Lance Dickson: Turning outrage into consequence: the AI cluster built for South Africa's next boycott

Listen here

Lance Dickson kept having the same conversation. A failing council. A price hike nobody could explain. Everyone complained. Nothing changed. So he built Ryze instead, a self-funded civic action platform designed to convert that outrage into organised economic consequence. In his latest interview with Irakli, the founder unpacks the six-agent AI cluster running underneath it, from a news-scanning watchdog named Argus to a legal gatekeeper named Shield, built so nothing goes live without clearing a legal check first. The mechanism isn't new. Pamphlets and word of mouth forced PUTCO to reverse a bus fare hike in 1957. Dickson just wants to run it at national scale. The urgency comes from the street, not the boardroom. "I literally got told yesterday from a petrol attendant that he doesn't know who to vote for," he says. He traces the instinct back to the country itself. South Africa, he says, is "known as one of the countries that have put boycotting on the map."

Edited transcript of the interview

Lance Dickson (00:00:00:00 - 00:00:25:22)

I literally got told yesterday from a petrol attendant that he doesn't know who to vote for. I don't think things should be like this. Especially in a in a beautiful country like this that we love. So it all starts from a point of accountability from self. Right. Working through a platform that will then give these corporations, these politicians, the accountability to start listening to us.

Irakli (00:00:26:00 - 00:00:56:06)

Lance Dickson is a Roodepoort-based AI technologist. Self-funded, with no venture money and no government backing. He's spent the past year building Ryze, a South African civic action platform, giving any citizen the tools to discover, join or start organised campaigns against corporations and government entities failing their communities. He calls it, in his own words, built to, quote, convert passive outrage into organised economic consequence.

Irakli (00:00:56:08 - 00:00:57:22)

Lance, welcome to BizNews

Lance Dickson (00:00:58:00 - 00:01:00:23)

Thank you so much for having me greatly appreciate it.

Irakli (00:01:01:01 - 00:01:17:19)

You've described this same conversation often whether it be at braais, in boardrooms or on WhatsApp, that a government or company fails the community. Everyone agrees and then nothing happens. When did you decide to take action into your own hands?

Lance Dickson (00:01:17:21 - 00:01:47:08)

Yes, 100%. So, I think this has been a constant theme amongst, you know, not only my, my peers, but I think amongst many South African citizens, we've, we've got the rage, right. It's constantly a talking point, but we do absolutely nothing with this. It always leads to a dead end. And you know, I actually got to a point where I thought, I don't think that is a valid enough point.

Lance Dickson (00:01:47:10 - 00:02:10:18)

I think we have the technology available to actually do something about it. And myself and my partner actually got to the point where we started doing something about it. So we started developing something that I think would bring more South Africans together. Because there's a there's a lot of us, we, I think are known as one of the countries that have, put boycotting on the map.

Lance Dickson (00:02:10:18 - 00:02:18:10)

And I think this platform perfectly sort of solidifies that positioning. So I'm extremely excited about it.

Irakli (00:02:18:12 - 00:02:36:13)

And you point to moments in South African history from the Alexandra bus boycott in 1957 to the consumer boycotts in the mid-1980s that were done with nothing but pamphlets and word of mouth. What convinced you that the same discipline now built on technology could work again today?

Lance Dickson (00:02:36:15 - 00:03:09:15)

So. So essentially, what happened in the past, shows us that the, the power of the majority is, is something that we can always leverage upon. So what motivated us was the fact that back in the day, there wasn't any digital platforms for these organisations, these citizens, to get together to show discontent against corporations and government entities, municipalities and so forth.

Lance Dickson (00:03:09:16 - 00:03:32:04)

So we've seen we've gotten a lot of sort of guidance from that standpoint. And due to the fact that we believe that there is a lot of change to be done in South Africa. There's a lot of potential. I love this country tremendously, and I think we have a lot of citizens that love this country tremendously as well.

Lance Dickson (00:03:32:06 - 00:04:11:03)

So, it's sort of naturally developed into an idea of combining what we're good at what we've always been good at, which is holding, organisations accountable. So I think, you know, because everybody spends the majority of time online, everybody has access to a laptop, everybody has access to a cell phone. It just brought us to a natural conclusion that, building some sort of a platform we could all meet up and show this discontent was, I think, the natural way to go.

Lance Dickson (00:04:11:08 - 00:04:21:13)

And that led us to, to the Ryze Civic Action Platform and might I add, also a first of the kind not only for South Africa, but for the world as well.

