Today we're launching Sleuth Studio.

In the first piece in this series, we described a structural problem: the infrastructure powering competitive and strategic intelligence in the life sciences doesn't support the work it's supposed to enable. Databases offer structure without context. AI tools offer speed without accountability. Consulting engagements offer synthesis without continuity. And in-house teams stitch it all together manually, every time, building intelligence that never compounds.

We built Sleuth to close that gap. We've spent the past two years running it underneath our advisory engagements with pharma and biotech teams across BD, S&E, and competitive intelligence. Today we're opening it up. The same infrastructure that has powered those engagements is now available directly, as a product you can use yourself.

Here's what we built and why it works.

Sleuth's foundation: connected intelligence

There's a reasonable argument that the data layer in biopharma is heading toward commodity. With capable enough agents and enough compute, you can collect data from public sources, normalize it, and link it together. If assembling a snapshot of structured data is all a platform does, that advantage is narrowing.

Snapshots are useful. Sometimes you just need to know what's out there before going deeper. But there's a far more powerful system available when you move beyond snapshots to a living model of how the biopharma landscape connects.

Existing databases have relationships between data. You can search for a drug and see its modalities, indications, trials, and the companies behind them. That structure is useful for orientation. But there's an immense long tail to how deeply and broadly those relationships can be resolved, and the difference in what becomes possible at the far end is enormous.

Think about what the full picture involves: drugs connected to mechanisms, mechanisms connected to clinical programs across every phase and geography, programs connected to specific cohorts, endpoints, and readouts, companies connected to portfolios, partnerships, financing histories — and all of this connected to the unstructured evidence that gives it meaning: publications, patents, posters, filings, transcripts, and regulatory signals. The depth and completeness of that web determines whether you get a real answer or a partial one.

We take this seriously. Building and maintaining this connective tissue at scale — across tens of thousands of sources, including hard-to-reach markets like China, proprietary datasets, and clinical data structured all the way down to specific cohorts and endpoints — is the core of what Sleuth does. Every layer of additional resolution, every foreign-language filing translated and entity-matched, every deal term connected to the competitive context it implies, widens the gap between a useful snapshot and a system that can power the kinds of decisions our industry makes. That ongoing commitment is the foundation everything else depends on.

This same infrastructure is designed to integrate an organization's own data (company decks, deal memos, scientific evaluations) into the same connected layer. Most enterprises are sitting on years of accumulated proprietary intelligence that is functionally invisible: unsearchable, disconnected from the external landscape, and decaying with every team change. When that internal corpus lives inside the same entity-resolved graph as the external world, it stops being an archive and starts behaving like institutional memory that compounds rather than leaks.

An analytical layer that earns your trust

With this foundation in place, what the AI layer can do becomes categorically different.

Before Sleuth, every AI tool used in biopharma faced the same bottleneck: before it could reason about a question, it had to reconstruct the context. Assemble the entities, infer the relationships, figure out what's current. Most of the computation goes toward building a picture that may or may not be complete, and the user has no way to tell.

When the underlying data is already mapped, connected, and validated, that entire step disappears. The AI layer starts from a picture that's been maintained, not improvised, with the lineage to show you where an answer came from and where the evidence is thin. It can focus on what actually matters: interpreting the landscape, surfacing what's changed, and helping you get to a decision. That's what Sleuth Studio is built on.

Proven in biopharma engagements

This isn't a launch built on a demo. We've been using this infrastructure to power real engagements for nearly two years — competitive landscapes, mechanism deep dives, deal benchmarking, indication prioritization across dozens of indications and hundreds of assets.

In a traditional consulting engagement, most of the time and cost goes toward assembly — collecting, reconciling, structuring — before anyone begins the actual analysis. With Sleuth, that work is already done. The data is structured, connected, and current. The domain judgment, the interpretation under ambiguity, the willingness to take a position on an unclear readout — that's still there when you need it, through our Concierge team.

And because everything happens in the same system, every analysis compounds. When you reopen a landscape you built two months ago, it's already been updated with everything that's changed since. The landscape one team member ran six months ago doesn't disappear into a slide deck — it persists in the system, searchable and connected, so the next person asking a related question starts from that foundation, rather than from scratch. The interpretation an expert encodes on an ambiguous readout becomes part of the organization's working knowledge. That is the difference between a tool that produces outputs and a system that accumulates intelligence.

Introducing Sleuth Studio

Sleuth Studio gives you direct access to the same connected intelligence infrastructure.

Start with a structured, entity-resolved picture of a space rather than a blank search bar. The data is already connected, validated, and sourced — so you're analyzing from the first minute, not rebuilding the picture from scratch. Ask questions and get answers grounded in real data, with the lineage visible so you can see exactly why the system believes what it believes. When something is wrong or uncertain, the system tells you, rather than generating a confident answer and hoping you catch the error.

Cut data across programs, mechanisms, endpoints, companies, and deal terms in a single environment. Build competitive landscapes. Benchmark clinical data. Track how a landscape evolves over time instead of rebuilding it every quarter.

The sprint that used to take two weeks of stitching together databases and documents can start from a baseline that's already been maintained. The marathon — keeping the competitive picture current as the world moves — happens in the background instead of in your spare time.

And when the question is ambiguous enough that you need a human to take a position, our Concierge team works inside the same system. Every interaction builds on what came before. The database and the consultant are no longer separate tools that don't share a language. They're the same thing.

We built this because we were the people living with that constraint — triangulating between sources that didn't talk to each other, rebuilding context that should have carried forward, watching good intelligence decay in slide decks.

We'll be sharing more in the coming days about what's under the hood — the knowledge infrastructure, the design principles, and how we got here.