts (that are resolved without waiting) and the monitoring (that a record of what was agreed upon is kept).
How to document onboarding sessions
The first few weeks are filled with meetings: presentations, training sessions, one-on-one follow-ups. The problem is that Taking notes while accompanying breaks the connectionAnd without notes, the agreement is lost.
Here's a tool like Voicit It fits well: record the session —in person or by video call—, transcribes it and generates a summary with the agreements and next steps, just like a meeting minutesThe new employee can review what was said at their own pace, and the manager can retrieve up to a 80% of the time which I would dedicate to writing notes. It is also a way of save time in meetings recurrent.
How to measure if your onboarding is working
In short: It measures time to becoming productive, retention at 90 days and one year, new employee satisfaction (surveys at 30 and 90 days), and the HR administrative burden. Without data, the process doesn't improve.
| KPI | What does it tell you? |
| Time to productivity | How long does it take to perform autonomously? |
| 90-day retention | Whether the integration takes hold or falls apart quickly |
| 12-month retention | The medium-term effect of the accompaniment |
| Satisfaction (30 and 90 days) | How the person experiences it, not how we think they experience it |
| Management hours per onboarding | How much administrative time is being freed up? |
| Itinerary compliance | Whether the plan is actually implemented or abandoned |
Common mistakes when using AI in onboarding
A chatbot doesn't replace having a person greet you. It automates the process, never the interaction.
Sending everything on the first day is counterproductive: deliver each item when it is needed.
An automatic summary or journey needs a human read before it is sent.
Any recording requires informing the person and having their agreement.
Onboarding lasts about 90 days; dropping it earlier wastes everything before.
Without KPIs you won't know whether the process improves or just seems to.
Actionable summary
- ✅ Automate logistics, questions and follow-up; protect the relationship.
- ✅ Think in 90 days, not two.
- ✅ Remotely, make explicit what is implicit in the office.
- ✅ Document the sessions without ceasing to be present.
- ✅ Measures time to productivity and retention from day one.
- ✅ Always review what the AI produces before it reaches the person.
Frequently Asked Questions
What is AI-powered onboarding?
It is an onboarding process in which artificial intelligence takes care of repetitive tasks —documentation, access, reminders, frequently asked questions, meeting summaries— so that the team can dedicate their time to supporting the person.
How long should an onboarding process last?
The usual and recommended timeframe is around 90 days, structured as follows: preboarding, the first 30 days (understand), 30-60 days (contribute), and 60-90 days (consolidate). Ending it after two weeks is one of the most common mistakes.
What should NOT be automated in onboarding?
The relationship: the first-day welcome, the introduction to the team, the one-on-one meetings with the manager, the feedback, and the evaluation. Automate the process, never the interaction.
How do you measure if onboarding works?
With KPIs such as time to become productive, retention at 90 days and 12 months, new employee satisfaction at 30 and 90 days, and management hours per onboarding.
How to do remote onboarding?
Making explicit what is learned on its own in the office: 1:1 presentations, written documentation, recorded sessions for review, and more frequent 1:1 meetings during the first few weeks.
Can onboarding sessions be recorded?
Yes, by informing the person and obtaining their agreement, explaining what it will be used for (remembering what was agreed upon and being able to review it) and complying with the GDPR.
What tools are useful for onboarding with AI?
HiBob and Talmundo for workflows and itineraries, Leena AI or an internal chatbot to answer questions, and tools like Voicit to record, transcribe and summarize sessions and 1:1s.
Can AI replace the person responsible for onboarding?
No. AI reduces administrative work, but guidance, judgment, and trust are human. When used properly, it frees up time precisely for that.
CEO and co-founder of Voicit. Works every week with HR teams that document their processes with AI.
Related articles
- 1:1 meetings between manager and employee
- Remote and hybrid work: trends for HR
- How to record meetings and generate minutes with AI
- Ethics and Technology in Human Resources