The AI platform for global finance
Meridian reads your dataroom, builds the model and writes live formulas straight into Excel — with a citation for every number. An in-depth, working teardown of the category leader built as a real product, not a mockup.
3.4M
Analyst hours saved / week
+32% YoY
128,400
Models built on platform
+46% YoY
19.2M
Documents parsed
+28% YoY
1.9s
Median task latency
-14% YoY
Helios Grid Systems
$118.23
FY30E revenue
$7.6B
+9.5% CAGR
FY30E EBITDA
$2.2B
28.4% margin
Net debt / EBITDA
2.1x
down from 2.4x
Agent activity
- Parsed 6 filings · 812 pages
- Reconciled FY24 EBITDA → $1,106M
- Built 214 linked formulas
- Flagged NWC peg 40bps above peer set
Guardrails
Deployed by 40,000+ analysts, bankers and investors
An in-depth review of rogo.com
Before writing a line of code we reverse-engineered the category leader: what it does, why it wins, and where the product surface still has gaps. Everything below informs the build that follows — and where a third product could differentiate.
| Capability | Reference platform | Impact |
|---|---|---|
| Prompt-to-model three-statement build | Native, agentic | High |
| Excel add-in with live formula write-back | Yes — core differentiator | High |
| Dataroom Q&A with page-level citations | Yes, source-linked | High |
| Financial data connectors | Capital IQ, FactSet, PitchBook, Bloomberg | Medium |
| PowerPoint / deck generation | House-style templates | Medium |
| Enterprise security posture | SOC 2 II, no training on customer data | Foundational |
| Workflow automation / scheduled agents | Emerging | Medium |
| Transparent per-seat pricing | Sales-led | Medium |
Strongest in document-grounded reasoning and spreadsheet fidelity. Weakest where procurement and conservative IT teams push back: opaque packaging and no offline mode.
Wedge: trust through citations
Every generated number links back to a page in the source document. For regulated finance workflows this is the difference between a demo and a deployment — it is the single most defensible part of the product.
Wedge: meeting analysts inside Excel
Rather than asking finance teams to abandon spreadsheets, the add-in writes live formulas into the workbooks they already own. Adoption friction collapses because the artifact of record never changes.
Gap: scenario governance at portfolio scale
Single-model workflows are excellent, but cross-portfolio consistency — one assumption set driving 60 models — needs a shared assumption registry with approval state.
Gap: offline and air-gapped use
Deals close on planes. A fully local inference tier with deferred sync would meaningfully widen the addressable market for conservative institutions.
Gap: transparent packaging
Pricing is sales-led with no public tiers. A self-serve professional tier would convert the long tail of boutique funds that will never book an enterprise pilot.
Agentic modelling
Describe the business and the agent builds a fully linked three-statement model with a defensible driver set — then explains every assumption it made.
Dataroom intelligence
Upload or connect a dataroom and interrogate it in natural language. Every answer ships with a page-level citation you can click straight into.
Excel & PowerPoint native
Work where the deal lives. The add-in writes live formulas into your existing workbooks and generates committee-ready pages that stay linked to the source.
Market data fabric
One normalised schema across filings, market data and alternative datasets, with lineage on every field so compliance can reconstruct any number.
Workflow agents
Schedule recurring work: earnings updates, covenant monitoring, portfolio marks and IC memo refreshes run on their own and post to your channels.
Enterprise control plane
SSO, granular entitlements, regional data residency and a zero-retention model posture. Your data is never used to train shared models.
Six surfaces that replace the manual analyst loop
Each capability is a real, working surface in this build — not a marketing tile. Explore the dashboard, interrogate the dataroom, then build and export a model.
- Driver-based build from a single prompt
- Formula-level traceability back to source
- Auto-generated sensitivity and scenario grids
- Cross-document entity resolution
- Table extraction from scanned exhibits
- Red-flag detection against a deal checklist
- Live formulas, not hard-coded values
- Deck generation with house style templates
- Round-trip: model → deck → model
- Capital IQ, FactSet, Bloomberg, PitchBook
- Segment-level fundamentals normalisation
- Full field-level lineage and audit trail
- Cron, event and webhook triggers
- Approval gates with full version history
- Slack, Teams and email delivery
- SOC 2 Type II, ISO 27001, GDPR
- VPC / private cloud deployment
- Model-level access logs and red-teaming
Connect the context
Point the platform at a dataroom, a data warehouse, or a market data subscription. Ingestion normalises every page, table and footnote into a queryable graph.
Ask in plain language
“Build the model”, “what changed this quarter”, “where is the risk”. The agent plans its own steps, calls the right tools and shows its work as it goes.
Review with citations
Each output carries a citation chip back to the exact page and a formula trace back to the driver. Challenge anything and the agent re-derives it in front of you.
Ship to Excel or PowerPoint
Export a live, formula-driven workbook or a committee-ready deck. Everything round-trips so the artifact of record stays current.
Your desk, instrumented
A live command centre across deals, models and market data — with the audit trail a risk committee will actually accept.
