New: Your Basalt workspace for Odoo. MCP, proactive agents, visualisation. Learn more

Basalt

For companies running on Odoo. Put AI to work.

Your Odoo is live. Get back the time it was meant to save.

We find the workflows taking days from your team, deploy the right agents and measure the time returned.

WorkspaceOdoo
  • Quote from a PDFDone
  • Quote follow-ups preparedDone
  • Supplier confirmation matchedDone
  • Stock variancesIn progressreading today’s stock moves

An agent prepares. A person reviews and sends.

Illustrative interface

AffexioEkipESL BrainsFairglowPrestant RealtyUbees

The problem

You run Odoo. Yet your team still loses days to repetitive work.

The work is in Odoo, but teams still rebuild it across modules, email and spreadsheets.

12 days

time lost

Rebuilding margin, cash or stock positions in Excel. Pulling the reports you need. Making sense of and finding the right data.

Every week

15 clicks, multiple tools

work scattered

The team ends up tracking part of the work outside Odoo.

To start a project

4 Odoo modules

context split

Manually searching for all the context before answering a customer question.

For one answer

Basalt Core

Four phases to return time, one workflow at a time.

We start with priority work and only build what has a real ROI.

01 · MAP

Map

We observe the real work and count the days it takes.

02 · DEPLOY

Deploy

We put the priority agents into production on your Odoo.

03 · MEASURE

Measure

We compare before and after, workflow by workflow.

04 · EXPAND

Expand

We choose what comes next from the results.

Industries

From factory floor to client door. We’ve got you.

The workflows change by industry. The method stays the same.

Use cases

What the agent prepares in your Odoo

Choose the module where your team’s time disappears. Each example shows what the agent prepares, what your team keeps in hand and what we measure.

Workflow 1

Customer order PDF turned into a draft quote

Context: the agent reads the attachment in the sales inbox and identifies the customer.

  • Lines matched to customer product codes
  • Order history used as a reference
  • Quote created as a draft, never confirmed

Workflow 2

Quote follow-ups sorted

Context: the agent compares every unanswered quote with that customer’s usual response time.

  • Quotes already handled removed from the list
  • Draft prepared, the salesperson still sends
  • Extension suggested before validity expires

Workflow 3

A regular customer slows down

Context: the agent measures each account’s ordering rhythm over two years, without a global threshold.

  • Rhythm measured account by account
  • Change compared with its own history
  • Activity assigned to the salesperson

Workflow 1

Supplier confirmation matched

Context: the agent retrieves the confirmation from the purchasing inbox and compares it with the purchase order.

  • Confirmation attached to the purchase order
  • Prices, quantities and dates compared
  • Activity opened on changed lines

Workflow 2

Late supplier line, downstream impact prepared

Context: the agent checks each morning for purchase lines whose receipt date has passed.

  • Late lines raised every morning
  • Follow-up prepared, the buyer reviews and sends
  • Impact carried to the customer order

Workflow 3

Supplier price drift

Context: the agent compares each received price with the supplier record and twelve months of orders.

  • Variance measured against the last price
  • Buyer alerted before confirmation
  • Increases grouped by supplier each month

Workflow 1

Delivery slip matched to the receipt

Context: the agent reads the delivery slip and matches it line by line with the open receipt.

  • Quantities, lots and serials prefilled
  • The warehouse operator still validates
  • Variances isolated, message kept as a draft

Workflow 2

Negative stock, cause found

Context: the agent detects when a location goes negative and traces it back to the move that caused it.

  • Product, location and move identified
  • Missing inbound move checked upstream
  • Valuation impact calculated

Workflow 3

Cycle counts ranked by risk

Context: the agent ranks locations and products by variance risk before the team starts counting.

  • Locations ranked by variance risk
  • Week sized to the time available
  • Each variance matched to its move

Workflow 1

Overdue invoices sorted by cause

Context: the agent traces an overdue invoice through the order, delivery and related exchanges.

