Over the past year we have spoken with more than a hundred investment teams about how they build conviction, and where screening and diligence tend to break down. Three themes came up again and again. Teams need to trust the numbers before they can trust the narrative. They need the analysis to reflect their own strategy rather than a generic template. And they want to stay firmly in control of the judgment that defines their craft, rather than hand it to a black box.

This release is our answer to all three. Some of what follows is visible the moment you log in; much of it lives deep in the platform, in how Clarenza reads, calculates, reconciles and challenges. Below are the ten most significant upgrades, ordered by how much we expect them to change the way your team works.

1. Accuracy you can reproduce: deterministic analysis at the core

Every number in a memo should be traceable, and every conclusion should be repeatable. Language models are excellent at reading and reasoning, but they are not calculators, and left alone they can return a slightly different answer to the same question on a different day. That is not acceptable in a document that goes in front of an investment committee.

We have moved the parts of diligence that should never vary (calculations, metric derivation, ratio analysis, thesis checks) onto deterministic solvers, while our agents focus on the work that genuinely requires reading and interpretation. Underneath sits a new Clarenza ontology that organizes knowledge in three layers: Clarenza Core, your fund thesis, and deal-specific overrides. It keeps claims separate from metrics, so a founder's assertion is never silently treated as a verified figure.

The practical result is simple. The same data room, under the same thesis, produces the same analysis. And when something does look wrong, the error is reproducible, which means it can be found, explained and fixed rather than shrugged off as model variance. That is the difference between a tool you double-check and a tool you can rely on.

2. Financial metrics that reconcile across the entire data room

Real data rooms disagree with themselves. Revenue in the management presentation differs from the audited accounts; the cohort file defines churn differently from the board pack; a quarter quietly gets restated between two versions of the model. Finding these inconsistencies is some of the most valuable work in diligence, and some of the most tedious.

Clarenza now extracts financial KPIs across the whole VDR, normalizes them to consistent definitions and periods, and reconciles them against one another. Where sources conflict, you see the conflict, with every version linked back to its document. Discrepancies, anomalies and figures that move without explanation are flagged rather than averaged away.

Instead of an analyst spending days building a reconciliation spreadsheet by hand, the team starts with a clear map of where the numbers agree, where they diverge, and which questions belong on the list for management. The decision about which number to believe stays with you. Clarenza simply makes sure you always know there was a choice to make.

3. Thesis configuration with gating and scoring

Your strategy is what distinguishes your fund, and your analysis should reflect it. Thesis configuration already let teams shape memos to their strategy through templates, refinement and versioning. This release takes it considerably further.

Gating lets you define the hard criteria a deal must meet (sector, stage, geography, ticket size, financial thresholds, or any condition your mandate imposes), so opportunities outside the mandate are identified early and transparently, before anyone has invested hours in them. Scoring lets you weight the dimensions that matter most to your fund and see how each opportunity measures up, with every score traced back to the evidence behind it.

Because thesis versions are tracked, you can follow how your criteria evolve and understand how a deal would have looked under an earlier version of your strategy. This is configurability as a principle: two funds looking at the same company should reach their own conclusions, not a homogenized one.

4. A sharper, more targeted Red Team

Conviction that has never been challenged is fragile. The Red Team runs an adversarial review of every deal, building the disconfirming case your memo might underweight, and in this release it has become both more rigorous and more useful.

Challenges are now anchored in your thesis and in the specific evidence in the data room, rather than drawn from generic risk lists. Kill conditions are explicit and tied to their sources. And the review separates risks that are genuinely disqualifying from those that need mitigation and those that simply need a question answered, so the team knows where to focus.

The effect is that committees receive a structured bear case before the meeting rather than discovering it during one. It also gives junior team members a model of how experienced partners stress-test a deal, which helps them grow as investors rather than simply consume conclusions.

