August 31, 2026
Weekly Data & Architecture sync
Person #42, Person #39, Person #41, Person #43
Person #42
Person #39
Person #41
Person #43
Person #41: Yaa, hello? Can you hear me? I think my mic is acting up again. Testing, testing. Person #39: Yes, we can hear you. Good morning everyone. Let's get started, we have a lot to cover and I know some of us have back-to-back. Person #42: Wait, is Zayed on? I don't see his name here. He was supposed to— Person #41: He sent a message in the group, he's stuck with the Chairman's office. Something about the GITEX briefing. Person #43: Okay. Fine. Let's wait for him. Person #39: We can start with the data governance policy. It's been sitting with legal for—um, three weeks now. I need a status before I escalate. Person #41: Wallah ya akhi, I told you, legal is backed up with the procurement law revision. Everything's been in a queue since Ramadan. Person #42: Actually, I had a quick chat with Reem from legal last Thursday. She said the comments from ADEC came in late, that's why it's stuck. They wanted the section on cross-entity data sharing to align with the federal data law. She's expecting to send it back to us by next Sunday. Person #39: Next Sunday okay. That's fine. Inshallah. But I want to flag that we have the World Governments Summit in February and the Chairman wants to announce the open data platform pilot there. If this policy doesn't get signed off before November, the timeline becomes tight. Person #41: What does the open data platform have to do with the governance policy, yaani? They're separate things. The platform is technical. The policy is—well, it's policy. Person #39: No, no, they're interlinked. If we don't have the policy ratified, the entities won't share data onto the platform. You know how it is, everyone's nervous about privacy. The policy gives them cover. That's the whole point. Person #41: We tried this in 2016, you know. The open data portal. We had, I think, 40 datasets. We had weather data, traffic data, some education numbers. Nobody used it. We ended up quietly shutting it down in 2018. Person #42: The difference now is the AI. With the LLM programs, there's actual demand for structured data. The 200 AI use cases in the strategy, they all need data. So it's not just for transparency anymore, it's fuel for the models. Person #41: Yaa, okay, but who's going to maintain it? The entities can't even keep their own websites updated. You think they're going to maintain open datasets? Person #39: That's why we have the data quality framework. Once the governance policy is in place, we enforce the standards. It's—okay, let me zoom out. We have three things here. The policy, the platform, and the quality standards. They're a package. We're not launching anything until all three are ready. Person #43: So we're waiting on legal, then. What about the executive council? Has anyone briefed them? Person #39: Not yet. I have a meeting with the Undersecretary on Thursday. I'll brief him and see if he wants to take it up to ADEC. Person #42: If you want, I can prepare the slide deck for that. I have the template from the last architecture review board. Person #39: Yes, that would help. Kindly—thank you. Person #41: You know what I think about the architecture review board, tab. Person #42: You've made it very clear, boss. Multiple times. But I have to say, the ARB is useful for tracking decisions. We have like 30 open decisions from the last year, and only half have been implemented. If we didn't have the records, we wouldn't even know. Person #41: That's my point. Why record decisions if nobody's going to follow them? It's bureaucratic theater. Person #39: It's not theater. It's accountability. Some of those decisions are waiting on budgets that haven't been approved. We're waiting on the autumn budget cycle. That's not the ARB's fault. Person #41: Fine. Fine. Moving on. Can we talk about the cloud migration? Because I have a licensing renewal coming up and I need to know if we're extending our current data center footprint or— Person #39: Wait, before we move on. Does anyone have the final count on entities that have signed on to ADERP? I know we were at 39 last month, but— Person #42: I think we hit 42 last week. DMT came on board. And the Department of Culture and Tourism is scheduled for next month. I can check the dashboard. Person #39: No, I mean the signed commitment. Not the actual rollout. Because some entities are dragging their feet on signing. They say they want to see the first wave results before committing. Person #41: That's what we said. Or I said. I mean, we're in wave one now. If the first wave is successful, the others will come. If not, then we have bigger problems than signing documents. Person #43: From my side, the data quality coming out of ADERP is still an issue. We're seeing, like, 30% of records with missing or inconsistent fields. The finance module is fine because the Department of Finance is strict, but the HR data is messy. People are still entering data in the old system and then manually transferring. Person #39: That's exactly what the governance policy addresses. Data