AI capability
Decide what AI should do for you — before you spend on it.
A readiness engagement on your own data and systems: an honest assessment, a ranked roadmap, a written AI policy, and a plan for the teams who will use it. You leave able to tell a good AI proposal from an expensive one.
The engagement
- AssessData, tools, skills, governance
- RoadmapUse cases ranked; what to postpone
- GovernA written AI policy, drafted with you
- EnableTraining aimed at named teams
01 Start from your goal
What are you trying to accomplish?
Pick the closest. You will see the path we would suggest — including when the right answer is a course, or skipping straight to a build.
Understand AI — Get a clear, sober picture first.
- Leadership briefingA half-day on what AI can and cannot do for an organisation like yours — delivered through GeXOn Training Center.
- A short conversationWe look at where your data and processes are today.
- DecideWhether, where and how much — with no obligation to do anything.
Where it leaves you
You leave able to tell a sound AI proposal from an expensive one.
Train leadership — Give decision-makers a shared vocabulary.
- Leadership sessionOpportunities, risks, costs and governance — without the hype.
- Your shortlistUse cases specific to your organisation, not generic examples.
- Optional assessmentTest the shortlist against your real data and systems.
Where it leaves you
A leadership team that can set direction and challenge vendors.
Train teams — Hands-on skills for the people doing the work.
- Needs by teamWhich roles need which capability — not one all-staff session.
- Hands-on coursesOn your site or remote, in Nepali or English, on your own kind of work.
- Safe-use rulesWhat must never go near a public tool, taught in every course.
Where it leaves you
Staff who use AI well — and know what not to do.
Automate workflows — You already know the job you want gone.
- Pick the workflowDocuments, support, leads, integration or internal knowledge.
- Scope itA written scope with a fixed price, after a free call.
- Build itRunning in your systems in 3–10 weeks.
Where it leaves you
Skip the strategy phase — go straight to a working automation.
Build AI internally — Grow the capability to build and maintain it yourself.
- AssessData, tools, skills and governance — honestly, including where you are not ready.
- RoadmapBuild-versus-buy for each use case, and the skills each needs.
- EnableTechnical training for your engineers and analysts, and co-building if you want it.
Where it leaves you
A team that can build, judge and maintain AI work without waiting on anyone — including us.
Assess AI readiness — Find out what AI should do for you, before you spend.
- Assess · 2 weeksYour data, tools, skills and governance.
- RoadmapUse cases ranked by impact and feasibility, with what to postpone.
- Govern & enableA written AI policy and a training plan aimed at named teams.
Where it leaves you
A costed roadmap and a policy — in six to eight weeks.
02 Why it exists
Most AI projects stall after the pilot. Rarely because of the model.
They stall because the data was not ready, nobody owned the result, and nobody inside could say so before the money was spent. This engagement is built to prevent exactly that.
- What happens today
- Staff paste work into public chatbots with no policy, and a vendor proposes something nobody internally can evaluate.
- Why it matters
- Data leaks quietly, money goes to the most confident proposal, and pilots that work have no one to extend them.
- What changes
- A ranked list of what is worth doing, a written policy your team helped draft, and people who can judge the rest.
- What improves
- Spend goes where it pays back, risk is written down, and the capability stays with you — not with us.
03 The readiness engagement
Four stages, six to eight weeks.
Before anyone builds anything, we work out what your organisation can realistically do with AI today, what it should do first, and what it must write down before staff go near a public tool.
Engagement 6–8 wks
- ✓Assess
- ✓Roadmap
- ✓Govern
- ✓Enable
Step 1 · Assess
- Data sourcesMapped ✓
- Records duplicated3 systems ⚠
- Written AI policyNone ⚠
- Staff already using AIMost, unofficially
Data, tooling, skills, governance. If the honest answer is that you are not ready, that is the finding — and it is cheaper here than after a build.
Step 2 · Roadmap
- Start hereInvoice handling
- SecondEnquiry routing
- LaterInternal Q&A assistant
- Not yetDemand forecasting
Ranked on impact against feasibility. The list of what to drop matters as much as the list of what to build.
Step 3 · Govern
- Acceptable useWritten ✓
- What must never be pastedListed ✓
- Human approval gatesDefined ✓
- ETA 2063 · GDPRAligned ✓
Your staff help draft it, which is the only version anyone follows. Written down, so it outlasts whoever was in the room.
Step 4 · Enable
- BoardAI for Business Leaders · half day ✓
- FinanceAI Automation Essentials · 2 days ✓
- OpsWorkflow Automation (Build) · 3 days ✓
- ITData & AI Readiness · 2 days ✓
Training aimed at the teams the roadmap actually named — not an all-staff session that teaches everyone the same thing badly.
-
Assess
Two weeks looking at four things: the data you hold and what state it is in, the tools you already run, the skills actually present in the building, and what governance exists. We interview the people doing the work, not only the people who commissioned the review.
-
Roadmap
Candidate use cases scored on impact against feasibility, with the reasoning shown. Some get marked "not yet" and the reason is named — usually not enough history, or data spread across systems that disagree with each other. Knowing what to postpone is most of the value.
-
Govern
A written acceptable-use policy: what staff may use, what must never leave your systems, where a human has to approve before anything executes, and how long anything is retained. Drafted with your team rather than handed to them, and aligned with Nepal's ETA 2063 and GDPR practice.
-
Enable
Now the training is aimed at something specific. The roadmap named which teams need which capability, so each group gets the course that matches its actual job — rather than one all-staff session that teaches everybody the same thing badly.
- A readiness assessment across data, tools, skills and governance
- Use cases ranked by impact and feasibility, with what to postpone
- A written AI acceptable-use policy your staff helped draft
- A training plan aimed at named teams, not the whole org
04 Delivery
Where it runs, and in which language.
Designed first for organisations in Nepal, and structured so the same material works for international teams without being watered down for either.
On site, anywhere in Nepal
Your building, your systems, your network. For organisations outside the Kathmandu valley this is usually the format that works — and the one where the most gets done.
Remote, for international teams
Live sessions scheduled against your timezone, with exercises run on your own systems over a screen share.
English, Nepali, or both
Leadership sessions often in English and floor-staff sessions in Nepali — with the same substance in both.
Training through GeXOn Training Center
Courses for leaders, whole teams and operations staff are run by our training arm. GeXOn Training Center ↗
05 Questions
Before an assessment.
What do we get at the end of a readiness engagement?
A written assessment of your data, tools, skills and governance; a roadmap of use cases ranked by impact and feasibility, including what to postpone and why; a written AI acceptable-use policy drafted with your team; and a training plan aimed at named teams.
How long does it take?
Usually six to eight weeks. The assessment itself takes about two; the rest depends on how many departments are involved.
Do we need technical staff for this to be worth it?
No. The assessment is aimed at leadership and operations as much as IT. The roadmap will tell you which technical roles matter for the use cases you choose.
Can you work in Nepali?
Yes — in Nepali, in English, or both in the same room. Most organisations want leadership sessions in English and floor-staff sessions in Nepali, and the material is written so neither group gets a thinner version.
Where does the training happen?
Courses are delivered through GeXOn Training Center, our training arm, on your site anywhere in Nepal or remotely for international teams. The roadmap decides which courses each team needs.
Will our data be exposed during the work?
Assessment work runs in your own accounts and environments. Where it touches live records we agree the scope in writing first. Our data protection page explains how we handle information generally.
Tell us who needs to be in the room.
A 30–45 minute call is enough to tell whether you need an assessment, a course, or a build — and roughly what each would take.