The Composable Stack and Valuation Math

The architectural shift underway in IP technology is being discussed almost entirely as a product story. It is also a valuation story, and the valuation implications are not yet reflected in most investment theses in the sector.

For twenty years the dominant architecture has been the monolithic IP management system: one platform handling docketing, portfolio management, analytics, and renewals. The Model Context Protocol, an open standard now governed by the Linux Foundation, allows specialized tools to interoperate without a single platform sitting underneath them. Questel has confirmed MCP adoption. FluidityIQ has launched the first MCP-native patent search tool. The transition has moved from thesis to observable market behavior.

Three forces are now acting at the same time, and each affects a different part of the underwriting model.

The Bundle Premium Is Eroding

Platform multiples in IP technology have historically been underwritten by retention, and retention has been underwritten by switching cost. The effective switching cost for a monolithic IPMS runs at three to five times the annual license fee. A firm paying $300,000 per year faces $900,000 to $1,500,000 in configuration, training, data migration, and workflow rebuild before it can move. That figure is the real asset behind the retention line.

When a customer can replace one module inside the IPMS with a specialized alternative through MCP, and keep everything else connected, the cost being defended collapses from the platform to the module. A platform reporting 95% gross retention before MCP may report 85% to 90% after. The product did not get worse. The customers who stayed because leaving was expensive can now leave without a full migration.

The underwriting consequence is that gross retention has to be decomposed by cause rather than accepted as a single figure. Two companies reporting identical retention can carry entirely different exposure, depending on how much of that retention was earned through preference and how much was enforced through architecture. Before MCP the two were indistinguishable in the data, because no customer had a cheap way to reveal the difference. That is no longer the case.

There is a second-order effect on revenue quality. As the unit of purchase moves from the annual platform license toward the workflow step, pricing shifts from annual commitment toward usage. Patent search moving from an embedded IPMS module to an MCP-native tool delivers a 40% to 60% cost reduction for the buyer. That reduction comes out of somebody's ARR, and it comes out of the component of ARR that was least contested and therefore most profitable.

The Specialization Premium Is Emerging

The mirror image of the first force is that the penalty historically attached to point solutions is being removed. A best-of-breed prosecution tool no longer needs to build a full platform in order to compete for the prosecution workflow. It needs to be the best prosecution tool and to connect to everything else through the protocol. The result is a smaller company with a stronger competitive position on the dimension that determines whether it is selected.

This inverts a diligence heuristic that has been in use for a long time. Breadth of product surface was treated as a proxy for defensibility, on the reasonable assumption that a buyer replacing one capability had to replace all of them. Under composable architecture, breadth carries development and maintenance cost without conferring the protection it used to confer. An acquirer paying a premium for product surface is paying for an asset whose function has changed.

The specialization premium is not automatic, and it does not accrue to every narrow vendor. Riseon's April 2026 stack audit mapped 1,973 platforms and services across 24 categories, with AI-native entrants tripling in the drafting and search categories over fourteen months. Narrowness is a crowded position. The premium accrues to the subset of specialists whose depth rests on something an alternative cannot assemble quickly: proprietary data, or domain expertise embedded in the product rather than in the marketing. Where the depth rests on being early, the premium is temporary and should be underwritten as such.

The Cross-Sell Assumption Requires Stress-Testing

The roll-up thesis in IP technology follows a consistent structure: acquire adjacent capabilities, sell the combined platform into the combined customer base, expand net revenue retention without proportional acquisition cost. The structure depends on an assumption that is rarely stated explicitly, which is that the customer evaluates and buys at the level of the platform.

Composable architecture moves the evaluation unit down to the workflow step. A prosecution workflow decomposes into invention disclosure intake, prior art search, patentability assessment, claims drafting, specification drafting, filing, office action receipt and routing, response drafting, response review, response filing, and grant and maintenance. Each of those steps becomes a separately contestable purchase decision. The cross-sell then has to win on the merits of each step against the best available specialized alternative for that step, without the bundle carrying it.

The same discipline that enables step-level buying also exposes what the incumbent relationship previously concealed. Typical multi-vendor stacks carry 40% to 60% functional overlap, and across multiple audit exercises the actual technology spend has exceeded the estimated spend by 20% to 40%. A combined platform sold across a combined base is precisely the configuration in which overlap is highest and least visible. The buyer conducting a first structured audit is likely to find it, and the finding arrives during a renewal conversation.