ownership. Who is accountable for each domain. Person #43: I know, habibi, but I need data now. We have the analytics backlog. The dashboards for the 70 entities are supposed to be refreshed by the end of Q3. If the data is bad, the dashboards are bad. And then we get the calls from the entities complaining. Person #41: What's the backlog count now? For analytics, I mean. Person #43: We're at, like, 200 requests. Some of them are small, they just need a report. Some of them are big, they want a whole new dashboard. We have a team of 15 people. We're stretched thin. And we lost two people last month to the private sector. Person #41: Yaa, I saw that. G42 poached them? I heard they're paying 40% more. Person #43: Walla, I don't know exactly. But yes, we lost two data engineers. And I have requests for more GPU access from the AI use case teams. They want to train models, but our current infrastructure is at capacity. Person #39: We have the new GPU cluster coming in, right? When is that— Person #41: December. It's in the budget. But there's a catch. The vendor is behind schedule. We ordered it in June, they said 16 weeks. Now they're saying 20 weeks. So we're looking at mid-December at best. January if there's a shipping delay. Person #39: January is cutting it close. The strategy says we need the AI use cases running by Q1 next year. If there's no compute— Person #41: I know. I know. I'm on their case. I have a call with the vendor tomorrow. I'm going to give them a deadline. If they can't deliver, we look at alternatives. Person #42: What are the alternatives? Another vendor? Person #41: We could go to the private cloud providers. But that's a bigger cost. And the sovereignty requirements... you know, the data has to stay in country. It's tricky. We could do a hybrid model. But the legal framework is not fully clear yet. Person #39: That's another thing the governance policy should address. Cloud sovereignty. Data residency. It's all in the policy. Person #41: Wait, is the policy about data governance or about cloud? Those are two different things. Person #39: They're related. The policy has sections on data classification, data sharing, data residency. So yes, cloud falls under it. Person #41: Okay, but the cloud procurement is a separate process. I can't wait for the policy to decide on the vendor. Person #39: No, no, I'm not saying wait. I'm saying the policy will provide the framework for future decisions. Right now we're making ad hoc choices, which is fine, but we need to formalize. Person #42: I have a question about the Capability Map. We're doing the entity architecture assessments, and I want to make sure the map is up to date. Because some entities have merged or changed their mandates. Like the Department of Community Development absorbed the social services authority. The map still shows them as separate. Person #39: Good point. The capability map is maintained by my team. We update it every quarter. But we follow the entity governance structure. Until the executive council formally approves the merger, we keep them separate. Person #41: That's bureaucracy. They've been merged for a year now. Everyone knows they're merged. The employees even sit in the same building. Person #39: Rules are rules, habibi. We can't make assumptions. If we update the map before the formal approval, we might have to redo everything. Better to wait. Person #42: Okay, but the assessments for those entities are delayed. We're waiting on the organizational structure clarity. Do you want me to put them on hold? Person #39: Yes, put them on hold. We'll revisit once the approval comes through. When is the ADEC meeting? Person #41: I think it's in two weeks. The Chairman mentioned it last week. They're going to discuss the merger approvals and the 2026 budget. Person #39: Two weeks. Okay. So we have time. Let's—sorry, I have a message. One second... The Chairman's office just asked for slides on the open data platform pilot. By end of day today. Person #41: What? End of day today? It's Monday. And it's, what, 9:20 now? We have seven hours to prepare slides. Person #42: I can draft it. I have content on the technical architecture from the GITEX briefing. And I can adapt it. Person #39: But we need to clarify the scope. Does the Chairman want the technical overview or the strategic value? Because those are different decks. Person #41: I'll send a message to his office. Actually, you should send the message. You have the relationship with them. I'll text you the contact. Person #39: Fine. I'll ask. But I think we can prepare a general overview and then customize. Let's not wait for the clarification. We could lose half a day. Person #42: I'll start with the general deck. Can I get the data from the pilot evaluation? The numbers we have from the six entities? Person #39: Um, the evaluation is still ongoing. We have preliminary results. We have, let me think, we had an 80% reduction in request times for the traffic data. And we had 45 external developers registering for the API access. But those are early numbers. The full report comes next month. Person #41: 45 developers? That's nothing. We