Irakli (00:04:21:15 - 00:04:40:05)

First of its kind and something that has been similarly done before in respect to, if you recall, Data Must Fall with Amandla.mobi changing data rules for for twenty million people through the campaign. What makes Ryze different and what does the infrastructure look like?

Lance Dickson (00:04:40:06 - 00:05:14:13)

Okay. So essentially it's built on immediate, sort of online sentiment. The main sort of pillar behind it is the AI intelligence, right? So we've actually built an AI cluster that, which is essentially the AI backbone for the application that's constantly scanning South African news. It's constantly monitoring social sentiment. It's validating campaigns and legal risk.

Lance Dickson (00:05:14:15 - 00:05:47:23)

It's building entity accountability profiles. And it's all with AI-grade intelligence. So it's not just a petition sort of platform. We, we looked at the numbers. This is actual, 24/7 running intelligence. So, the nice thing about it is every campaign that is started, every campaign that is joined, it's not built on a whim of, you know, this might be the right thing to do.

Lance Dickson (00:05:48:05 - 00:06:12:14)

It's actually shows, whether this is the right way to go with regards to the boycott campaign. So everything, shows, whether it is legal or not. We have a companion app as well, called Lexio, that is essentially built upon the backbone of the Constitution. as well as the Bill of Rights in every single branch of law.

Lance Dickson (00:06:12:16 - 00:06:35:09)

So it's constantly validating against it, you know, so that legal compliance is maintained at every given point as well. So accountability, not only on the other end, but on the, actual initial end, you know, the person starting these campaigns that the accountability is, is you all tracked as well. All within the legal standpoint also.

Irakli (00:06:35:11 - 00:06:59:07)

So walk us through, as an example, what happens when a contentious campaign is started and it may not be, it may not be supported by evidence. How does AI provide those guardrails in that process? I guess not just from the campaign side, but also on the other side, against the, kind of corporation or government entity that is holding accountable how does that mechanism work?

Irakli (00:06:59:08 - 00:07:00:08)

Perhaps an example.

Lance Dickson (00:07:00:12 - 00:07:23:02)

So, so, so like I said, the so the main, sort of USP of the entire platform is the, the AI backbone. Right. So, so we have a cluster of six AI Ryze agents with a main one being Catalyst. Okay. So catalyst is the orchestrator. It's the master agent that sort of coordinates all the other agents as well.

Lance Dickson (00:07:23:08 - 00:07:58:22)

It manages the intelligence board. It assigns all the tasks, it synthesizes everything. into a daily briefing. It escalates urgent alerts. So think of it as sort of a mission control. So we get that entire platform on on our end, right? All the distillation of all of these reports. But then the user that's actually starting these campaigns will get it into a sort of extremely distilled response in terms of, whether there is accountability to be held right.

Lance Dickson (00:07:58:22 - 00:08:24:14)

And, and whether we work within that sort of legal frameworks as well. Okay. So all of that is guided by the other five agents with the main news intelligence agent being Argus. So this agent essentially is just tasks to, monitor the news 24/7. Right. That's your Daily Maverick. News24, TimesLIVE, GroundUp, EWN and BusinessLIVE.

Lance Dickson (00:08:24:16 - 00:08:52:18)

It essentially just extracts all of that top stories, extracts all of the key entities. And then it nothing is built on rumour, right? Everything is built on real news that, that could be, sort of searched online. So, so we, we, we try to build as many accountability mechanisms into the actual cluster as possible so that there's no misinformation spread at any particular time.

Lance Dickson (00:08:52:19 - 00:09:28:11)

So that's that's the second agent. The third agent. Vox is the social sentiment agent. So this agent essentially tracks the social sentiment between all of the social platforms. That being, X, formerly Twitter, Reddit, public Facebook groups, you know, all of that sort of, social online groups as well, just to see what the general public is sort of, feeling around a particular issue that could be bothering them, compiling those reports and then passing it over to, to catalyst.

Lance Dickson (00:09:28:12 - 00:09:55:22)

So you see, there's multiple sort of levels to this thing. There's another agent as well, the Veritas agent. So that's the accountability data. Okay. So it builds essentially accountability profiles, you know, targeting entities pulling from the CIPC, the IEC Municipal Money, the Auditor-General, National Treasury procurement data, all of these sort of aspects as well. So that's another layer that we built into it.

Lance Dickson (00:09:55:22 - 00:10:29:08)

The next layer is a shield, right. This does the trust and legal screening. So essentially every single campaign goes through, shield before it goes live. Right. This is where it checks for any, factual basis defamation risk, legal exposure, blocking campaigns that, could, could, sort of, have some sort of legal liability, just to ensure that accountability, not only on the other end is, is sort of, vice versa.

