Here are all the essential ERP and Odoo news of May you should not miss.
News from Odoo HQ
Official updates and posts from Odoo’s headquarters, Odoo employees, and Odoo founder Fabien Pinckaers.
OWL 3.0 introduces major architecture changes
Odoo just shared the development notes for Owl 3.x, the upcoming version of its custom UI framework. Because the Odoo codebase has grown so much in recent years, the developers hit a few structural limits. Owl 3 addresses these by introducing better typing, predictable performance, and a reworked reactivity system.
It includes several breaking changes, but the end goal is to make future development simpler and more stable.
Key changes include:
Signal-based reactivity: Owl 3 replaces useState with "Signals," computed values, and effects. Instead of using implicit proxies (which can sometimes cause hard-to-find bugs), Signals explicitly track when and where data changes. This makes UI updates much more predictable.
Plugins replace 'env' and Services: The traditional env object and Odoo services are being phased out in favor of a new Plugin system. Plugins are self-contained, type-safe building blocks that are easier for IDEs to read and autocomplete.
A reworked Props system: Component props will now be accessed through a direct props() function call rather than static objects. Also, the compiler will automatically detect and optimize arrow functions (the .alike behavior), preventing unnecessary child component re-renders without needing manual setup.
Strict template rendering: To prevent scoping bugs, templates will no longer implicitly guess where a variable comes from. Accessing component data will now require an explicit this. prefix (for example, writing <t t-out="this.value"/> instead of just value).
Cleaner template directives: The old t-esc directive is retired in favor of t-out. Additionally, t-slot has been renamed to t-call-slot to make its purpose clearer, and t-call is now restricted to <t> nodes.
Simplified References & Forms: Both t-ref (for accessing DOM elements) and t-model (for two-way data binding) have been rewritten to work natively with the new Signals system, simplifying how data is passed between parent and child components.
Looking ahead to the Owl 3 migration:
Since these are structural breaking changes, migrating from Owl 2 to Owl 3 will take some effort, especially updating templates to include the new this. prefix.
To help, Odoo plans to provide automated scripts to handle the majority of the template migrations. They are also building temporary compatibility layers (like a useEnv hook) so that developers won't have to rewrite their entire architecture on day one. The team's expectation is that these updates will lead to a simpler and more maintainable codebase.
Native voice dictation added to AI chat for Odoo 20
Voice-to-text is officially coming to Odoo's AI chat. Odoo's Fabrice Henrion recently teased an "Ask AI" window in listening mode on LinkedIn and shared a new GitHub commit that confirms a native voice dictation feature.
The update does not just add a microphone button. It improves Odoo's core voice recorder system to support the feature across the whole platform.
What to expect:
Voice-to-text dictation: Users can speak directly to the AI assistant. You can search your database or give commands without typing.
Upgraded recording UI: The voice recorder interface has a fresh design. It now includes a live audio visualizer, a clear timer, and dedicated confirm and cancel buttons.
Under-the-hood improvements: The core code is now much more stable. It handles audio monitoring, file encoding, and recording limits more reliably.
Odoo 19.3 release notes are here
Odoo has published the release notes for version 19.3. Here is a breakdown of some of the most interesting updates:Â
Artificial Intelligence
AI is being woven into the core workflow to act as a digital assistant for data and design.
AI Record Management: You can task AI agents to create new records or update existing ones. This allows the system to populate data or change statuses based on your natural language instructions.
AI Image Generation: Odoo now features built-in AI image generation. You can generate original visuals directly within the platform for product placeholders or marketing materials.
AI Website Assistant: The website builder includes an assistant that helps with choosing themes and generating content. It acts as a digital designer to help you launch or refresh pages in a fraction of the usual time.
Mobile & Offline Productivity
Universal Offline Actions: Offline capability is no longer limited to specific apps like POS. You can now create, edit, or archive records across the entire Odoo ecosystem without a connection. Your changes are saved locally and automatically synced the moment you’re back online - perfect for uninterrupted work in "dead zones" like large warehouses or remote field sites.
Translation Mode: A dedicated Translation mode has been added, which can be quickly accessed via the command palette to manage multi-language content on the fly.
Accounting & Online Payments
These changes aim to centralise your financial tasks and improve error correction.
Direct Bill Payments: You can now pay vendor bills directly from the Odoo interface through integrated service providers. This removes the need to manually log into your banking portal to settle every invoice.