For an investor, this does not invalidate the roll-up thesis. It changes how the synergy line should be built. Cross-sell modeled at the platform level and cross-sell modeled step by step produce materially different numbers, and only the second one survives contact with a buyer who has architectural visibility.

Where the Two Instruments Apply

Riseon Advisory maintains two instruments that bear directly on this question, and they answer different halves of it.

The Riseon Commercial Risk Index scores a company across ten commercial dimensions, each from 1 to 5, producing a composite between 10 and 50. It is designed to be applied during pre-investment diligence, during the first 100 days of ownership, and quarterly thereafter. Two independent assessors reviewing the same company produce scores within three points of each other. The framework structures judgment rather than replacing it.

Dimension 8, Competitive Position Durability, is the dimension on which the composable shift acts directly, and its scale was written for this transition. A score of 1 describes an advantage based on lock-in or incumbency, with no workflow-step differentiation and clear vulnerability to MCP-enabled substitution. A score of 3 describes an advantage based on product capability with moderate differentiation, some MCP compatibility, and a stable position. A score of 5 describes an advantage based on proprietary data or domain expertise that alternatives cannot replicate, with full MCP compatibility and a strengthening position. Dimension 10, Market Timing, sits immediately adjacent: an architecture built for a pre-MCP market with no composable roadmap scores 1, and a product native to MCP and AI scores 5.

A target can carry a respectable composite while scoring 1 on dimension 8. In an asset underwritten primarily on retention, that single dimension carries far more weight than its one-tenth share of the composite implies. It should be read separately from the composite and stated separately in the investment memorandum.

The IP Stack Maturity Model addresses the other half, which is timing. It describes five levels of customer-side architecture, from a single-vendor monolithic stack at Level 1, through bolt-on and API-connected configurations, to a composable stack at Level 5 running eight to fifteen specialized tools through an orchestration layer with no platform dominance. Most organizations sit at Level 2 today, running an IPMS alongside one to three point solutions with no integration architecture. No IP practice operates at Level 5.

Read together, the two instruments answer the question an investor actually needs answered. RCRI dimension 8 measures whether the target's position survives substitution. The maturity distribution of the target's customer base tells you when substitution becomes practically available to those customers. A vendor whose base is concentrated at Levels 1 and 2 has time. A vendor whose base is moving through Levels 3 and 4 is already being evaluated step by step, whether or not the vendor has noticed. The competitive window before the architectural gap becomes structural runs 12 to 18 months, which falls inside a typical hold period rather than beyond it.

What to Ask a Target

Seven questions, put plainly, will surface most of the exposure described above.

1. Decompose gross retention by cause. What proportion of the customer base has an economic reason to stay that is independent of switching cost? Which accounts would remain if replacing a single module cost nothing?

2. Which workflow steps does the company win on capability alone? For each one, what is the evidence: competitive win rate at that step, displacement of a named incumbent, retention measured on that module specifically rather than on the contract.

3. What is the MCP position in the shipped product, as distinct from the roadmap? Where does the product connect natively, where does it still require proprietary integration, and what exactly would a customer have to give up to replace one component.

4. Where does defensibility sit: proprietary data, domain expertise, or contractual and configuration inertia? If the honest answer is inertia, the company scores 1 on competitive position durability regardless of what the retention figure currently reports.

5. What is the architectural maturity of the customer base? What proportion of accounts have deliberately decoupled at least one workflow from their IPMS, and what proportion have run a structured stack audit in the last two years.

6. For a roll-up, at which specific workflow steps does the combined offering beat the best available specialized alternative, and what functional overlap exists across the combined portfolio at the step level.

7. What happens to reported ARR and to gross margin when the unit of purchase becomes the workflow step and pricing moves from annual commitment toward usage.

Pricing the Transition

The composable shift does not compress or expand valuations across IP technology uniformly. It redistributes them. Value moves from breadth toward depth, from enforced retention toward earned retention, and from platform-level relationships toward step-level performance. Some assets currently trading at a platform multiple hold a position that only supported that multiple while substitution was expensive. Some assets currently discounted for lack of product surface hold the exact position that the new architecture rewards.

The underlying diligence question is a familiar one about revenue durability. What has changed is that the mechanism through which durability will be tested is now identifiable, and the test arrives inside a normal hold period rather than after it. The instruments to score it exist. The discipline required is to apply the score before the market applies its own.

Sacha Lafaurie, Founder & CEO, Riseon Advisory