Lance Dickson (00:10:29:08 - 00:10:54:08)

And reflected back to the user that's actually starting the campaign. So that's, that's the other layer of accountability. And then the last one, Irakli, is the scribe. So that's the content generation agent. It essentially just summarizes everything into brief action points, structured content. Everything is essentially captured in the back end. And then users can actually draw these reports as well.

Irakli (00:10:54:09 - 00:11:09:11)

There's a lot of talk currently around open models and closed models. How does your solution approach that problem and how do you account for bias in your models, especially scaling these agents?

Lance Dickson (00:11:09:12 - 00:11:36:22)

Yeah. So it's all, closed-source, the models that we're using. We train it to be as bipartisan as possible. The app itself has no bias towards any side, whether that be a municipality, any particular, politician, any corporation. It is bipartisan by, by nature. And that is one of the sort of accountability points that we've sort of trained from the off.

Lance Dickson (00:11:37:00 - 00:12:12:12)

There's actually one of the first things that we train the models on, and that is only to work on pure evidence, right? So essentially and mitigating any sort of polarity between what could be real and what could be fabricated, because we know that the, some, some actors, some, some bad actors out there. We know the, the platform has the possibility of being taken advantage of and that's why we pulled all of these accountability points into the system itself, right down to the training of the model as well.

Irakli (00:12:12:14 - 00:12:29:22)

I'm glad you mentioned that. It is a closed, model, given the cost of using AI is growing globally and there is sticker shock in some cases. How do you manage the cost side given the complexity of the architecture?

Lance Dickson (00:12:30:04 - 00:13:01:00)

So, so because so the like you mentioned earlier, everything is privately funded. We haven't, you know, gotten any outside funding at all. So it's not cheap. It it does cost money. Our agency is essentially covering all of these costs. But in terms of, mitigating the costs as much as possible. We have like I said, we have used our, our own, server infrastructure.

Lance Dickson (00:13:01:01 - 00:13:25:01)

And this way as well, it allows us to mitigate the, the, security aspect as well. Data leakage is one of the most sort of, important aspects when it comes to AI agents. So, so what we've sort of done to mitigate this as well is to, like I said, have everything on our own servers running through our own servers as well, so that everything is managed by us.

Lance Dickson (00:13:25:01 - 00:13:47:21)

And all of the data, does not escape at all. The major thing with that is it's helped us to keep the costs low. It's helped us keep the costs manageable as well. And, but yeah, these expenses are growing. I think with regards to the platform getting bigger and bigger, we're going to have to be a bit more smarter with the way that we deploy these agents, because remember they are working 24/7.

Irakli (00:13:48:01 - 00:14:01:02)

But it seems you're on the right track and keeping it local and also, on local servers and also managing that compliance element and thinking ahead and managing, you know, personal information. What's next for Ryze?

Lance Dickson (00:14:01:03 - 00:14:29:07)

So, look, essentially we want to, try and get this in in front of as many eyeballs as as possible. Our beta will be our live beta will be ready within the next seven to 10 days. Thereafter we have the Play Store application as well. That will take another two weeks. Thereafter. So essentially the main goal, Irakli, is to just give the power back to the citizens.

Lance Dickson (00:14:29:07 - 00:14:54:06)

Right. We will be constantly, myself and my partner be constantly speaking to people on the streets, like whether that be at a petrol station or whatever. The case may be. And there's this sentiment, you know, everybody is sort of on the same page. We're in this together and, and people don't know where to where to go to, I literally got told yesterday from a petrol attendant that he doesn't know who to vote for.

Lance Dickson (00:14:54:08 - 00:15:16:12)

I don't think things should be like this. Especially in a in a beautiful country like this that we love. So it all starts from a point of accountability from self. Right. Working through a platform that will then give these corporations, these politicians, the accountability to start listening to us. Right. That's what we want.

Lance Dickson (00:15:16:12 - 00:15:34:18)

We want to be heard. And I think this is going to be that wake-up lesson for them to start hearing us out. And the only way we can do that, is to try and get this in as many hands as possible. And that's my that's my goal. So that's my mission. Over the next few months.

Irakli (00:15:34:20 - 00:15:42:16)

Wonderful. Lance Dickson, helping South Africans hold companies and councils to account before they even reach the ballot. Thank you for joining us.

Lance Dickson (00:15:42:22 - 00:15:44:04)

Thank you so much for having me.