  • Customer order reference checked
  • Overdue invoice classified by real cause
  • Blocked case routed to the right person

Workflow 2

Supplier invoice variance calculated

Context: the agent takes an invoice already flagged by Odoo and compares it with the order and receipt.

  • Price and quantity variances split
  • Unordered charges listed separately
  • Tolerance applied before approval

Workflow 3

Review before month end

Context: over the last five working days, the agent checks what could block the monthly close.

  • Missing documents and drafts listed
  • Tax checked against fiscal position
  • A correction suggested for each issue

Workflow 1

BoM corrected from actual use

Context: the agent compares quantities consumed on completed manufacturing orders with the BoM.

  • Actual use compared with the BoM
  • Only repeated variances retained
  • Revision suggested with orders attached

Workflow 2

Component shortfall, trade-off prepared

Context: each morning, the agent finds components that do not cover the week’s manufacturing orders.

  • Manufacturing orders linked to customer orders
  • Allocation calculated with replenishment time
  • Delayed reservations released after approval

Workflow 3

Subcontracting follow-up prepared

Context: the agent matches what left for the subcontractor with what has come back.

  • Outbound and return moves matched by order
  • Stock held at the subcontractor flagged
  • Message prepared, a person reviews and sends

Workflow 1

Project budget drifts, amendment prepared

Context: the agent compares logged hours and sold quantities milestone by milestone.

  • Logged hours compared with hours sold
  • Out-of-scope work found in message threads
  • Amendment prepared at contract rates

Workflow 2

Hours checked before invoicing

Context: every Monday, the agent compares the previous week’s timesheets with the sales orders.

  • Hours without an order line identified
  • Work above the allowance flagged
  • Reallocation applied only after approval

Workflow 3

Dormant project, close prepared

Context: the agent lists what remains open on a project with no activity for four weeks.

  • Sold quantities not yet invoiced listed
  • Items to invoice and cancel separated
  • Closing sheet ready for approval

The platform

One place to use, control and monitor your agents.

Your team stays in Odoo. Access, approvals and results remain visible.

Choose a platform feature

Claude reads only the records it needs and prepares the changes you allow.

A dedicated user with permissions you control.

Ask a question that needs context from more than one Odoo module.

Answers use only the records available to that user.

Each agent prepares one defined task. Your team keeps the decision and the send.

Prepared work and human approval remain traceable.

The useful information appears where the decision is made.

Built on your modules, not another export.

MCP connectorbounded access
SalesInventoryAccountingClaude
Illustrative diagram.
Chatthree modules
“Which customers have both a blocked order and an overdue invoice?”Sales · open orders3Inventory · unreserved lines5Accounting · overdue invoices2
Illustrative interface. Figures are examples.
Agent · Quote follow-updraft
  1. The agent prepares the follow-upQuote, last exchange and relevant linesdone
  2. You review and sendThe draft waits for you. Nothing leaves first.your turn
  3. The agent logs the sendIn the quote thread, with the follow-up datenext
Illustrative interface. A person still sends.
Management viewMonday

6workflows tracked

This week

Quotes to review
38
Invoices to reconcile
21
Stock variances
9
Needs review
6
Illustrative interface. Figures are examples.

Our process

From the first call to the first workflow in production.

Three concrete steps. You know what we inspect, what we build and what we measure.

Video call with a member of the Basalt teamDiscovery call30 min

Step 1

First call

We look at one task that comes back every week and the Odoo modules behind it. In 30 minutes, we can usually tell whether it is worth mapping.

Action planIn progress
Time today2 days/wk
First scopeSales
  • Quotes from PDFs
  • Follow-ups to prepare

Step 2

Baseline and action plan

We observe the task, measure the current week and write the first scope. You approve the baseline before we build.

This task comes back every week and needs the same checks each time.

We will measure the baseline, prepare the workflow and keep the final approval with your team.

Step 3

Production and measure

We deploy the workflow, watch how the team uses it and compare the result with the baseline.

Your Odoo modules

From the quote to accounting, inside your Odoo.

The agents work in the modules your team already uses.

