Ola Yeenca

@SpaceXSI Community Ambassador

[Ola Yeenca @Ola_Yeenca]

so we kind of lost the plot for a while. we were so deep into conversa that sme analytica became “the whatsapp bot company”. that was never the idea. we’re going back to what we actually wanted to do, go into a business, connect ai to what they already run, and leave it working.

[SME Analytica @smeanalytica]

Your AI pilot didn’t fail because of the model. It failed at the handoff nobody wrote down.

We’ve been building SME Analytica for more than three years, and in that time we shipped Conversa, which connects an estate agency’s WhatsApp to its CRM, and Facturia, which we use every day to run our own invoicing. Both are in production. But honestly, we were working so much on Conversa that we drifted a bit from what SME Analytica was supposed to be from the start.

Because when we talk to companies, they almost never ask us for another tool. What they have is an ERP, an Excel, some old system nobody wants to touch, and an AI idea that stays a pilot because nobody connects it to all of that.

So we’re going back to that. Our engineers go into your operation, integrate AI with what you already have, and leave it deployed and working. Conversa and Facturia are still here, and now they’re the proof that we know how to do it.

And it’s not just for estate agencies, because the work always starts the same way, understanding how your business works on the inside, whether that’s a workshop, a public office or a finance team.

Embed. Integrate. Ship.

[Ola Yeenca @Ola_Yeenca]

https://t.co/h7BOyZhcIP

[Ola Yeenca @Ola_Yeenca]

No way I twisted my ankle

[Ola Yeenca @Ola_Yeenca]

This wan go soon waka enter bush 😂🙂‍↔️

[“𝐓𝐨𝐩-𝐓𝐢𝐞𝐫 𝐌𝐲𝐤𝐞𝐡𝐥“🃏🇳🇬 @Its_Mykehl]

😉

[Ola Yeenca @Ola_Yeenca]

https://t.co/l3ZgNDJJkB

[Ola Yeenca @Ola_Yeenca]

@HerRoyally 😂😂

[SME Analytica @smeanalytica]

in our setup, one privacy rule gave us 0 email leaks in 50 agent trials. harder tests: 0/120.

but when a vendor required the email, 0/11 tasks finished. the agent stopped.

we need to test the human handoff too. grok 4.6, mocked tools, fake data.

[Ola Yeenca @Ola_Yeenca]

@Ibn_momoh The wicked run whilst no one chaseth 😮‍💨

[Ola Yeenca @Ola_Yeenca]

@_thatgyalzozo @_kofs Congratulations Bonita

[Ola Yeenca @Ola_Yeenca]

@HerRoyally I come from the lineage of twins… it’s part of my ancestry 💯

[SME Analytica @smeanalytica]

why do so few ai pilots make it into the daily work?

s&p global's 451 research surveyed 1,006 it and business leaders in north america and europe at the end of 2024: 42% of companies had dropped most of their ai initiatives before production, up from 17% a year earlier.

the mit "95%" is shakier than the headlines. it's 52 interviews and 153 surveys and the authors call it directional, so we show the funnel instead.

the same reasons keep coming up: data that isn't ready, nobody owning it, a demo that never meets real systems, cost without a clear return, and security and legal worries. every number has its source on screen.

what stopped your last pilot?

[Ola Yeenca @Ola_Yeenca]

this is from February. Sánchez said it in a speech at the world governments summit in dubai on 3 feb 2026, it’s not “new rules starting next week”, and an announcement isn’t a law.

if you run a business in spain, check the date before you panic.

[JackTheRippler ©️ @RippleXrpie]

🚨SHOCKING: 🇪🇺🇪🇸 NEW RULES IN SPAIN!

Spanish PM Pedro Sánchez:

"Starting next week, my government will implement the following actions:

1. We will change the law in Spain to hold platform executives legally accountable for many infringements taking place on their sites.

2. We will turn algorithmic manipulation and amplification of illegal content into a new criminal offense.

3. We will implement a hate and polarization footprint system to track, quantify, and expose how digital platforms fuel division and amplify hate.

4. Spain will ban access to social media for minors under the age of 16. Platforms will be required to implement effective age verification systems — not just checkboxes, but real barriers that work.

5. My government will work with our public prosecutor to investigate and pursue the infringement committed by Grok, TikTok, and Instagram."

[SME Analytica @smeanalytica]

claude code was putting people's account email in its user-agent on web calls. the email got into its context to stop it inventing @author tags, so fixing one bug made the leak

a hook that pattern matches the email can't catch a script the agent writes itself. if an agent touches client data, outbound calls need an allowlist

[Ola Yeenca @Ola_Yeenca]

@Idntgiveafuvk The FAANGS

[Ola Yeenca @Ola_Yeenca]

@Idntgiveafuvk @4thsonXhail So humans will still be needed maybe not heavily dependent, also more like researchers and or babysitters

[Ola Yeenca @Ola_Yeenca]

@Idntgiveafuvk @4thsonXhail I hear your point but you should also understand that, they ai agents can’t improve themselves… and they need to be properly trained, which is not possible because we have good and bad data’s

[Ola Yeenca @Ola_Yeenca]

@Idntgiveafuvk AI is just a the start of what’s yet to come… beside, tech is always and will keep advancing, evolving and increasingly be innovative… we still have AGI, super intelligent and quantum intelligence

[SME Analytica @smeanalytica]

facturia first thing again. no demo energy, just has to be right. that's the bar i use for any agent work now

[SME Analytica @smeanalytica]

every small company has that one rule nobody wrote down. the client who always gets a different rate, the thing the manager fixes by hand every friday.

an agent rewriting the system won't know it exists. run the old one next to the new one for a few weeks before you switch