Bulk "Reset to Draft": In list views, you can now select multiple records and reset them to Draft in one go. This is a massive time-saver for fixing data entry errors across dozens of invoices or orders at once.
Payment Method Updates: New integrations include Wero and expanded Stripe support for Alma and Swish. As per your note, SOFORT has been removed from the supported list.
Logistics, Barcode & Label Design
These updates focus on warehouse precision and retail transparency.
General Label Design: Odoo has updated the core tool used to design labels across the system. This means you have more control over the layout and look of product tags, shipping labels, and asset stickers.
Retail Price Transparency: In the Point of Sale, you can now include the base unit price directly on product labels. This helps customers compare the value of different sizes (e.g., "price per litre") and ensures your shop stays compliant with local pricing regulations.
Advanced Barcode Scanning: The Barcode app now supports manufacturing operations and pre-encoded packages. You can now scan a package and the system will automatically recognise everything inside it without you needing to scan every single item individually.
Untracked Goods: You can now handle goods within packages that do not require individual serial or lot tracking. This simplifies movements for high-volume, low-value items.
Manufacturing & Maintenance
Kanban View for Orders: Manufacturing orders can now be managed via a visual Kanban board. This gives production managers a "bird’s-eye view" of which jobs are in progress and which are stalled.
Component Replacement: There is now a streamlined process for swapping components or handling changes mid-production. This is essential for keeping the production line moving when specific parts are out of stock.
DRC delegation visits Odoo HQ ahead of 1st African Odoo Experience
Odoo recently hosted 19 officials from eight public institutions in the Democratic Republic of the Congo (DRC) to showcase how digital solutions can streamline the public sector. Â
The group travelled over 10,000 km to see first-hand how Odoo collaborates with African institutions. The visit aimed to give these representatives practical insights to help speed up the DRC’s ongoing digital transformation. Â
The meeting sets the stage for Odoo’s growing presence in Africa, ahead of the first-ever Odoo Experience event in Kenya on 3 and 4 September.
Odoo hints at new AI agents for version 20
In a recent webinar, Odoo showcased the newest AI features from its 19.1, 19.2, and highly anticipated 19.3 updates. They also dropped some hints about Odoo 20 and the future looks very agentic. Â
New features include:
- Update records via chat: Use simple instructions to update your existing database records. Â
- Create new records: Ask the AI to build new entries, like customer profiles or sales opportunities, from scratch. Â
- Generate images: Create or edit product photos, such as putting a basic picture of an item into a realistic background.Â
- 'Vibe code' websites: Build a full website layout and theme just by describing what you want. Â
- Web search: AI agents can now look up information online, meaning you can share a specific URL to use as a design reference. Â
- SEO optimisation: Automatically write SEO-friendly titles, descriptions, and keywords to help your site rank better on search engines. Â
- Upload documents in bulk: Upload multiple files at once to teach your AI agent about your business. Â
- Refresh AI knowledge manually: Click a button to update what your AI has learned from your documents, instead of having to delete and re-upload them. Â
- Ask AI about documents: Summarise or find specific details inside a file directly from the preview screen, saving you from reading through the whole text. ​
Looking ahead to Odoo 20:Â
The next major update will focus heavily on making AI a core, independent part of your daily operations.
Odoo is also hoping to expand the AI models you can choose from. Right now, the software uses ChatGPT and Gemini, but the team is actively trying to add others. A key addition would be Mistral, which European clients have been requesting so they have a local alternative.
The developers have a tight four-month deadline before the September launch. Adding support for these new models is a major goal for version 20, but if they run out of time, it might be delayed until version 21.
Community talk
All the hot topics and discussions Odoo users are buzzing about this month.Â
Odoo 20 to expand HMRC integration beyond VAT
Attention British Odoo users: word on the street is that version 20 is bringing some massive accounting upgrades to the UK.
Odoo expert William McMahon shared some early news online. While Odoo has handled MTD for VAT for a while now, it looks like they are finally expanding their HMRC integration to cover the rest of the British tax ecosystem. If all goes according to plan, the UK will get native support for:
MTD for Income Tax (word is this is already built and just waiting on HMRC approval)
CT600 (Corporate Tax) filings
Year-end accounts
Self-assessment submissions
This would be a huge step forward for UK localisation and enterprise adoption. But as a few users rightly pointed out, there is an obvious catch: this is only a real win if Odoo actually commits to supporting and maintaining these tools every time HMRC decides to change the rules.
Migrating legacy Odoo data into live systems - a recipe for disaster?