said we'd have a thousand by the end of the year. Person #39: That's the projection, yes. But if our marketing is good—the WGS announcement, the GITEX momentum—we could accelerate. Person #42: Also, we have the partnership with the startups. The four startups from the innovation lab are already using the platform. They gave us some positive feedback. Person #41: Positive feedback from startups that want to be our vendors. Not exactly unbiased. Person #39: Let's stay positive. We have a story to tell. The platform is working. The data is flowing. We just need to package it well. And we need to give the Chairman confidence that this is worth announcing at the World Governments Summit. Person #43: I should mention, from the analytics side, we have some interesting use cases with the traffic data. We're building a predictive model for congestion. Early results show we can predict rush hour traffic with 85% accuracy. That's a good demo for the summit. Person #39: Excellent. That's a strong story. Can we include that in the slides? Show a live demo or a screenshot? Person #43: We can do a screenshot. A live demo might be risky if the internet fails. You know how it is. Person #41: Yaa, no live demos. I've seen too many things fail in front of ministers. Static slides are safer. Maybe a short video, pre-recorded. Person #39: Agreed. Video is good. Do we have the video from the GITEX demo? Person #43: We have a rough cut. The editing is not finished, but it's usable. I'll check with the team. Person #42: I'll take the draft and the screenshots and the video link. And I'll put together a deck that covers the vision, the progress, and the roadmap. By end of day. Person #39: Perfect. Kindly share it with me as soon as it's ready. I'll review it and send to the Chairman's office. Person #41: What about the budget impact? Should we include the cost numbers? The Chairman might ask. Person #39: Good point. Do we have the total budget for the platform? Person #41: The approved budget is 42 million for two years. But we've spent about 18 million so far. The remaining is for scaling and maintenance. I'll include that. Person #39: Okay. Now, what about the data governance policy? Should we mention that in the slides? Or keep it separate? Person #42: I'd keep it separate. The slides are about the platform. The policy is a governance issue. If we mix them, we dilute the message. Person #41: But the policy is the key to scaling. Without it, we can't get data from the big entities, like DOH or DMT. The platform is only as good as the data it has. Person #39: Let me think... Okay. We'll include one slide on the policy, just a brief mention. "Data governance framework in progress" with a timeline. That way, the Chairman knows it's on our radar, but we don't overcomplicate. Person #42: I'll add that. "Policy under final legal review, expected sign-off by Q4. Person #41: Is that true? Expected sign-off by Q4? Person #39: We're hoping. If legal doesn't delay more. Inshallah. Person #43: So what do we do about the entities that are not engaging with the platform? We have, like, 10 entities that haven't even attended the onboarding sessions. Person #39: They'll come. Once they see the traffic data and the analytics value, they'll come. It's a matter of time. Person #41: Or we need to force them. The Chairman could send a directive. That would make them move. Person #39: Let's not escalate yet. We give them until the end of October. Then we escalate if needed. Person #42: I have a list of the entities and their engagement status. I can share it with the Chairman's office if they ask. Person #39: Good. Include that in the appendix of the slides. Just in case. Person #41: So, this is a lot for end of day. Can we get an extra pair of hands? Maybe someone from the comms team? Person #39: I'll ask the Undersecretary. He can reassign someone. Or we can use the internal design team. They did a good job with the TAMM 4.0 launch materials. Person #42: Yes, the design team is good. I'll loop them in. Person #39: Okay. Anything else on the platform before we move on? Person #41: No, I think we covered it. Just the file. And the... the video. Yalla, we'll deliver. Person #39: Great. Now, the second item on the agenda. The entity architecture assessments. Person #42, you had an update? Person #42: Yes. We've completed 12 assessments so far. Out of the 53. The first batch is done. The second batch of 10 is in progress. We're on track to finish all assessments by the end of the year. But— Person #41: But what? There's always a but. Person #42: But we're facing some resistance from the entities. They see the assessment as a checklist rather than a value-add. We're getting feedback like "we're too busy" or "we don't have the data." We need to change the narrative. Person #39: How do we change the narrative? Person #42: I think we need to show them the benefits. Maybe a "before and after" example. An entity that used the assessment to improve their operations and saw results. Person #43: We have the example from the Department of Energy. They used the assessment to identify inefficiencies in their permit