Assets under coverage
$412.8B
+8.4% QoQ
Live feeds from 14 custodians
Active deal processes
86
+11 this month
32 sell-side · 54 buy-side
Models under management
2,914
+164 MoM
Version-controlled, audited
Agent citation accuracy
99.2%
+0.6 pts
Audited on 12,400 gold tasks
| Holding | Invested | Fair value | MOIC | IRR | Mark |
|---|---|---|---|---|---|
NC Northwind Cloud Vintage 2021 | $145M | $412M | 2.84x | 41.2% | Up |
CI Cascade Industrial Vintage 2020 | $320M | $508M | 1.59x | 17.4% | Up |
IP Ironwood Packaging Vintage 2022 | $210M | $236M | 1.12x | 5.8% | Flat |
AP Atlas Payments Vintage 2021 | $188M | $512M | 2.72x | 36.6% | Up |
VS Vertex Semiconductor Vintage 2019 | $402M | $621M | 1.54x | 14.8% | Up |
HP Harbor Point Logistics Vintage 2023 | $276M | $301M | 1.09x | 4.1% | Flat |
BG Bluepeak Growth III Vintage 2022 | $96M | $74M | 0.77x | -12.8% | Down |
MC Meridian Consumer Vintage 2020 | $354M | $468M | 1.32x | 11.2% | Up |
09:42:18 · Agent · model-builder
Wrote 214 formulas to Helios_DCF_v7.xlsx
09:41:55 · A. Mehta
Approved WACC assumption change 9.4% → 9.2%
09:38:02 · Agent · dataroom
Cited 3 sources for covenant headroom answer
09:31:44 · System · guardrail
Blocked external egress of model payload
09:22:10 · L. Bergström
Flagged change-of-control clause for legal review
09:12:37 · Agent · credit
Refreshed covenant sweep for Ironwood Packaging
Build it here. Download it as Excel.
A fully linked DCF and three-statement engine running live in your browser. Tune the drivers, stress the scenarios, then export a genuine .xlsx workbook with formulas — not a frozen CSV.
Implied share price
$118.23
Market $78.40
Implied upside
50.8%
Undervalued
Enterprise value
$17.67B
8.2x FY5 EBITDA
Equity value
$14.42B
10.0x FY5 P/E
FY5 revenue
$7.58B
9.5% CAGR
TV % of EV
76.0%
Terminal growth 2.8%
| Income statement ($M) | FY2026E | FY2027E | FY2028E | FY2029E | FY2030E |
|---|---|---|---|---|---|
| Revenue | 5,398 | 5,981 | 6,550 | 7,087 | 7,583 |
| Gross profit | 3,369 | 3,732 | 4,087 | 4,422 | 4,732 |
| Gross margin | 62.4% | 62.4% | 62.4% | 62.4% | 62.4% |
| EBIT | 1,306 | 1,448 | 1,585 | 1,715 | 1,835 |
| EBIT margin | 24.2% | 24.2% | 24.2% | 24.2% | 24.2% |
| EBITDA | 1,533 | 1,699 | 1,860 | 2,013 | 2,154 |
| EBITDA margin | 28.4% | 28.4% | 28.4% | 28.4% | 28.4% |
| Net income | 1,032 | 1,144 | 1,252 | 1,355 | 1,450 |
2,400+ models, one click to load
Every template is a full driver set, not a screenshot. Load one into the builder above and it becomes yours — then export it straight to Excel.
- Monthly / quarterly toggle
- Revolver plug
- Dividend & buyback schedule
- NRR ramp 114% → 121%
- Gross margin 78% → 84%
- Terminal FCF margin 31%
- 18 pre-loaded peers
- Quartile statistics
- Implied value waterfall
- 5.1x leverage at close
- 2.8x MOIC base
- 24.6% gross IRR
- Synergy phasing 40/70/100%
- Mix of cash / stock / debt
- Accretion bridge
- Construction draw schedule
- Lease-up absorption
- LP / GP 8% pref
- Net leverage bridge
- Covenant headroom
- Downgrade stress grid
- Vintage pacing engine
- J-curve & DPI / TVPI
- Net-of-fee IRR
Ask your documents. Get citations, not vibes.
Every answer is grounded in a page-level citation you can open. Try a prompt below — the assistant is simulated client-side but the reasoning surface, tables and citation trail are all real UI.
Documents indexed
8
7 fully processed
Pages parsed
914
incl. scanned exhibits
Tables extracted
41
94% auto-validated
Median answer time
2.4s
p50, 812-page corpus
Summarise the key diligence risks in Project Cascade and quantify the covenant headroom under a 200bps EBITDA shock.
09:41
I reviewed 184 pages of the CIM, the executed credit agreement and the quality-of-earnings report. Three risks clear my materiality threshold, and covenant headroom survives the shock — but only via the portability carve-out.