A recent post on Reddit has rehashed the debate around a classic Odoo headache: moving historical data from an old database directly into an active system. The discussion kicked off when a user wanted to pull nine months of records from Odoo 11 Community straight into a running Odoo 19 Enterprise setup.
The Odoo community was quick to wave red flags. Trying to merge data across that many version jumps is notoriously messy. As several developers pointed out, standard upgrade tools simply aren't built for dumping partial records into a live database without causing system conflicts.
Instead of attempting a straight migration, the consensus is to treat this as a highly controlled data import project. The safest route recommended by Odoo veterans includes:
Exporting only the exact records you need.
Using custom scripts to map the data to the new architecture.
Testing heavily in a staging environment to avoid nightmares like duplicate contacts or broken accounting entries.
Odoo.sh deletes staging instances after 30 days
Odoo.sh users, take note: From 31 May, staging instances will be deleted after just 30 days instead of 90 days up until now.
Once that month is up, you will have to rebuild the instance or start over from scratch.
Odoo didn’t actually announce this change. Partners discovered it tucked away in the documentation. Where the text once said staging databases were "intended to last up to three months," it now states they are "automatically deleted after one month". If you want to use the branch again, you’re now required to rebuild it. Â
The move has frustrated the community online. Users have criticised Odoo for the lack of notice and argued that fitting every test into a 30-day window is simply not realistic.
Odoo’s leadership is standing by the decision, however. They argue that staging is meant for quick code fixes rather than long-term storage. Product owner Fabrice Henrion stepped in to defend the move, claiming that testing a code change shouldn't even take a week, let alone 30 days.
But for many Odoo partners, that logic doesn't hold up. They’ve pointed out that things like user training, complex data migrations, and full UAT cycles often take months to complete, making the new 30-day limit a significant hurdle.
AI-driven Odoo development - blessing or curse?Â
After the recent integration of Claude into Odoo.sh, some lively debates have kicked off online: are companies losing control of their own software?
The new setup gives businesses huge coding power. Projects that once took months and cost thousands can now be knocked out in minutes. For companies on a tight budget, it feels like striking gold, as it makes testing new ideas cheap and risk-free.
But veteran developers are urging caution. They warn that relying purely on AI often creates messy "spaghetti code" and hidden bugs that are a nightmare to fix. In core business systems, stability beats speed. Without strict oversight, 'vibe coding' quickly snowballs into massive technical debt.
Ultimately, this argument is likely here to stay. Much like the long-running debate over Odoo Studio, the clash over AI-driven development will probably become a permanent fixture in the community. After all, AI-powered customisation has really only just started, and its influence is only going to grow.
News from the ERP worldÂ
Headlines from the ERP world you shouldn’t miss.
EU bans US cloud services for sensitive data
The EU Commission has officially banned the use of US cloud services for sensitive data in the healthcare, finance, and justice sectors. This move affects major providers like Amazon Web Services, Microsoft Azure, and Google Cloud, which currently control about 80% of the European market. The ban is a response to the US Cloud Act, which allows American authorities to access customer data regardless of where it is stored.
How the industry is adapting:
Strict new rules: Alongside the ban, a new regulation introduces eight sovereignty targets and a "sovereignty score" to rate cloud infrastructure providers.
Local alternatives step up: European providers are quickly filling the gap. T-Systems' 'T Cloud Public' has just been approved for government use, keeping all data strictly within German data centres.
A new European CRM: T-Systems and BSI Software are building a purely European customer relationship management platform. This gives banks and insurers an option completely free from US data rules.
Fujitsu and Anthropic team up: Fujitsu is bringing the Claude AI model into its secure infrastructure to offer a compliant option for regulated sectors.
Clear AI deadlines: The EU AI Act timeline is now fixed. Rules for AI transparency take effect in December 2026, and high-risk AI systems must comply by December 2027.
Rising security threats: These rules arrive at a critical time. Banking trojans increased by 196% early this year, and total GDPR fines have now passed the €6.11 billion mark. Many companies are also struggling to safely control autonomous AI agents.
Looking ahead: The tech sector faces a busy few years. By June 2027, the updated BSI (Federal Office for Information Security) standards will become mandatory for providers. Meanwhile, industry groups and politicians are pushing hard for "ethical AI" to reduce Europe's heavy reliance on foreign hardware and software imports. The real test will be whether smaller businesses adopt local alternatives like Hetzner or Mistral AI over the coming years.