5. Automated DDQ, from questionnaire to answered draft

Due diligence questionnaires remain one of the heaviest and most repetitive parts of any process. Clarenza now imports your existing DDQs and answers them from the evidence in the data room and your private memory, citing the source for every answer and clearly marking where information is missing, incomplete or ambiguous.

It goes beyond the template, too. Based on the gaps, red flags and reconciliation issues it has surfaced, Clarenza suggests deal-specific questions, so your follow-up list to management is shaped by what this particular deal needs rather than by last year's checklist. Your team reviews, edits and approves; the mechanical work of locating answers across hundreds of documents is done before they start.

6. An evidence explorer and document previewer built for verification

Every claim in a Clarenza memo links to its source, and the redesigned evidence explorer makes following that trail effortless. Open any claim or metric and the document previewer takes you to the exact passage, table or figure, highlighted with bounding boxes on the original page.

From there you can move between related pieces of evidence, compare how a figure appears across different documents, and inspect the full reasoning chain the agents followed to reach a conclusion. Verification turns from a search exercise into a single click, which matters most at exactly the moments that count: in IC preparation, and when a partner asks where a number came from.

7. Deeper ingestion and new connectors for documents and meeting notes

The quality of any analysis depends on what goes into it. Document ingestion and extraction are now substantially more powerful, handling complex layouts, dense financial tables, scanned pages and long-form reports with greater fidelity, and extracting claims and metrics with cleaner structure.

Alongside our existing integrations for market data enrichment and CRM, new connectors for Notion and Granola bring your internal research and meeting notes into the platform. Founder calls, expert interviews and internal discussions become part of the deal's evidence base instead of living in separate tools. Everything flows into your tenant-isolated private memory, where it compounds with every deal and remains available to the whole team.

8. Collaboration and sharing for the whole deal team

Conviction is rarely built alone. Teams can now comment directly on memos, claims, and evidence, mention colleagues to bring them into a specific question, and tag deals, documents, and findings to keep work organized across the Deal Library. Discussions stay attached to the evidence they concern, so the reasoning behind a decision is preserved long after the deal closes.

When the work needs to travel beyond the platform, improved export formats carry your memo into the formats your committee and stakeholders already use, with structure and source references intact. New sharing options let you put specific artifacts in front of the right people, at the right level of access, without exporting the whole deal.

9. Stronger compliance, security and workspace controls

Clarenza was built for confidential deal data from day one: EU-sovereign models only, confidential compute, zero data retention, no training on client data, compliance with the EU AI Act and readiness for DORA. This release puts more of that control directly in the hands of your administrators.

Compliance and security controls are more granular, and workspace settings now offer greater configurability over the elements firms most often want to adapt, from defaults and templates to access and how the platform behaves for each team. For firms that have held back from AI because confidential deal data cannot go into consumer tools or US cloud providers, these controls make the conversation with compliance a much shorter one.

10. A redesigned interface around the way deal teams work

Finally, Clarenza looks and feels different. The UI overhaul reorganizes the product around the natural flow of a deal, from intake and screening through diligence to decision, so the next step is always clear and the evidence is never more than a click away.

Navigation is simpler, dense information is easier to scan, and the memo, its evidence and the team's discussion now sit side by side. Most of the upgrades above live deep in the platform. This is the one you will notice in your first minute.

What ties it all together

Each of these upgrades serves the same purpose: moving programmable machine work off your team so analysts and partners can spend their time on judgment. Clarenza does the reading, reconciling, cross-checking and drafting. Your team decides what it all means. Every edit, correction and override they make is stored in your firm's private memory, so the platform becomes more attuned to how you invest with every deal, while the reasoning behind each conclusion stays visible and open to challenge.

We believe this is how AI should enter private capital: not as a replacement for investment expertise, but as the infrastructure that lets that expertise reach conviction faster, and with more confidence.

Try it on your own deals

The best way to judge any of this is on a live deal of your own. Our beta program is open to European private equity, venture capital, growth equity and family office teams, and we would be glad to have you join. Test Clarenza on your deals for free for 30 days, and tell us what you think.

‍