Hesham Khalafallah
In recent years I have attended and taken part in a great many conferences, forums and workshops on artificial intelligence, at universities, government bodies and companies. Let me be clear from the outset: I am not against these events. Public, scientific and technical dialogue about AI matters; it helps build a deeper understanding of the technology and its implications, and it always produces meaningful outcomes and recommendations.
What concerns me is what happens next. Are these recommendations implemented? Do the outcomes turn into actual work? The question I always ask once the sessions end is: what will change tomorrow morning in the way this organization works?
All too often, the irony is painful. An organization spends considerable sums on hosting a conference, on venues and logistics, and then says there is no budget to implement the recommendations, despite their importance and clear return. The organization is left without a single application that actually works, even partially. And when it finally decides to act, it brings in a ready-made external system without building any internal capability or understanding.
So this article is not meant to diminish anyone’s efforts. It is meant to focus on the practical side: how do recommendations become reality?
Why do recommendations stay on paper?
From what I have seen, the same reasons recur almost everywhere:
• A budget for discussion, but none for execution: resources are allocated to the event itself, with no clear line item for implementing its outcomes.
• No executive owner: everyone talks about AI, but no one is responsible for applying it and measuring the results.
• Starting with the big dream: a five-year strategy instead of a small, real problem that can be solved within weeks.
• Missing data: AI needs organized data, and many organizations still keep theirs on paper or in systems that do not talk to each other.
• Total dependence on vendors: the organization buys the tool, the vendor leaves, and the knowledge stays outside.
Five steps to turn recommendations into practice
1. Start with a real problem in daily work
Don’t ask, “How can we use AI?” Ask instead, “Which task takes up most of our team’s time?” It might be answering repetitive inquiries, reviewing documents, preparing reports or sorting requests. That is where the real opportunity begins.
2. Deliver a small application in weeks, not years
Choose one use case, apply it in one department, within a short period. A small application that works is better than a big strategy waiting for approval.
3. Build an internal team that owns the knowledge
Experts and vendors are helpful, but there must be a small team inside the organization that understands what is happening, learns, develops and follows up. An organization that does not own the knowledge will always depend on those who do.
4. Measure the impact in numbers
How many working hours did we save? How much did service time fall? Did quality or user satisfaction improve? Numbers are what convince management to keep going, and what separate a successful pilot from a media showcase.
5. Scale once the results are proven
When the first application succeeds, extend it to other departments and add new use cases. This is how AI gradually moves from a pilot project to a natural part of the way the organization works.
The missing link: from recommendation to execution
• When an organization stops at recommendations: outcomes are written up and filed, a comprehensive strategy is awaited, and success is measured by the success of the event itself.
• When an organization carries on to execution: it turns every important recommendation into a project with an owner, a budget and a deadline; it experiments, measures results and scales gradually, and it measures success by the number of processes that have actually changed.
Here I propose a simple idea: set aside a share of the budget of every conference or forum to implement at least its first recommendation in the months that follow, so that the outcomes do not remain mere ink on paper.
From my experience: an idea that became a platform
When we started working on Judicia AI, we did not begin with a conference or a vision document, but with a specific problem: how can we help specialists carry out legal research in Saudi regulations accurately and quickly? Instead of staying at the discussion stage, we built the platform and launched it. It is up and running today, and we are working on its development and commercial launch. What I learned from this experience is that execution itself is the best teacher: real problems do not appear in conference halls, but in actual use.
Conclusion
Conferences and workshops open doors and build understanding, but an organization’s maturity in AI is ultimately measured by the number of processes that have actually changed in its daily work. Start small, start now, measure the results, and build the knowledge inside your organization.
If your organization wants to move from talking about AI to actually applying it, I would be glad to identify the right first use case with you and set a realistic plan to implement it.