Agent trace
- Parsed 6 documents · 812 pages, 41 tables extracted
- Reconciled FY24 EBITDA across CIM, QofE and audited statements
- Modelled a −200bps EBITDA shock against the leverage covenant
- Cross-checked restricted-payment basket and change-of-control language
| Covenant | Level | Headroom (base) | Headroom (−200bps) |
|---|---|---|---|
| Total leverage | ≤ 7.25x | 0.85x | 0.09x |
| Interest coverage | ≥ 2.00x | 0.94x | 0.61x |
| Fixed charge coverage | ≥ 1.30x | 0.42x | 0.26x |
3 cited sources
Answers are generated from indexed sources only. Zero retention · no training on your data.
Work that runs while you sleep
Schedule the recurring parts of the job — earnings updates, covenant sweeps, portfolio marks — and keep a full trace of what the agent did and why.
| Run | Status | Duration |
|---|---|---|
Q4 earnings update — Helios Grid Systems RUN-4821Equity researchModel update | Running | 4m 12s |
Covenant headroom sweep — Ironwood Packaging RUN-4818CreditMonitoring | Completed | 2m 48s |
Dataroom red-flag review — Project Cascade RUN-4815DiligenceLegal | Needs review | 11m 03s |
Portfolio mark-to-market — Growth Fund III RUN-4811FundValuation | Completed | 6m 31s |
Comps refresh — Semis basket (18 peers) RUN-4809CompsMarket data | Queued | — |
IC memo first draft — Project Atlas RUN-4804Investment committeeDrafting | Completed | 3m 57s |
S&P Capital IQ
Market data
FactSet
Market data
Bloomberg
Market data
PitchBook
Private markets
Microsoft Excel
Productivity
Microsoft PowerPoint
Productivity
Snowflake
Data warehouse
Databricks
Data warehouse
NetSuite
ERP
SAP S/4HANA
ERP
Slack
Collaboration
Microsoft Teams
Collaboration
Intralinks
Dataroom
Datasite
Dataroom
Built for the risk committee
Finance data is the most sensitive data in the building. Meridian is engineered so that security review is a formality, not a six-month project.
SOC 2 Type II
Annual audit, continuous control monitoring, evidence on request
ISO 27001 & 27701
Information security and privacy management certified
Zero training on your data
Contractual and technical guarantee — customers never enter a shared training set
Regional data residency
US, EU, UK, UAE, Singapore and Australia processing regions
Bring your own key
Customer-managed keys via AWS KMS, Azure Key Vault or GCP KMS
Private deployment
Single-tenant VPC inside your cloud account, no egress of raw data
Multi-tenant SaaS
Logical isolation, regional processing, 99.9% SLA
Single-tenant VPC
Dedicated compute inside our cloud, private networking
In your cloud
Deployed into your AWS / Azure / GCP account, no egress
United States
us-east-1 / us-west-2
European Union
eu-central-1
United Kingdom
eu-west-2
UAE
me-central-1
Singapore
ap-southeast-1
Australia
ap-southeast-2
The same work, 98% faster
Benchmarked against a 12-person deal team across a full quarter. Every number below is hours of analyst time per cycle.
Manual baseline
81h
per reporting cycle
With Meridian
1.9h
97.7% reduction
Annualised capacity returned
3,797 hours
≈ 2.1 full-time analysts
Priced per seat, published openly
One of the gaps the teardown flagged in the category is opaque packaging. Here it is on the page, with a live seat calculator.
- 150 agent runs per month
- 10 model builds with Excel export
- Public filings + market data
- 5 GB dataroom storage
- Community support
- Unlimited agent runs
- Unlimited model builds + Excel / PPT export
- Capital IQ, FactSet and PitchBook connectors
- Excel add-in with formula write-back
- 200 GB dataroom storage
- Priority support, 4-hour SLA
- Everything in Professional
- SSO / SCIM, granular entitlements
- Private VPC or on-prem inference
- Customer-managed encryption keys
- Unlimited dataroom + audit retention
- Named solutions architect, 99.9% SLA
14-day trial, no card
Full Professional feature set including Excel export.
Free for students
Verified academic accounts get Analyst tier at no cost.
Volume pricing at 100+
Blended seat rates and pooled model allowances.
What finance teams say after 90 days
A first-draft model that used to take an analyst two days now lands in under ten minutes — and it can show me the source for every single assumption.
Priya Raghunathan
Head of Investment Research, Kestrel Capital Partners
We ran 41 datarooms through it during diligence season. The citation trail is what got our risk committee comfortable, not the speed.
Daniel Okonjo
Managing Director, Private Equity, Northgate Equity
The Excel add-in writes live formulas into our house template. That single detail is why adoption went from pilot to 400 seats in a quarter.
Elena Vasquez
COO, Corporate Development, Halcyon Industries
Questions we get in every security review
Still stuck? The sales engineering team answers technical questions directly, with no gatekeeping.
Bring your next model. Leave with an Excel file you actually own.
Start free, connect a dataroom, and build your first three-statement model in under ten minutes.
No credit card · SOC 2 Type II · Your data is never used for training