process. They reduced processing time by 30%. We have a case study. Person #39: Perfect. Let's use that. Add it to the assessment package. And maybe we hold a workshop with the next batch of entities to walk them through the process. Person #42: Good idea. I'll propose a date. Maybe mid-September? Person #39: Let's aim for the 10th. That gives us time to prepare. Person #41: The 10th is a Thursday. Friday is the weekend. Half the entities might be in a rush to leave early. Person #39: Okay, the 14th then. Monday. That's better. I'll send the invitation. Person #42: I can prepare a short video explaining the assessment. Sort of like a "what to expect" video. That might ease their concerns. Person #39: Good. Do it. Person #41: We have too many videos. Everything is a video. Person #42: It's the modern way, boss. People don't read anymore. Especially— Person #41: I read. I read. Person #39: Okay, let's stay on track. What about the reference architecture? Is that still on schedule? Person #42: Yes, the reference architecture for the integration layer is 80% complete. We're working on the APIs and the data exchange standards. We should have the final version by November. Person #39: November. Good. And the integration platform? We had issues with the vendor. Person #41: The integration platform is fine. The vendor is fine. We had a small issue with the licensing terms, but we resolved it. We're proceeding with the rollout. Person #39: How many entities are on the platform now? Person #41: We have 15 entities live. We're adding two per month. By the end of the year, we should have 25 live. Person #39: That's a bit slow, isn't it? We wanted 40 by the end of the year. Person #41: We're limited by the vendor's capacity. They can only onboard so many per month. We could increase the pace if we pay for more resources, but that's a budget decision. Person #39: Let's discuss in the budget meeting next month. For now, we keep the pace. We don't want to overload the entities. Person #43: Related to the integration platform, we're having issues with the data quality from the entities that are live. We're seeing, like, 15% of data records with errors. The entities blame the platform, the platform blames the entities. Classic. Person #39: That's where the governance policy comes in. Data quality standards. Clear for everyone. But until then, we have to manage manually. Person #41: We're doing the best we can. I have two people on the quality issue full-time. Person #39: Thank you. I appreciate it. Now, the next item. Any updates on the AI use case pipeline? Person #43: From my side, we have 40 use cases in development. Out of the 200 in the strategy. The top three are the traffic prediction, which I mentioned, the school placement optimization, and the health resource allocation. We're showing progress, but we need more compute. Person #39: We've discussed the compute. December is the timeline for the new cluster. What about the team? Do we have enough people to develop the use cases? Person #43: We have the people, but they're stretched. We have the data scientists, but we need more product managers. People who can translate the business needs into technical requirements. Person #39: Can we hire? Person #43: The HR approval process is slow. We submitted three requests last month. No response yet. Person #39: I'll talk to the Government Talent people. Maybe they can expedite. They keep saying they want to improve the hiring process. Let's test them. Person #41: Good luck with that. They've been saying that for two years. Person #39: We'll see. Inshallah. Anything else on AI? Person #43: We're facing a challenge with the entities' data sharing for the health resource allocation use case. The Department of Health is hesitant. They have—you know, they have concerns about privacy. Even though the data is anonymized. Person #39: That's the governance policy. Again. Everything comes back to the policy. We need it signed. Person #41: We know. We know. You said that. Person #39: Sorry. I'm just... it's the blocker for so many things. Okay. Let's move on. We have the slides to prepare. Let's keep the rest of the updates brief. Person #43, anything else? Person #43: No, I think I've covered. Just the dashboard refresh. We're on track. The entities, the 70 dashboards, we're 80% updated. The remaining 20% are with entities that haven't provided their data yet. We're chasing them. Person #39: Okay. Chase harder. The Q3 deadline is... actually, when is the Q3 deadline? Person #41: End of September. Two more weeks. Person #39: Two weeks. Can we do it? Person #43: We can try. But the entities... you know how they are. They don't prioritize our requests. Person #39: I'll send a reminder to the entity heads. A formal letter. That should help. Person #41: A letter from the DGE. That always gets attention. Person #39: Okay. Person #42. Any final points? Person #42: Just one. We have the architecture review board meeting next week. The agenda includes the new data platform for the education sector. Some of the entities are proposing their own solutions instead of using the common platform. We need to