A financial data foundation for Claude that your whole company can build on.
Tick & Tie connects Claude to your financial data, defines your key metrics once in a custom semantic layer, and certifies the results against your official reports after every month-end close. Three weeks for most stacks, at an affordable fixed fee.
We work with finance teams already on Claude Team or Claude Enterprise. Everything we build lives in your own Claude tenant.
From first call to certified numbers.
Intro call (30 minutes).
We map where your key metrics live, determine which approach fits your stack, and provide an affordable fixed-fee quote. You decide from there.
The build (three weeks for most stacks).
Connect, Define, Certify, inside your own Claude tenant. You keep everything we build.
Keep it right (optional).
Monthly recertification after every month-end close and every new Claude version, plus add-ons when you want them. See what comes next →
Connect.
Connecting Claude directly to your ERP is a good first step, but it won't show Claude all of the numbers you actually run the business on. Adjusted EBITDA, ARR and management P&Ls depend on adjustments and mappings that live in a warehouse, a planning tool or an analyst's spreadsheet. So we connect Claude to every source that holds a key number, and every user sees only what their permissions allow.
Define.
We build the semantic layer: the set of agreed metric definitions that sits between your raw data and everything built on it. Adjusted EBITDA, ARR, management and business-unit P&Ls each get one definition, written in a form Claude reads, with instructions for when to pull a metric and when to ask which one you mean. Your controller and FP&A lead own the definitions; we write them down and wire them in.
Certify.
Our evaluation suite asks Claude a bank of finance questions with known answers and checks every answer against your official reports. It runs before handover, then again after every month-end close and every new version of Claude, so you know the numbers still tie out.
Three weeks, week by week
| Week | What happens | You see |
|---|---|---|
| 1 | Access and connection. Definitions workshop with your controller and FP&A lead. | Every key metric listed, with its owner and its agreed definition |
| 2 | We build the semantic layer and connect Claude to it. | Your team asking Claude for adjusted EBITDA and getting your number |
| 3 | The evaluation suite runs against your official reports; we fix every miss and hand over. | A certified pass rate, and a foundation your teams can build on |
Larger stacks take longer: several entities, data that first needs landing in a warehouse, or ARR that must reconcile to the GL. The intro call sets the timeline and the fixed fee before anything starts.
We build on what you already have.
| Where your numbers live today | How we build the semantic layer |
|---|---|
| A warehouse (Databricks, Snowflake, BigQuery) | Metric tables or views in the warehouse; Claude queries them |
| Your ERP only | Metric definitions applied to data pulled from the ERP through a secure connector |
| A BI model (Power BI, Looker, Tableau) | We govern the measures you already have and expose them to Claude, rather than rebuild |
| An existing semantic layer (dbt, Looker, warehouse metric views) | We check its definitions against finance's, connect Claude and certify |
| A planning tool (Anaplan, Adaptive, Pigment, Planful) | We align the semantic layer with your FP&A model, so both give the same answer |
| Spreadsheets | We move the spreadsheet logic into a small governed store with written definitions |
Where it helps, we also land your data into a warehouse first. Vendor AI layers such as Snowflake Cortex Analyst, Databricks Genie and Fabric data agents work alongside Claude; we don't replace them.
Connects with the systems your finance team already runs on.
ERP and accounting
- NetSuite
- Sage Intacct
- Dynamics 365
- Workday
- SAP S/4HANA
- Oracle Fusion
- QuickBooks
- Acumatica
- Infor
- Epicor
Data warehouse
- Snowflake
- Databricks
- BigQuery
- Redshift
- Azure Synapse
BI and analytics
- Power BI
- Tableau
- Looker
- Qlik
- Sigma
Planning
- Anaplan
- Adaptive
- Pigment
- Planful
CRM and billing
- Salesforce
- HubSpot
- Dynamics 365
- Zoho
- Pipedrive
- Stripe
- Bill.com
HRIS and payroll
- Workday
- Rippling
- BambooHR
- UKG
- Paycom
- ADP
- Paychex
- Gusto
- Paylocity
- Justworks
If your system has an API, Claude can reach it.
Everything stays in your Claude, and it's all yours.
- The semantic layer, with every definition written down and owned by your finance team.
- The connection, running in your own Claude Team or Enterprise tenant under your Anthropic contract.
- The evaluation suite, with its questions, answers and pass history, ready to rerun.
- Yours to keep. Ours to keep certified. Everything lives in your tenant and belongs to you, with no AMW platform to log into and nothing to license. Each month-end close, reorg and new Claude version can move a number, so we rerun the suite and fix what drifts.
Every engagement is led by a former finance executive and built by a forward-deployed engineer working inside your Claude tenant. Adam Ware owns the scope, runs the definitions workshop and signs off on every certified number. The engineer builds the connections, the semantic layer and the evaluation suite.
An affordable fixed fee, set on the intro call.
We price the foundation by your stack, not by the hour: how many sources, how many metrics, whether data needs landing first. On a 30-minute call we confirm your systems and give you a fixed fee and timeline. No retainer is required to start.
FAQ
Do we need to be on Claude already?
Yes. We work only in Claude Team or Claude Enterprise accounts, because everything we build lives in your tenant. If you're not on Claude yet, that's the first step.
Where does our data go?
Nowhere new. The semantic layer sits in your systems, Claude runs in your tenant under your Anthropic contract, and AMW never hosts your financials. Users see only what their existing permissions allow, which matters most for restricted employee and customer data.
What access do you need from us?
Read access to the systems that hold your key metrics, a Claude admin to approve the connection, and two to three hours from your controller and FP&A lead for the definitions workshop. We send the full list after the intro call.
What if our numbers live in spreadsheets?
That's common, and it's often why Claude's answers drift. We move the spreadsheet logic into a governed form with your team, so the definition stops living in one analyst's workbook.
Can't our analysts build this?
They can build Skills, and they should. What usually doesn't get built is the foundation underneath: the connection, the governed definitions, and the monthly test against the official numbers. That's the part that decides whether you can trust what teams across the company build.
What happens when Claude changes?
The evaluation suite reruns on each new version of Claude. If anything stops tying out, we see it before your team does. Monthly recertification covers this after handover.
What if we stop after the foundation?
You keep everything we built. Without recertification, though, nobody is checking that it still ties out after each month-end close, and that's the job monthly recertification does.
Ready to give your teams numbers they can build on?
In 30 minutes we'll map where your key metrics live, determine which approach fits your stack, and provide an affordable fixed-fee quote.