Prodigy AI Solutions

🚀 We’re excited to share that we’re participating in the CINECA training course on HPC Build Systems & Package Managers - a three-day hands-on program focused on the tools powering high-performance scientific software.

The course covers:
⚙️ Makefiles
⚙️ GNU Autotools
⚙️ CMake
⚙️ Python packaging
⚙️ Spack for HPC environments

As we continue building AI systems and graph-based intelligence platforms like Verbis Graph, strengthening our expertise in scalable software infrastructure and HPC workflows is incredibly valuable.

Always learning. Always building. 🚀

Why do different industries need different ontologies?

Because the same word can mean completely different things depending on context.

Take the word "asset."

In finance, an asset could be cash, securities, receivables, or property.

In manufacturing, an asset could be a robot arm, production line, machine, or sensor.

In legal documents, an asset might appear inside contracts, due diligence reports, or disputes.

The word is the same.

The meaning isn't.

This is one of the biggest challenges in enterprise AI.

Modern AI systems are very good at finding similar text. They're getting better at connecting entities through GraphRAG and knowledge graphs.

But similarity isn't the same thing as understanding.

A system can retrieve two related paragraphs and still miss the business meaning behind them.

That's why industries develop domain-specific ontologies.

Healthcare uses clinical concepts and relationships.

Finance uses financial concepts and regulatory relationships.

Legal systems use obligations, rights, parties, jurisdictions, and clauses.

Manufacturing uses machines, sensors, production lines, batches, and maintenance events.

The ontology provides a shared understanding of what things are and how they can relate to one another.

But there's another important lesson:

Ontology alone doesn't solve everything.

If your extraction layer doesn't understand document structure, tables, sections, and context, your graph can still create incorrect relationships.

Good AI systems need both:

1. Accurate extraction
2. Meaningful ontology

At Verbis Graph, we've been exploring how document hierarchy, layout-aware extraction, and ontology constraints can work together to preserve context and improve reasoning.

Not because graphs are fashionable.

Because understanding meaning matters more than connecting keywords.

What industries have you seen struggle most with domain-specific knowledge representation?

We’re excited to bring **Verbis Graph Investigator** to the Google for Startups AI Agents Challenge.

Built on **Verbis Graph**, our graph-based knowledge layer, the project explores a new kind of AI agent: one that investigates with evidence, not just retrieves text.

First use case: helping investors and venture funds turn startup data rooms into grounded diligence insights.

From documents to graph intelligence.
From claims to evidence.
From search to investigation.

We’re always looking for ways to improve Verbis Graph verbisgraph.com, and one of the next big steps for us is adding an ontology layer.

This may sound like a very technical thing, but the impact is actually very practical.

Today, many AI systems can search documents, extract information, and even connect entities through graphs. That already helps a lot.

But in real enterprise environments, the hardest part is often not finding similar text. It is understanding what the information actually means in context.

For example:

a policy is not the same as a procedure
an obligation is not the same as a recommendation
a risk is not the same as a control
a symptom is not the same as a diagnosis

A graph shows that things are connected.
An ontology helps explain what kind of things they are and how those relationships should be interpreted.

For Verbis, this matters because we work with complex document collections, especially in areas where trust, traceability, and consistency are essential.

Adding ontology can help us improve:

answer precision
multi-document reasoning
explainability
consistency across teams and documents
productivity in knowledge-heavy workflows

In simple words, it helps people spend less time searching and interpreting, and more time making decisions.

That is the direction we believe enterprise AI should take: not just retrieving more information, but understanding knowledge in a more structured and useful way.

🎓 Something we keep seeing across EdTech platforms in Europe and Latin America and we think it's worth talking about openly.

Most education platforms have solved the product problem. The content is good. The UX is solid. The teachers are talented. But there's a silent operational crisis happening underneath: the data infrastructure hasn't kept up.

Think about it this way. A typical EdTech platform generating 50,000+ learner interactions a week is running on a CRM, an analytics tool, a support inbox, and a billing system - all operating in parallel, all telling a slightly different story. Nobody on the operations or pedagogical team can see the full picture in real time. At-risk students get identified too late. Reporting takes days. Decisions get made on gut feel and partial data.

The good news? This is a solvable infrastructure problem. Not a product problem.

What we build at Verbis Graph verbisgraph.com is the intelligence layer that sits across those systems: normalizing data, resolving duplicate records, watching behavioral signals continuously, and surfacing insights in plain language — so that a pedagogical coordinator, a retention manager, or an operations lead can actually understand what's happening with their learners. No SQL. No analyst waiting time. No spreadsheet archaeology.

The result isn't just efficiency (though cutting manual reporting time from 15 hours a week to under 5 matters a lot). It's visibility. The kind of visibility that lets you intervene before a student disengages, before a family decides not to renew, before a cohort-level problem becomes a revenue problem.

In a subscription education business, retention is everything. Even a 2-point improvement compounds meaningfully over 12 months. And the data to achieve it? It's already there. In your systems. Right now. verbisgraph.com

If you're building or scaling an EdTech platform and this resonates - we'd love to talk. Drop us a message or share this with someone who needs to hear it. 👇

Big ideas need big engines. ⚙️

Our commitment to the builder community is simple: providing the high-performance computing (HPC) power and expertise required to scale the next generation of AI.

With the strength of the Leonardo supercomputer, we’re enabling developers to push the boundaries of the Verbis Graph Engine, turning ambitious concepts into enterprise-ready reality.

Every successful startup begins with the right foundation.
SBDC helps founders navigate business structure, strategy, and growth - empowering thousands of startups, including ours, to move forward with confidence.

While developing Verbis Graph, our retrieval‑based knowledge system, we realized how challenging it actually is to integrate AI solutions into AWS and Microsoft infrastructure — especially if you want to publish on the major cloud marketplaces.
Because of that, we decided not only to build our own product, but also to help other teams who want to be present on AWS Marketplace and Microsoft Marketplace. https://aws.amazon.com/marketplace/pp/prodview-ia2d65ogwkojm

At the same time, for our own internal operations, we’ve been building automations and AI assistants that will soon be unified into a single automation + AI agents flow. These tools are already helping us scale faster — and we’re starting to share some of them publicly.

Available now:

Our first free n8n automation templates https://n8n.io/creators/kroll-elly/

More automations coming soon, and we’d love to hear what would help your workflow

We’re also building a multilingual AI co‑sell assistant to support global partners and customers in real time. Since we’re targeting a global market, this tool is becoming essential for us.

If automation or AI agents could save you time, I’d love to hear what you’d want us to build next.

In Pharma and Life Sciences, the challenge is rarely a lack of information.

The real challenge is that critical knowledge is spread across too many documents, systems, and teams. Regulatory content, clinical materials, SOPs, quality records, scientific publications, and internal reports often remain disconnected, making it harder to find the right information quickly and use it with confidence.

This leads to slower decisions, duplicated effort, repeated internal requests, and unnecessary friction across departments.

Engine helps organizations turn fragmented documents into connected, trusted, and usable knowledge.

For teams that want to evaluate the value in a practical way, we also offer a 1-month PoC/Pilot using real documents and real use cases.

Explore Verbis Graph Engine here: https://engine.verbisgraph.com/pharma-life-science/en