Maqsut Narikbayev University — a needs-first integration

Technology second. Understanding first.

At MNU the work did not begin with a camera feed. It began with a series of meetings.

The team sat down with two departments separately, because they do not have the same job and would not have given the same answers in one room.

The security department. Discovery meetings to map the existing surveillance workflows, understand how incidents are currently escalated and resolved, and identify where real-time detection would add something the department did not already have.

The psychological support unit. Separate sessions on student welfare priorities, to make sure the deployment complemented the pastoral structures already in place rather than cutting across them. A safeguarding system that arrives as a surveillance programme makes the welfare team's job harder, not easier.

What was deployed

Three modules, each chosen because those conversations had identified a specific need for it:

• Weapon detection — real-time alerts when firearms, knives or hazardous objects are detected across campus camera feeds.
• Smoking detection — automated detection of smoking in restricted zones: lecture halls, corridors and public spaces.
• Aggressive behaviour detection — pattern recognition identifying physical altercations, intimidation and repeated conflict behaviour on campus.

AMAN Secure was integrated into the cameras the university already owned. No hardware was replaced.

How it unfolded

1. Stakeholder discovery meetings. A series of sessions with the security department and the psychological support unit — mapping existing workflows, identifying pain points, and defining what a meaningful deployment would look like here specifically.
2. Needs-based module selection. Three detection modules chosen directly from the priorities those meetings identified.
3. AMAN Secure integration. Deployed onto the university's existing camera infrastructure, with no hardware replacement required.
4. Testing and validation. Internal tests of algorithm performance in real campus conditions.
5. Leadership presentation. A formal presentation of results, findings and recommendations to university leadership.

Tested in live campus conditions

The pilot was evaluated on three things, in the environment it would actually have to work in:

• Detection accuracy across diverse campus environments and lighting conditions.
• False positive rates, so security staff receive alerts worth acting on rather than noise. A system that cries wolf is switched off within a term, and then it protects nobody.
• System responsiveness, measured across different scenarios to confirm real-time capability.

What changed

The needs-assessment approach meant the deployment addressed real institutional priorities rather than a generic checklist — and that showed up as buy-in.

• The modules deployed answered problems the university had actually named
• Stronger buy-in from both the security staff and the psychology team
• Accuracy, false positive rates and responsiveness validated in live conditions
• Formal results and recommendations delivered to university leadership

MNU also produced something reusable: the needs-assessment model established here became the repeatable framework for how AMAN starts a deployment.