decide whether to enforce the common platform or allow exceptions. Person #39: We enforce the common platform. No exceptions. We can't have every entity building their own thing. Person #41: But what about the education sector? They have specific needs. The unified platform might not fit all their requirements. Person #39: They can adapt. The platform is flexible. We have customizations. We've done it for other sectors. The education sector is not special. Person #41: I'm just saying. We should hear their concerns before we dismiss them. Person #39: Okay, let's hear them at the ARB meeting. We'll listen. But the decision is clear. Common platform. Yalla. I think we're done here. Let's focus on the slides. Person #42, send me a draft by 2 PM. I'll review. We'll have a final version by 4. Person #43, the video. Person #41, the budget numbers. Any questions? Person #41: What about the policy? Should I chase legal? Person #39: No, I'll handle it. I have a call with them this afternoon. I'll push for the Sunday timeline. Person #41: Okay, habibi. Good luck. Person #39: Thank you. Anything else? No? Okay, let's go. We'll reconvene on Wednesday to review the slides. Person #43: The video, yes. I’ll have the draft cut by three. Just the rough version, not the final graphics. We still don’t have the voiceover from the studio, so I’ll put the placeholder music track and the subtitles. It’ll be enough for the board to get the feel. Person #39: Good. That works. It’s about the feel, not the final cut. Person #41: And the budget numbers. You want the savings figure from the procurement gate, right? The one we quoted at the World Governments Summit. Or do you want the updated number from the last quarter? Person #39: Updated. Use the updated quarter. The old number was… what, 420 million? We’re at 510 now. Don’t quote the old one, the Chairman will notice. Person #41: Noted. 510. I’ll dig it up. But the finance system still hasn’t closed the quarter, so it might be 507, 508. I’ll put a footnote. Person #39: A footnote. Fine. Better than a wrong headline. Person #42: And the slide title, do we go with “Enabling a Connected Government” or the one the comms team suggested, “Smart Government, Smarter Services”? They wanted the second one for the social media posts. Person #39: No social media language on the board deck. “Enabling a Connected Government.” That’s the message. The comms team can do whatever they want for LinkedIn—just don’t put it in the deck. Person #42: Okay, okay. I’ll keep it formal. Person #43: Also, I’m going to need the final policy draft. The one we’re still waiting on legal. The video has a segment on data sharing, and I referenced article twelve. If the wording changes by Sunday, I have to reshoot the segment. Tell me by Thursday, please. Person #39: I’ll get you a decision. My call with legal is at two. I’ll send you a note right after. Person #43: Two o’clock. Perfect. I’ll keep the evening free. Person #41: And the education platform? We still haven’t decided what to tell the board. I think we have to be careful. They’ll ask about the integration with the federal systems. The Ministry of Education has their own standards, and if we force our platform, we might have to build bridges anyway. Person #39: We build bridges anyway. That’s the cost of federation. But we don’t let them start a parallel project. If they have specific needs, we accommodate them in the roadmap. The ARB will hear them. But we enforce the principle. Person #41: Yaa, I’m not arguing the principle. I’m just saying, let’s have a message ready. A one-liner. So we don’t look like we’re dismissing them. Person #39: Good point. Person #42, put a one-liner in the deck. Under the ARB agenda item. Person #42: Consider it done. “The common platform remains the default, sector-specific requirements will be addressed through targeted extensions.” Something like that. Person #39: Perfect. Anything else? The policy call at two, the slides by four, the video by three tomorrow. Wednesday review at ten. I think we’re set. Person #43: Let me just say one thing. The data quality issue from ADERP. The entities are still arguing over who owns the master data. It’s holding up the dashboards. We’re stuck. I need the data governance policy signed, or we’ll be showing the board dashboards with gaps. And that’s not a good look. Person #39: I know. That’s the call at two. I’ll push hard. The policy is ready, it’s just legal wording. We’ll get it done. Person #43: Okay, habibi. Just making sure it’s on your radar. Person #39: It is. I’ve been on it all week. Sunday, we’ll have it. Person #41: And if legal doesn’t budge? What’s the fallback? Person #39: Then we escalate. The Undersecretary can speak to their general counsel. But I don’t think it’ll come to that. They know this is a priority. They just want their language in it. Person #41: Fair enough. Yalla, I have a meeting at ten about the data center cooling. I’m out. Person #39: Yes, go. We’re done here. Good work, everyone. I’ll send the notes. Person #42: Thanks, everyone. Person #43: Bye. Person #41: Hello? You there? Uh, I think he dropped. Anyway, I’m off.