The MNPI compliance stack splits in two directions at once

Expert-call compliance used to be a one-vendor problem: pick GLG, Guidepoint, or Third Bridge, trust their compliance team, move on. That model is breaking. Three forces are pulling the stack apart, and the resulting map has at least four distinct layers, each with its own incumbent and its own gap.

The first force is legal. An Illinois federal court allowed a Walgreens investor lawsuit to proceed in 2026 partly on the strength of a recorded conference statement, a signal that recorded management and expert commentary is now fair game in 10b-5 actions. Recall the SEC v. Mathew Martoma prosecution: the consult was the case. The Walgreens decision extends that logic to the recording itself, not just the call.

The second force is vendor proliferation. General-purpose meeting recorders (Tactiq, Fathom, Read.ai, Otter, Bluedot, Granola) have spent the last two years bolting on enterprise compliance features and selling into research and IR teams that previously got their transcripts from an expert network or did without. Most of these tools were built for sales calls. The compliance posture is being added after the fact.

The third force is the expert networks pushing back. GLG announced ISO 27001 certification to anchor a security narrative. Guidepoint leans on the compliance scaffolding around its roughly 100,000-transcript library. AlphaSense has built MNPI-redaction tooling around the Tegus catalog it acquired. Each is staking out a different layer of the stack.

The four layers of the stack

Pre-call screening. Who the expert is, what they're cleared to discuss, and what topics are off-limits. This is the layer expert networks have always owned, and it's the hardest for general-purpose recorders to replicate. A Tactiq or a Fathom records whatever the user puts in front of it. They have no view on whether the person on the other end of the call is a current employee of a public company in a quiet period.

In-call recording and transcription. This is where the proliferation is loudest. The general-purpose tools are good enough that research teams use them for internal calls without thinking. The expert networks bundle recording into the call platform itself. The buyer choice here is increasingly about format and latency, not capability.

Post-call transcript review. The expert networks run compliance teams that scrub transcripts before delivery. AlphaSense has automated parts of this for the Tegus catalog with MNPI-redaction features that flag and mask sensitive language. The general-purpose recorders mostly do not do this. A research analyst recording an industry expert on Fathom gets the raw transcript and the raw legal exposure.

Buy-side internal review. The layer with the fewest vendors and the most risk. Once a transcript is in a hedge fund's research system, who decides what stays in the IC memo and what gets stripped? This is currently a process problem, not a product problem, and it is the layer most exposed by the Walgreens ruling.

Where the vendors actually compete

The expert networks are not really competing with Tactiq and Fathom on recording. They are competing on the screening and review layers, which the general-purpose tools cannot easily replicate without a compliance team and an expert-vetting process. The ISO 27001 certifications and the redaction features are a way of saying: we own more of the stack than the meeting recorder does.

The meeting recorders, in turn, are not really competing with the networks on expert sourcing. They are competing for the recording slot inside calls the buy-side is already taking, often calls the network never knew about: industry contacts, ex-colleagues, advisors. The category the recorders are eating is not "expert network"; it is "untranscribed conversation," which is a much larger pool.

The AlphaSense move, with Tegus underneath it, is the most direct attempt to own multiple layers at once: a curated transcript catalog (sourcing), a recording and transcription pipeline (capture), and automated redaction (review). The bet is that integration across layers beats best-of-breed at any single layer.

What to watch next: how the Walgreens docket develops on appeal, whether other 10b-5 plaintiffs cite recorded expert commentary as a basis for pleading scienter, and whether the SEC weighs in on transcript retention standards for registered investment advisers. The legal layer moves slower than the vendor layer, but it is the legal layer that sets the ceiling on what the vendors can sell.