Build the capability inside your own team
Two to three days, on your own systems and your own recurring work — not a slide deck. Your team leaves able to build a first automation themselves, and to tell a good AI proposal from an expensive one.
What's actually going on at your organisation?
Recommended
See the full outline ↓Pick the situation closest to yours — every course below ends on your own systems, not a sample dataset.
Three ways to get there — this is the one that builds the muscle
Some problems are best solved by having us build it (Professional Services) or run it for you (SolmiriX). This one is different: it exists so your own team can build, judge, and maintain AI work without waiting on anyone — including us.
Gartner expects around 60% of agentic AI projects to fail through 2026, and the most common reason is not the model. It is that the data was not ready and the organisation had no one able to say so out loud before the money was spent.
Nobody internally can extend it
A proposal is persuasive, a pilot works, and then it stalls — because whoever bought it cannot maintain it, judge it, or say what should come next without going back to whoever built it.
- Staff use AI tools unofficially, with no policy and no guardrails
- Leadership cannot separate a real proposal from a confident one
- The pilot works; nothing after the pilot does
Capability stays with you
We teach the same methods we use on our own delivery work, on your systems, and we write down what your organisation decided so it survives the people who were in the room.
- Your team builds and keeps working automations
- A written acceptable-use policy your staff helped draft
- A ranked roadmap you can act on without us
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.
The readiness engagement
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.
You may recognise this: half your staff are already pasting work into a chatbot, there is no policy saying whether that is allowed, and a vendor is proposing something ambitious that nobody internally can evaluate. What changes: you get a ranked list of what is worth doing, a written policy your team helped draft, and enough internal understanding to say no to the rest.
- Assess
- Roadmap
- Govern
- Enable
-
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
Six courses, by who is in the room
Each one is built around what that group actually does all day. They can be taken on their own, or as the enablement half of a readiness engagement. Every course finishes on your own systems.
AI Foundations for Professionals
Anyone, any department. No technical background assumed.
- What AI is, and the things it is confidently bad at
- Writing prompts that produce usable output
- Checking an answer before you rely on it
- What must never be pasted into a public tool
You leave with: five tested prompts for your own recurring tasks.
AI for Business Leaders
Executives, board members, department heads.
- What is real in 2026 and what is vendor noise
- Reading an AI proposal for the parts that will not work
- Cost, risk, and where the liability actually sits
- Choosing a first use case: impact against feasibility
You leave with: a ranked shortlist of your own use cases, and the questions to put to any vendor.
AI in the Workplace
Whole teams, taken together rather than as individuals.
- Day one on your department's real recurring tasks
- Drafting, summarising and translating — English and Nepali
- Where a human must stay in the loop, and why
- Day two writing your own team's ground rules
You leave with: a written team AI policy, and tested prompts for each role in the room.
AI Automation Essentials
Operations and administrative teams.
- Spotting which repetitive work is worth automating
- Document handling, data entry and routing
- Building a first automation with no-code tools
- Knowing when to stop and call an engineer
You leave with: one automation running on a real process of your own.
AI Workflow Automation
Operations leads, analysts, and internal IT.
- Mapping a workflow end to end before touching a tool
- Connecting systems: APIs, webhooks and queues
- Failure paths — retries, review queues, alerts
- Approval gates, and measuring whether it worked
You leave with: three automations, built by your team, running on your own systems.
Data & AI Readiness
IT, data owners, and compliance staff.
- What “AI-ready data” means in practice
- Auditing your own sources, quality and gaps
- Access, retention, and who may see what
- ETA 2063 and GDPR-aligned handling
You leave with: a readiness assessment of your own systems and a prioritised roadmap.
On pricing
There is no per-seat list price, for the same reason there is no package price anywhere else on this site: a half-day board briefing and a three-day build for nine engineers are not the same product. Group size, location, how much of it runs on your own systems, and whether it sits inside a consulting engagement all move the number.
Tell us who needs to be in the room and what they do, and you get a written quote with a fixed scope before anything is booked.
Where it runs, and in which language
Designed first for the Nepali market, and structured so the same material works for an international cohort without being watered down for either.
On site, anywhere in Nepal
We come to you — your building, your systems, your network. For organisations outside the Kathmandu valley this is usually the only format that works, and it is the one where the most gets built.
Remote, for international teams
Live sessions rather than recordings, scheduled against your timezone, with the same exercises run on your own systems over a screen share. Cohorts are kept small enough that everyone actually builds something.
English and Nepali
Taught in either, or both in the same room — which is what most Nepali organisations actually need, because the leadership session and the floor-staff session rarely want the same language.
नेपालीमा पढ्नुहोस् →Four commitments, all of them checkable
We are not going to show you a wall of client logos. Here is what to hold us to instead.
Taught by the people who build
The same team that ships DocTrace, UnifyCore and SolmiriX writes and delivers this material. The failure paths in the Build course are there because we have had to handle them in production, not because a curriculum template listed them.
See what we build →Your systems, not a sample dataset
Every course ends on your own data and your own tools. A workshop that only ever runs on a tidy demo file teaches people that their real data is the problem, which is exactly the wrong lesson.
We will tell you not to buy it
If an assessment finds your data is not ready, or that the problem is a process one that AI will only obscure, that is the finding you get — in writing. It is a cheaper conversation before a build than after one.
See how we work →You keep everything
The automations built during a course are yours, running on your own accounts, with no dependency on us to keep working. The written policy and the roadmap are documents you own, not slides you were shown.
AI Training We've Conducted
A Few Things Worth Knowing First
Do we need technical staff for this to be worth it?
Not for four of the six courses. Foundations, Business Leaders, In the Workplace and Automation Essentials assume no technical background at all — they are aimed at finance, admin, HR, operations and management. Workflow Automation and Data & AI Readiness do expect people who are comfortable with systems and data, which usually means internal IT or an analyst.
Can you teach in Nepali?
Yes — in Nepali, in English, or with both in the same room. In practice most Nepali organisations want the leadership session in English and the floor-staff sessions in Nepali, and the material is written so that works without one group getting a thinner version.
How much does it cost?
There is no per-seat list price. A half-day board briefing and a three-day build for nine engineers are not the same product, and group size, location and how much runs on your own systems all move the figure. Tell us who needs to be in the room, and you get a written quote with a fixed scope before anything is booked.
Do we have to do the consulting engagement first?
No. Plenty of organisations book a single course and nothing else, and that is a perfectly sensible purchase. The assessment is worth doing first when you have several departments involved, or when you are about to commit real money to an AI proposal and want to know whether it holds up.
How many people can attend?
The awareness courses work well up to around 25 people. The hands-on ones — Automation Essentials, Workflow Automation, Data & AI Readiness — are capped much lower, because the whole point is that every person leaves having built something themselves, and that stops being true in a large room.
Will our data be exposed during the training?
Work runs in your own accounts and environments. Where a course touches live records we agree the scope in writing beforehand, and one of the first things taught in every course is which categories of data must never go near a public tool. See our data protection page for how we handle information generally.