You Can't Buy Your Context Graph

Tony Seale — the knowledge graph engineer better known to half of LinkedIn as “the Knowledge Graph Guy” — posted something this week that we think every analytics leader should read twice:

Everyone is suddenly racing to sell you your own context. But a context you can buy is a context your competitors can buy too.

He’s right. And the timing is almost comical. Barely a week passes now without another “context layer,” “context graph,” or “context foundation” for your AI agents — your organization’s reasoning, its exceptions, its hard-won workarounds, packaged up and ready to plug in. The whole industry has finally agreed on the thing we’ve been saying since we started Spotonix: context is the missing piece.

But agreement on the problem has produced a wave of products that quietly get the solution backwards. So we want to be unusually direct about where Spotonix stands, because Seale’s warning cuts straight through our category — including, if we’re not careful, us.

The thing you can buy isn’t a moat

Seale’s argument has a hard edge to it. Context that arrives pre-packaged from a vendor is, by definition, available to everyone who writes the same check. If your “institutional knowledge” ships in a box, it isn’t institutional and it isn’t knowledge — it’s a dataset with a logo on it. As he puts it: the moment your context is defined outside your control, you are no longer a distinct system.

This matters enormously for business intelligence, because BI is where the temptation is strongest. Every analytics vendor would love to sell you a starter pack of “industry metrics,” a library of “best-practice definitions,” a pre-trained model that already knows what “revenue” and “churn” and “active customer” mean. It demos beautifully. It’s also worthless as a differentiator, because your competitor sees the identical demo and buys the identical definitions.

The uncomfortable truth is that the valuable context in any organization is the context nobody else has: how your finance team defines a habitual buyer, why your operations org excludes returns from net revenue in the Northeast region, the precedent from the Q3 board meeting that quietly changed how everyone reports retention. None of that comes in a box. It accumulates, from the inside, one decision at a time.

So what does Spotonix actually sell?

Not your context. We’re emphatic about this. We sell the machinery that lets you build and own your context — and we make sure the context that machinery produces belongs to you, not to us.

There’s a clean line between the two, and it’s the whole game:

We provide the engine — the discovery, composition, validation, and learning loop that turns ambiguous business questions into reusable semantic objects. That engine is our product. You can buy it. Your competitor can buy it too. Fine.

You accumulate the graph — the actual decision traces, interpretations, segmentations, and calculations that the engine produces as your analysts work. That graph is built from your data, your definitions, your reasoning, your resolved ambiguities. It is unique to you the day you start, and it diverges further from everyone else’s every single day it runs. That is the asset. And it never leaves your control.

To borrow Seale’s framing: you can buy the engine; you cannot buy the boundary. The boundary — the meaning that makes your organization a distinct system — only exists because it was woven from the inside.

Why a decision trace can’t be commoditized

Here’s the mechanism that makes this real rather than rhetorical.

When Spotonix answers a question — say, “Which stores are losing habitual buying customers?” — it doesn’t just produce a number. It captures the full reasoning trace behind that number:

• The interpretation. “Habitual buyers” resolved to customers with $500+/quarter spend, based on the segmentation your team validated last quarter — not a generic definition we shipped.
• The composition. The answer was assembled from named building blocks: your Habitual Buyers segmentation, a QoQ Customer Count Change calculation, a Store × Quarter dimension.
• The validation. The ambiguity between “frequency-based” and “value-based” was surfaced to a human and resolved deliberately, not silently guessed.
• The trace. Every block points back to its definition. Every step is auditable.

A generic context library can’t produce any of that, because every line of it is downstream of decisions only your people can make. A vendor can sell you a definition of “churn.” A vendor cannot sell you the forty clarifications your team made about your churn over six months — because those clarifications didn’t exist until your team made them. The graph is a record of judgment, and judgment doesn’t ship in a box.

This is also why text-to-SQL, and the current crop of “ask-your-data” chatbots, will never produce a durable asset. They generate a query, return an answer, and throw the reasoning away. Each question is an island of context that disappears the moment it executes. There’s nothing to own because nothing accumulates.

The part Seale gets exactly right: open standards

Seale doesn’t stop at “build it yourself.” He adds the sharper point: a boundary made of proprietary, vendor-specific formats is just a smaller cage. Context locked in our format would be no more yours than context locked in someone else’s.

We take this seriously, because it’s easy to claim “you own your data” while engineering things so the meaning is hostage to our software. Ownership that evaporates the day you stop paying us isn’t ownership. So the semantic objects Spotonix builds — segmentations, calculations, business analyses — are defined in terms of your identifiers, your warehouse, your definitions, and they remain legible and portable outside our runtime. The graph describes your business in your terms. We’re the engine that helps you write it down; we are not the landlord of what it means.

As Seale puts it, and we’ll happily steal the line: openness is not the opposite of ownership. It is the only durable form of it.

Where this leaves us

There’s a flywheel here, and it’s the reason this compounds instead of commoditizing. Question one builds two new blocks and takes forty-five seconds. Question fifty discovers six existing blocks, composes instantly, and is more correct than question one because it inherits every clarification that came before. The graph doesn’t just remember — it accelerates. And because it’s built from the inside, every day it runs your context looks less like anything a competitor could buy, not more.

So if you take one thing from Seale’s post and one thing from ours, let it be this. The race to “sell you your context” is mostly a race to sell you a thing that, the moment it’s purchasable, stops being yours. The context worth having is the context you build — discovered, composed, validated, and learned from your own decisions, in open terms you control.

We’d love to sell you the engine. The graph is yours.

---

Read our whitepaper: The Interpretation of Business Questions

Demo access: try Spotonix