AI adoption sprint for university teams
A four-week hands-on programme that teaches your language centre and staff to use AI with confidence rather than fear. Three working workflows, shipped during the sprint.
An AI adoption sprint is a four-week hands-on training programme in which a university team learns to use AI tools on its own real work, and finishes with working workflows rather than notes. HumAIn Connection runs these for up to six staff members, delivered in English, Dutch or French.
The problem with AI training
Most institutional AI training is a lecture. Someone presents the technology, the room nods, everyone returns to their desk, and nothing changes because nobody built anything. Three months later the institution concludes that AI did not deliver, when what actually happened is that no one ever used it on their own work.
The second failure mode is the opposite: staff are already using AI tools, quietly and without guidance, pasting student data into whatever service is free and convenient. The institution has no visibility into it and no policy that anyone follows. That is a bigger exposure than a slow adoption curve.
Language centres feel both of these more sharply than the rest of the university, because AI touches translation, assessment and teaching material directly. Staff there are often the most anxious about it, and with reason: it is their work that the technology is loudest about replacing.
What we deliver
Four half-day sessions for up to six people, built around your work rather than a generic curriculum.
- Three working workflows, shipped
- Not slides. By the end of the sprint the team has three automations or AI workflows running on real tasks it does every week, built during the sessions.
- Hands-on from the first session
- Participants work on their own material. The first thing they build is something they were going to have to do anyway.
- Honest limits
- We spend real time on what these tools get wrong, how to spot it, and which tasks should not be handed to them. Staff who understand the failure modes use the tools better and trust them appropriately.
- A playbook you keep
- Written documentation of what was built and how to run it, so the capability survives one person moving on.
- Delivered in English, Dutch or French
- In the language the team is most comfortable thinking in, which matters more for training than for anything else we do.
Who it is for
Language centre staff, international office teams, marketing and communications, and administrative teams carrying repetitive document and correspondence work. Up to six participants, because beyond that it stops being hands-on and becomes the lecture we are trying to avoid.
It works best when the group has real authority over its own processes. A team that can change how it works will finish the sprint with three things running. A team that has to file a request to change anything will finish with three good demos and no adoption, and that is a procurement problem we cannot solve for you.
What the four sessions cover
Session 1
Foundations and the honest limits. What these systems do, what they reliably fail at, and where student and staff data must not go. Participants identify the repetitive work in their own week that is worth attacking.
Session 2
Build the first workflow together, on real material the team brought. Most of the learning happens here, at the point where a tool does not do what the participant expected.
Session 3
Two more workflows, with the team leading and us reviewing. Quality checking and review habits are built in, since the output still needs a human who knows the subject.
Session 4
Consolidation and handover. The playbook is written, the workflows are documented, and the team agrees what it will and will not use these tools for going forward.
Who runs it
The sprint is delivered by the same people who build our production pipelines, which is deliberate. Training that is written by someone who does not ship the technology tends to teach the demo rather than the reality, and it shows the first time a participant hits something that does not work.
It is also informed by our linguist side, which matters when the room is full of language professionals who want a straight answer about what AI does to their profession. We give them one, including the parts that are uncomfortable.
Timeline
| Scoping call and participant selection | 1 week before |
|---|---|
| Four half-day sessions | Across 4 weeks |
| Playbook and workflow handover | End of week 4 |
| Optional follow-up review | 4 to 6 weeks later |
What drives the cost
We do not publish a rate card yet, because a number without a scope behind it is noise. What we can tell you is exactly what moves it, so you can size the work before you talk to us.
- Remote or on-site delivery.
- Number of participant groups, since each group is capped at six.
- Language of delivery.
- How much of the workflow build we do versus facilitate.
- Whether you want a follow-up review session after the team has been running the workflows alone.
Every engagement starts with a scoped pilot, priced and agreed in writing before any work begins. No open-ended retainers, and no per-word invoicing that nobody can forecast.
Questions we get asked
Why only six participants?
Because beyond six it stops being hands-on. Everyone in the room builds something during the sessions, and that is not possible in a lecture hall. If you need to train more people, we run more groups rather than a bigger one.
Do participants need a technical background?
No. The sprint is built for language and administrative professionals, not developers. Nobody writes code.
What do you mean by three working workflows?
Three tasks the team does regularly that leave the sprint partly or fully automated, running on the team's own tools and material. They are chosen by the participants in session one, so they are things that matter to the people doing the work.
Is this just prompt training?
No. Prompting is one session's worth of material at most. The larger part is workflow design, knowing where a human has to stay in the loop, and building review habits so the output can be trusted.
Can it be delivered in Dutch?
Yes, in English, Dutch or French. For training specifically, the language the team thinks in makes a noticeable difference to how much people are willing to ask.
What happens after the four weeks?
The team keeps the workflows and the playbook. Some institutions book a follow-up review a month or two later, once real use has surfaced the questions that only come up in practice.
Related services
Start with one pilot, and judge the review.
Tell us what you are trying to do and we will tell you honestly whether we are the right people for it.
Prefer email? hugo.megardon@megaaisolutions.com