What Does Syntheia Sell? Why Not Gen AI?
This post is about change.
“What do you sell?”
One of the most common questions we get asked is “what does Syntheia actually sell?” And it's one that has always given us trouble to answer.
For a long time, the answer was: a lot of things! Comparison tools, fund tools, a clause bank, a Word add-ins, email agents, custom work where it aligned with our roadmap, and underneath all of it, a platform that could do multiple things… but all held together by a single idea:
“Take deal documents, break them down to their component parts — the Lego bricks — and then reassemble those bricks as you need them. ”
That's the essence of what we do. That has never changed.
But, it's just a hard thing to explain quickly. We just took three paragraphs to tell you.
Gen AI, the God of Hammers
The question changed in the last two years. It has gone from "what do you sell?" to “why don’t you sell Gen AI?”
Why haven't you built a contract review tool that uses GenAI? A playbook generator? A negotiation intelligence platform? Wouldn't that add to what Syntheia can do?
And the answer is: yes, it would add to what Syntheia can do. But it wouldn't differentiate us from the tens of other tools already doing exactly that — most of them by plugging documents into a language model and returning the output.
We already make use of Gen AI — like inside our Email Agents. We are just not all-in on Gen AI.
The way we see it, GenAI is one capability among many. It has its place, but it is not the default answer to every problem. There is a saying that goes “to a hammer, every problem is a nail.” We want to wield more than just a hammer.
What we see in the market — and why we've been cautious
If you look at the legaltech market right now, you'll see a pattern. Multiple companies — well-funded, credible, with real customers — are building essentially the same product: AI-inferred intelligence from a corpus of documents.
The pitch is compelling — give us your documents and we'll tell you the answer. Gen AI will analyze the text. Gen AI will synthesize the information. You will know the chronology. You will know what market looks like. The tool will flag inconsistencies. The tool will flag the risks. Lawyers have always wanted this. Now there's something that can do it with auto-magic.
As at February 2026, the models have improved to such an extent that it is nearly indistinguishable from real intelligence.
Except, there are still severe limitations.
The problem starts when you give the model more data than it can hold in the context window. For a single document, or a small set of documents that fits in the context window, Gen AI is impressive and good. Once you exceed the context window, you have to compress information. If you exceed the context window by a lot, then you have do compressions of compressions, and summaries of summaries. But each pass costs fidelity. Something is always lost. The builders all know this.
Put in one document and ask for an analysis — you will get something useful.
Put in 10 documents and ask for an analysis — you will get something passable.
Put in 100 documents and ask for an analysis — the quality is unacceptable.
This is why we've been so cautious. A core principle for Syntheia is that we are providing lawyers with real text. Text traceable to the source document. Reliable data. Certainty.
Of course, there are people betting that GenAI will grow exponentially — that context windows will expand, costs will drop, and eventually the technology will handle very long documents across very large sets. That might happen. But even if that happens, it is not an efficient way to work. There is a lot to be said for a system that breaks documents into pieces, cross-references and links them properly, and retrieves just the parts you need. That is what we built.
The thing we built by accident
When we released Super Comparer as a self-service SaaS tool in August 2025, something unexpected happened.
Within the first couple of weeks, about ten or twelve legaltech companies reached out to ask about the API underneath it: Can we use your comparison engine in our tools?
We said yes, but explained what that actually involved — ingesting their documents, breaking them down, structuring them properly, then returning something useful.
Then, a few of them said: fine, but if you can do that, if you can give us a clause bank, if you can break documents down to that level, then what else can we do with it?
One company asked: can you deconstruct a DOCX file and reconstruct it? Give the right textblocks, move the Lego bricks around, change formatting, and put the DOCX back together?
We said: yes. That's actually how we build the comparisons in the first place.
And a funny thing happened. They didn't want the comparison engine. They wanted the layer underneath it — the ability to deconstruct documents into something that could feed into a large language model, have the model review and annotate, and then reconstruct it back into something lawyers could download and use.
We felt a tiny inkling that this was a bit like the Slack story. Slack set out to build a video game. To make themselves more efficient, they built an internal communication tool. The video game never shipped. The communication tool became one of the most widely used pieces of software in the world. Was this us? We set out to build tools for lawyers, and we accidentally built something that other legaltech companies had been looking for.
Three products, one foundation
Over the last six months, we've come to understand that what we built to power our own products can be isolated to power other people's products too. And that's led us to reorganise.
From one company — Syntheia — we are now three separate products:
Super Comparer handles document comparison. Many-to-many blacklines. Review changes across a document timeline. We will also be bringing our Smart Drafter Word add-in into this product line, because the add-in has always been focused on the same core questions: “have we agreed to this before?” and “What did we draft last time?” That's a comparison problem. It belongs here.
Fund Curator was built with law firms to serve private funds lawyers. It started with side letters and MFN elections — a fairly specific and genuinely tedious problem. We'll be announcing some collaborations shortly that bring the tool to more lawyers and address adjacent needs for private funds.
Syntheia — the mothership — is now the data layer. The infrastructure. The thing we spent years building that turned out to be valuable to more than just us. This infrastructure powers our own products, but now it is available to law firms and legaltech companies who need it to power theirs. Whether as the source of data to feed human experts or AI engines.
We will continue to release additional product lines as we build them with law firms and in-house teams. As they become mature, we will announce them.
So — what do we sell?
It all comes back to data. Garbage in, garbage out.
What we sell is certainty. The actual text, from the actual document, broken down in a way that's cross-referenced, linked, and retrievable at the level of detail lawyers actually need. Not a compression. Not an inference. The real thing.
That's always been the case. We have always sold certainty. We are now clearer about who needs it, and how we deliver it.
Syntheia.io · SuperComparer.com · FundCurator.com
We would love to hear your thoughts on this. Especially if you're a law firm thinking about what sits underneath the AI tools you're evaluating, or a legaltech company building something that needs a better data foundation, we are at hello@syntheia.io.
