Making Enterprise AI Actually Work: Why We Invested in Solid

Everyone wants AI.

For the last two years, we’ve been living through an AI gold rush. Every board deck has an AI slide. Every team has an AI initiative. Every vendor has an AI story.

Now comes the harder phase: realization.

Enterprises are discovering that AI doesn’t fail because models lack power; it fails because they lack context. A model can pass a Bar exam, but it doesn't understand your business.

The 'single source of truth' usually exists only in slide decks. In production, AI collapses when it hits the messy reality of enterprise IT: fragmented data, drifting logic, and tribal knowledge that never made it into the documentation. Without that business-specific context, even the most powerful model is just guessing.

That’s the gap Solid is built to close.

Solid has just launched publicly with $20M in seed funding, backed by Team8 and SignalFire, and a team of ~20 people. It is building what we see as a new category in the AI stack: AI enablement for enterprise data - a layer that gives AI a single, trusted, always-up-to-date understanding of how your business actually works.

The problem: AI doesn’t understand your data

Every enterprise wants the same three things from AI: to chat with their data instead of waiting on reports, to automate workflows that still depend on manual checks and decisions, and to deploy agents that can reason, decide, and act across systems.

On paper, the models can do all of that. In practice, they hit the same wall: AI doesn’t know what your data means.

It doesn’t know which numbers the business trusts, how metrics are really defined, how rules change between products or regions, or how data connects across systems. That “business meaning” is scattered across dashboards, SQL, documentation, and tribal knowledge, often inconsistently.

The result is predictable: AI gives confident but conflicting answers; workflows break when business logic changes; agents can’t act safely because they don’t know which numbers or rules to trust; and data teams become the bottleneck for every AI initiative.

Most organizations try to fix this manually: maintaining definitions in BI tools, semantic layers, and docs. It works in a small domain, then breaks as soon as definitions change, new metrics are introduced, or another team comes online.

This is the Data Understanding Gap. Until it’s solved, AI stays stuck in pilots and demos.

Solid’s thesis: AI enablement for enterprise data

Solid exists to solve this exact problem.

The core idea is simple: AI only works at scale if it has a single, continuously maintained understanding of your business data.

Solid provides that understanding by creating a single source of truth for business meaning, and then doing the hard part: automating how that meaning is created, tested, and maintained over time. Instead of asking data teams to constantly update definitions by hand, Solid learns the structure and meaning of enterprise data—including all the edge cases that never show up in clean demo environments—then automatically generates, tests, and deploys the semantic models AI needs to connect safely to that data. It keeps those models continuously in sync as metrics, rules, and business logic evolve, and it pairs this automation with human expertise to handle exceptions, capture nuance, and improve accuracy over time.

On top of this, Solid builds a context graph over an organization’s data: a live map of metrics, relationships, and rules that reflects how the business actually operates. AI systems, dashboards, and agents can rely on this context graph as the backbone for reasoning and decision-making.

Solid is not trying to replace the existing data stack. It works with modern platforms like Snowflake, Databricks, and BigQuery, and sits before AI or BI systems as the enablement layer that translates raw data into business-ready information AI can safely understand and use.

The impact is measurable. Across deployments, teams see AI+data accuracy rise from roughly 20–30% to above 85%, while cutting 50–70% of the work hours required to maintain and test semantic models and compressing AI initiatives from one–two years down to three–six months. This is AI that actually works in real enterprise conditions—not just in controlled demos.

Why now: from AI hype to AI that delivers

We’re in an interesting moment.

The initial AI euphoria, “just plug your data into a model and it will figure it out”, 

is giving way to a more sober reality. Enterprises are asking if they can trust the answers, whether systems will still work when a metric changes or a new region is launched, and whether they can operationalize AI without overloading already-stretched data teams.

This is where AI enablement comes in.

If the last decade was about building data infrastructure—collecting, storing, and processing data at scale—this decade is about enabling AI: giving it the context it needs to be safe, reliable, and operational.

Solid fits directly into that shift. It takes a problem enterprises have been wrestling with manually for years—keeping business meaning in sync across tools—and turns it into an engineering system. In doing so, it shows what the post-hype AI era looks like: less magic, more results.

The founders: two second-timers solving a problem they’ve lived

Our conviction in Solid is rooted in the founders and the problem they chose to solve.

Solid was founded by Yoni Leitersdorf (CEO) and Tal Segalov (CTO)—both second-time founders and alumni of Israel’s elite Unit 49.

Yoni previously built and exited a network security company, giving him deep experience building and scaling critical infrastructure products in demanding environments. Tal previously built and exited a data company, bringing hands-on experience with data platforms, semantics, and the realities of enterprise analytics.

They’ve known each other for years, worked together in one of the most technical units in the world, and share a clear view: AI only becomes trustworthy when the data underneath it is aligned, tested, and continuously maintained.

Solid comes directly from that experience. They saw how misaligned definitions quietly undermine trust in analytics and AI, how much time data teams lose to rework and clarification, and how, as AI adoption accelerates, the gap between what the business means and what the systems think it means only grows.

Their answer is not another dashboard or model. It’s a new layer in the stack: a dynamic, context-driven semantic foundation built for AI.

Looking ahead

Solid is still early, but the trajectory is clear.

With $20M in seed funding, a ~20-person team, and backing from Team8 and SignalFire, the company is focused on helping enterprises move from AI ambition to AI impact. That means enabling business leaders to make decisions based on numbers everyone trusts, giving teams the ability to chat with data and run AI-powered workflows without constant hand-holding, freeing data and analytics groups from low-value firefighting, and allowing AI systems to operate on a single, reliable understanding of the business rather than fragmented, conflicting definitions.

In a market where everyone talks about AI, Solid is working on the unglamorous but critical part: making sure AI actually knows what your data means. That’s why we chose to invest, and why we believe Solid will be a key player in the emerging category of AI enablement for enterprise data.