Odoo SalesOdoo CRMOdoo PurchasingOdoo InventoryOdoo ManufacturingOdoo AccountingOdoo Projects

Pricing

A clear cycle, with a published starting price.

The first cycle covers one team and the workflows agreed during Map.

Basalt Core

Monthly

Map, deploy and measure a first scope. No setup fee.

Starting from

€3,500

a month

Book a call

The call is for deciding what is worth measuring.

In the first cycle

  • Observe the real work
  • Fix the baseline
  • Deploy on your Odoo
  • Train the people involved
  • Compare before and after

Illustrative diagram of hours returned over the next three months

5 h

W1-4

14 h

W5-8

30 h

W9-12

50 h

W13+

Customers

What changed in their work.

Concrete results, in their words.

Within a few weeks, we had cut down several time-consuming processes, saved 10+ hours every week, and even won new contracts.
Founder & President
We've cut proposal preparation time by more than 60%, allowing us to send proposals much faster and focus our team's time on higher-value work.
Founder & CEO
In under two weeks, Basalt deployed a Company Brain, automated our marketing and technical reports, and built AI agents around our existing workflows.
Founder & CEO
Working with Basalt helped us adopt AI much faster than we would have on our own.
CFO
We really appreciated working with Basalt. The team was always available, responsive, and transparent throughout the project.
Chief Operating Officer
Today, we save hours every week, find information instantly, and spend our time on work that actually creates value.
Founder & CEO
Basalt redesigned our sales and proposal workflow so that most of the client information is transferred automatically into a ready-to-customize presentation.
Founder & CEO
We've automated several repetitive tasks, save hours every week, and can focus more of our time on growing the business.
CFO
It's clear how much effort went into exploring different approaches and pushing the use case as far as possible.
Chief Operating Officer
I was initially hesitant about AI, but Eliott took the time to understand my business and showed me exactly where it could create value.
Founder & President
Impressive work considering the complexity of our process. I'd recommend Basalt to any company looking to eliminate repetitive admin work and give their teams more time to focus on customers.
Founder & CEO
Our team now spends far less time on repetitive work and more time on what really drives the business.
Founder & President

Questions

What you need to know before the first call.

The first cycle

Where do you start?

With one recurring task that takes visible time from the team. We observe how it is done, which modules are involved and where the exceptions sit before proposing anything.

How long before the first workflow is live?

Most companies have a first workflow in production within the first month. Timing mainly depends on data access and the time needed to approve exceptions.

What do you need from our team?

Less than two hours to observe and scope the first workflow, then about 30 minutes a week with the people involved during deployment.

Do we need to finish our Odoo project first?

No. We work with the Odoo you use today. If a data issue prevents the workflow from working, it becomes part of the scope rather than being hidden.

Pricing and results

How much does Basalt Core cost?

Basalt Core starts at €3,500 a month. The exact scope depends on the team and workflows agreed during Map. There is no setup fee.

How do you measure the result?

We fix the baseline before deployment and compare the same workflow after it is live. The pricing curve illustrates a possible trajectory, not a guaranteed result.

What if the result is not there?

The measurement makes that visible. We fix the workflow, change the priority or stop expanding it. An agent without a result does not become an endless project.

What do we keep if we stop?

The configuration, agents, prompts and documentation built for your company. We then remove our access.

Data and control

Where is our data processed?

Odoo remains the source of record. Before production, we document the fields used, the providers that process them, retention and the access granted.

Can an agent send a message on its own?

No. The agent prepares, triages and logs. A person reviews and sends every customer or supplier communication.

What can an agent do without approval?

Only what you have explicitly authorised. Work touching a customer, supplier or money remains subject to human approval.

Do you replace our Odoo integrator?

No. Your integrator maintains and develops Odoo. Basalt handles the AI workflows built across your modules and works with the integrator when the scopes meet.

Bring us one workflow that takes too much time.

In 30 minutes, we will see whether it is worth mapping.

Book 30 minutes
  • Start with a real task
  • Look at the modules involved
  • Identify the control points
  • Choose a clear next step
Basalt | Get Back the Time Odoo Was Meant to Save