We shipped two significant updates in January 2026: an expansion to 34 detected clause types and a new deal room view that surfaces cross-document clause patterns. This post explains what changed and why we built it this way.
Product update posts are something we want to write honestly, which means being specific about what we built and what we are still working on, not just announcing features in the most favorable light. Here is what happened in Q1 2026 and what it means for how legal teams use Clausebeam.
Clause-type expansion: from 22 to 34 types
When we launched, Clausebeam detected 22 clause types focused on the core risk categories in SaaS agreements and NDAs: indemnification, liability caps, termination rights, confidentiality, representations and warranties, governing law, IP assignment, and a few others. That set covered the highest-frequency, highest-stakes clauses in the contract types most of our early users were reviewing.
The 12 new types we added in January are primarily drawn from feedback about where the gaps were most painful:
M&A schedule coverage: We added Material Adverse Change (MAC) clause detection, closing conditions, and representations survival period analysis. These are critical in M&A ancillary agreements and due diligence work, and they were the most-requested expansion from law firms using Clausebeam for transaction support. MAC clause language has become particularly important since the litigation arising from pandemic-era deal disputes that tested what "material adverse effect" actually means under Delaware law.
Dispute resolution depth: We expanded from a single "dispute resolution" type to separate detection of arbitration clauses, mandatory venue requirements, jury waiver provisions, and class action waivers. These are functionally distinct and carry different risk implications, and collapsing them into one type was making the output less useful for practitioners who needed to know specifically which element was present.
Data processing and privacy: We added detection for data processing agreement requirements, cross-border transfer mechanisms, and data subject rights obligations. As more commercial contracts involve personal data processing, these provisions have become routine negotiation items, and our NDA and SaaS contract coverage was incomplete without them.
IP licensing terms: Sublicense rights, exclusivity scope, field-of-use restrictions, and audit rights in licensing agreements. IP licensing was the contract type where our users most consistently reported gaps between what they needed to flag and what Clausebeam was catching.
How we validated the new types
Adding clause types is not just a matter of expanding a detection list. Each new type requires building out what "standard" looks like for that type, what deviations warrant flagging, and what the deviation flags should say. For the MAC clause expansion, we reviewed a meaningful sample of M&A agreements and read through the relevant Delaware case law on MAC interpretation to calibrate our deviation flags against what actually matters in practice.
We are not claiming the new types are as well-calibrated as indemnification or liability caps, which have had months of review data behind them. The MAC clause detection, in particular, is new territory for us and we expect to continue improving its precision over the next quarter. We flagged this internally and we are flagging it here: if you are using Clausebeam for MAC clause analysis on a time-sensitive transaction, treat the output as a first-pass that warrants careful attorney review rather than a definitive analysis.
Deal room view: what it does and why it matters
The deal room view is a different kind of update. It does not add new clause detection; it adds a new way of looking at what is already being detected across multiple documents at once.
The core problem it solves: when you upload 30 contracts from a deal room and Clausebeam analyzes each one, you previously had to read 30 individual clause flag reports to understand the pattern across the set. If 19 of the 30 contracts had indemnification clauses with the same asymmetric structure, that pattern was visible only if you went looking for it across all 30 reports.
The deal room view surfaces cross-document patterns automatically. It answers questions like: "Across the 28 vendor MSAs in this set, how are liability caps distributed? Are most at 1x, or is there significant variation? Which specific agreements have the most aggressive cap language?" or "Does indemnification asymmetry appear uniformly across the vendor stack, or is it concentrated in specific counterparties?"
For M&A due diligence, the deal room view changes how you triage a target company's contract portfolio. Instead of reading through every contract to identify the high-risk items, you start with the cross-document pattern view to understand where the systemic issues are, then read the individual reports for the agreements that show up as outliers. That is a fundamentally different workflow from reading document by document.
Current limits of the deal room view
The deal room view works best when the contracts in a set are of the same or similar type. Mixing an M&A agreement with a set of NDAs and SaaS MSAs produces a cross-document view that is harder to interpret because the baseline for "normal" differs across contract types. We have built in contract-type grouping to address this, but the grouping relies on our contract-type detection, which is more reliable for common types than for unusual or hybrid agreements.
The deal room view also has a practical size limit at this stage: it performs best up to around 50 documents in a set. Above that, the cross-document pattern surfacing can become noisy as the number of detected deviations grows. We are working on prioritization and filtering for larger deal rooms, but we want to be honest about where the current version works best rather than overclaiming the capability.
What comes next
Two things we are actively building: better handling of redlined documents and clause change tracking across drafts, and expanded M&A schedule coverage including representations qualification schedules. Both are squarely in the lane of where legal teams using Clausebeam for transactional work need more from us.
We also want to improve the attorney annotation layer, specifically the ability to mark a flag as reviewed and record a disposition note (accepted as-is, marked up, escalated) so that the Clausebeam output can travel with the review process rather than existing separately from it. That is the integration gap that practitioners tell us is the most significant friction point in using AI clause analysis in a real workflow, and it is where we are putting significant development effort in Q2.
If you are using Clausebeam and have specific feedback on either of these areas, or on the new clause types, we want to hear it. We are a small team and every piece of practitioner feedback goes directly into what we build next.
