Skip to content
MekongSI

Perspective

AI is changing what organisations do, how they are structured and what they buy.

Here is how we see the next few years, and how we have built MekongSI to stay useful as AI keeps moving.

Six shifts

Where AI is heading, and our response.

  1. Shift 01

    AI agents take on routine knowledge work

    Reporting, document handling and first-line analysis are increasingly done by software, supervised by people.

    How we respond

    We redesign workflows for human + AI teams and build agents with human review, audit trails and clear accountability.

  2. Shift 02

    Adoption, not technology, is the bottleneck

    Most organisations have tried AI; few have it working reliably inside real processes.

    How we respond

    Senior, forward-deployed teams who sit with your people, integrate with your systems and stay until it works in production.

  3. Shift 03

    Trust becomes the deciding factor

    Boards, donors and regulators now ask for evidence that AI is safe, fair and under control.

    How we respond

    Governance, independent evaluation and managed AI operations as standard services, not afterthoughts.

  4. Shift 04

    Data and AI sovereignty matter more

    Sensitive data must stay in-country or on-premises, and AI must work in local languages.

    How we respond

    Model-agnostic architectures that run in your cloud, on-premises or in-region, with Khmer and English in scope.

  5. Shift 05

    Clients buy outcomes, not hours

    Buyers increasingly expect to pay for defined deliverables and results rather than time spent.

    How we respond

    Fixed-scope packages, managed capacity and outcome-based partnerships where baselines are clear.

  6. Shift 06

    Skills are the scarcest resource

    Every role is changing; organisations need people who can direct, check and improve AI.

    How we respond

    DataU training and capacity transfer built into engagements, so capability stays with you.

The organisation of the future

Smaller teams, more leverage, clearer accountability.

As AI agents take on routine analysis, reporting and document work, organisations will run with fewer hand-offs and smaller teams, each directing a set of AI tools. The scarce skills become judgement, domain knowledge and the ability to check and improve what AI produces.

That makes three things decisive: data that AI can trust, governance that lets people move quickly without adding risk, and teams trained to work this way. Those are exactly the capabilities we build with clients, through transformation, data and AI engineering, governance and DataU training.

For development organisations and governments, it also means doing more with constrained budgets while protecting the people behind the data, which is why responsible AI and data sovereignty sit at the centre of how we work.

Built to last

Built to keep pace with AI, so your systems do too.

Model-agnostic by design

Architectures that can swap in better or cheaper models as they arrive, without rebuilding the system.

Evaluated continuously

Every AI system ships with evaluations that are re-run when models, prompts or data change.

AI-assisted delivery

AI supports our research and engineering under the same data-protection rules as everything we do: faster turnaround, with confidentiality intact.

People who direct AI

Consultants and engineers who frame the right problems, check what AI produces and take responsibility for results.

Ways to engage

Buy results, not hours.

We price around deliverables, capacity and outcomes, so you pay for results rather than time.

Fixed-scope packages

Best for

A clear first step, a defined deliverable or a donor-funded activity.

How it works

Agreed scope, timeline, deliverables and price before work starts.

Programmes and managed capacity

Best for

Multi-phase transformation, platforms, and ongoing data and AI work.

How it works

Phased delivery, or a dedicated pod billed monthly against agreed capacity and service levels.

Outcome-based partnerships

Best for

Repeatable processes with a measurable baseline, such as report production or document processing.

How it works

Fees linked to agreed results, after a baseline is measured together.

See how this translates into our services.

Plan your organisation's next three years with AI.

Tell us what you are trying to achieve. We will tell you honestly whether we are the right partner, and what a sensible first step looks like.

Book a scoping callExplore services