GoodData.AI

05/14/2026

Ungoverned definitions. Inconsistent logic. Context that differs by team, tool, or query….Sound familiar?

When AI agents don't have a stable foundation to reason from, outputs become unreliable, and organizations stall before they ever reach production.

This is the context problem.

And it's the real barrier between AI experimentation and AI that organizations can actually act on.

On May 27, Peter Fedoročko and Jan Panský are hosting a 30-minute live conversation built for data leaders and practitioners who are done with pilots and ready to think seriously about what production-grade AI actually requires.

Save your seat: https://hubs.ly/Q04gpgZl0

05/13/2026

Over-skilling your AI agents is a real problem.

If the LLM already knows how to do it, writing a skill for it doesn't add value. It adds noise.

The real opportunity is encoding what's specific to your business: your processes, your logic, your edge cases.

Peter Fedoročko and Rosta Striz explain where the line is. Watch here: https://hubs.ly/Q04gpjyb0

05/12/2026

Launching an analytics agent shouldn't take a sprint.

With Agent Builder, you start from a template or plain-language description, connect it to your business data automatically, and get an observable framework that handles multi-step reasoning.

Code optional. Governance built in. Ready in minutes.

See how: https://hubs.ly/Q04fMvVl0

05/12/2026

What's the difference between AI that answers and AI that acts?

One waits to be asked. The other monitors continuously, applies your business rules, captures decisions, and surfaces what matters without a prompt.

Most analytics AI is still in the first category. Our newest whitepaper on AI architecture gets into why that gap exists and what closing it actually requires. Take a look: https://hubs.ly/Q04fMysf0

05/11/2026

Not every user needs the same AI agent.

With our new Agent Builder, configure multiple agents with custom skill sets, personalities, and user group access, all from a single hub.

Your sales team gets an agent focused on pipeline and revenue metrics. Operations gets one scoped to efficiency and logistics. Product gets another built around usage and retention data.

Want to limit certain capabilities for some users while giving others full access? You can set that up in just a few clicks.

Each agent stays grounded in your governed semantic layer, so responses are consistent and rooted in real business definitions.

Watch the demo from Principal Product Manager Rosta Striz. 👇

05/10/2026

Every time you need a different AI experience for a different user group, someone files a ticket, waits for a PR, and queues up a deploy.

That's the real cost of managing AI assistant configuration in code.

With GoodData.AI’s AI Hub, admins can define named agents — each with its own personality, skills, and knowledge access — and push changes live without a single line of code.

Read how it works:

Manage GoodData's AI assistant variants from a UI with Agent Builder. Configure skills, personality, and knowledge access per user group, no redeploy.

05/08/2026

🧑‍💻 Writing the best code used to be the priority. But not so much anymore. AI has changed what good engineering actually looks like.

So, what's taken its place? According to GoodData.AI CTO, Jan Soubusta: language skills, deep product understanding, broad architectural knowledge, the ability to direct AI well, and to recognise good output when it comes back.

For Jan, AI shows up across everything he does — software development, architecture design, maintenance, learning, presentations. If there's a task, AI is part of how he approaches it.

The old constraints are mostly gone. Slow onboarding, boilerplate code, and limited parallelism are all handled. Prototyping, large redesigns, and technical debt removal are now faster. He can run five times more tasks in parallel than before.

The parallelism is real. But so is the exhaustion. The human brain wasn't built for this, and Jan is finding that hitting the right balance takes time.

On the product side, the impact is visible. More features, faster. And room to expand into agentic analytics alongside the BI capabilities GoodData.AI has always delivered.

At GoodData.AI, this is what building with AI actually looks like. Not just faster ex*****on but a fundamentally different way of working. Stay tuned for more behind-the-scenes stories from our team.

05/08/2026

Big Data & AI World was two days of sharp conversations, big questions, and a few strong coffees.

We met engineers rethinking how AI fits into their data stack, product teams figuring out governance at scale, and everyone asking some version of the same question: how do we make this actually work?

Good to be in the room for it. See you next time! 👋

05/07/2026

"Skillification" is apparently the word of 2026 so far.

Everywhere you look: skills. Everyone's reading about them, hearing about them, being told they should be writing them. And along with that comes the noise: that agents are dead, that the companies that survive AI will be the ones that codify every process into skills.

Peter Fedoročko and Rosta Striz cut through it in this session. What skills actually are, why the architecture around them matters more than the hype, and how separating skills, agents, and tools is what makes agentic systems governable and maintainable at scale.

Watch the full session: https://hubs.ly/Q04fql5M0

05/06/2026

GoodData.AI is a direct reflection of what we’ve been building: the analytics foundation serious enterprise AI deployments need

Every decision has been heading here.

Welcome to GoodData.AI: https://hubs.ly/Q04fq5c50