Independent work · education and professional teams
Make AI competent in your context.
Most organisations already have experts. Their knowledge lives in heads, workflows, documents, examples and judgement. The problem is not access to AI. It is making AI usable in a specific domain, under review.
What this page is: how to work with me. The method is expertise encoding. The laboratory is NiallOS. Delivered programmes sit on Work.
What I do
Encode expertise. Keep people in the loop.
The starting point is always the same: what your experts actually know, how they decide, which sources matter, and what generic AI gets wrong in your domain. From there the work is encoding that into systems that retain it.
That might be tools a team uses daily. It might be knowledge infrastructure that makes organisational expertise queryable. It might be a training programme so people can encode their own expertise instead of only using off-the-shelf tools.
The design bet
Professional judgement can be made available at the point of work without being replaced. Source grounding and review are part of the system, not an afterthought.
Typical work
What an engagement can look like
These are kinds of work, not a menu of products with prices attached.
AI strategy and readiness
Assess whether leadership, policy, staff capability and operations can support AI use, then set a practical next step. For schools, see AI readiness for schools.
Knowledge capture and workflow mapping
Find the decisions, documents, examples and quality bars that actually matter, and map the repeated work around them.
Product design and prototyping
Design and test AI-supported tools against a real workflow, with enough context to be useful and enough review to be safe.
Document-heavy knowledge systems
Turn policies, research, proposals and working files into a searchable layer that people and AI can both use.
Source-grounded retrieval
Build synthesis workflows that find, compare and cite the right material, instead of relying on the last prompt.
Workshops and quality gates
Leadership and staff sessions, plus human-in-the-loop checks for client-facing or high-stakes output.
Who I work with
Education first, and teams whose product is expertise
Education organisations, professional services, L&D, research teams and development partners who want to move past generic AI adoption into something that reflects how they actually think and work.
Most of my project experience is across Sub-Saharan Africa and the Middle East, in settings where generic tools fall short: different curricula, languages, infrastructure and professional norms.
What this is not
Not a chatbot pitch. Not a SaaS signup.
- Not unsupervised AI delivery. People still specify, supervise and reject the work.
- Not a substitute for the programme pages. DELP, KFAS, ECOSOCC and the rest stay on Work, with public sources.
- Not a Cambridge product page. This is independent work. I do not present a future title as current.
Contact: hello@niallmcnulty.com · +27 76 739 1687. No form on this page. If email is easier, use that.
If the valuable knowledge is still in people’s heads
Start with a conversation. We can then decide whether the fit is Learn, Done with me, or Done for you.