Microsoft trials AI tool to speed up Dynamics 365 rollouts
Microsoft is launching a private preview of "Project Mia", a new tool designed to speed up how businesses set up Dynamics 365. The system uses AI agents to automate the tedious parts of configuring enterprise software.
The testing is starting this month with a small group of partners focusing on Finance and Supply Chain Management. Microsoft plans a wider public preview later this year, aiming for a full launch before the end of 2026. The technology will eventually expand to cover Sales, Customer Service, and Contact Centre platforms too.
Asana buys StackAI for $75m to automate cross-app tasks
Asana has acquired StackAI for around $75 million, shifting its focus from simple project management to broader workplace automation. The deal was announced alongside the company’s latest quarterly earnings.
StackAI is a no-code platform built to integrate different software tools. While most AI assistants are trapped inside a single app, StackAI lets users build workflows that connect separate systems, like Salesforce, Slack, and Google Workspace. Asana has already been testing the platform internally, using it to automatically pull data from various marketing tools to track its SEO spending.
The acquisition comes at an important moment for Asana. While its quarterly revenue grew by 9.5% to $205.1 million, its share price has been under pressure as investors look for proof that AI features can drive growth.Â
The move puts Asana in direct competition with the likes of Salesforce, Monday.com, and Zapier, which are all rushing to build similar automation tools.
SAP presents AI strategy following API lock-out backlash
- Built-in AI Assistants: SAP is putting over 50 AI assistants directly into core modules like finance and HR. Its main interface, Joule Work, uses Anthropic's language models to handle multi-step tasks.
- Walled Garden Data: To justify blocking outside AI, SAP claims its native tools are safer because they are tied directly to its own Business Data Cloud and a new Knowledge Graph.
- €100m Fund & n8n: To stop developers from using external platforms, SAP launched a €100 million fund to help partners build custom agents. These tools will plug into 'Joule Studio', which now integrates the automation platform n8n.
- On-Premises U-Turn: In a major concession to unhappy users, SAP dropped its cloud-only rule for AI. Older on-premises systems can now connect to these tools, provided the customer has started moving to the cloud.
While the strategy addresses the current skills shortage, most of these features are still in early testing stages. Crucially, SAP has not yet revealed any pricing models, leaving clients in the dark about what these tools will actually cost once free promotional periods end.
Hackers breach GitHub and steal internal code
Open-source giant GitHub has confirmed a significant security breach after a group of cybercriminals used a "poisoned" Visual Studio (VS) Code extension to compromise an employee's device. The attackers claim to have exfiltrated nearly 4,000 internal private repositories, which they are now attempting to sell on dark web forums. Fortunately, GitHub’s initial investigation confirmed the blast radius is contained: only internal source code was taken, and no customer data was compromised. Â
A huge supply chain campaign
This breach isn't an isolated incident. It is part of an aggressive, automated supply chain campaign that has already hidden credential-stealing malware in over 500 open-source software tools.Â
By hijacking trusted developer utilities, the hackers steal active API keys to plant self-spreading malware - a vicious cycle that has already touched high-profile organisations like OpenAI, Anthropic, and the European Commission.
SAP invests in German AI startup n8n
SAP has just poured a reported $60 million into the Berlin-based AI automation platform n8n. The deal doubles n8n’s valuation to $5.2 billion, making it one of Germany’s most valuable AI companies right now.
This isn't just a financial investment. SAP is actually building n8n’s tools directly into its own AI development platform, Joule Studio. This will give users a visual, drag-and-drop canvas to build and connect their AI workflows right inside SAP's system.
The move is a big step toward what SAP is calling the "autonomous company". Pitched at its Sapphire event this week, the ultimate goal is to reach a point where AI assistants can take over complete, everyday business processes, working right alongside human staff.
Anthropic targets SMBs with 'Claude for Small Business'
Anthropic has launched Claude for Small Business (CSB), a new desktop plugin designed to automate administrative tasks like payroll, accounting, and forecasting.
Featuring native connectors for tools like QuickBooks, PayPal, HubSpot, and Canva, Claude can now pull data directly from your existing systems. It uses 15 specialised workflows to handle manual tasks such as:Â
reconciling QuickBooks cash positions with PayPal to build 30-day forecasts,
generating month-end close packets, or evenÂ
launching marketing campaigns based on real-time sales data.
Please note: unless you're on a Teams plan, Anthropic opts you in to training its models on your chat data by default, so you'll need to manually opt out if privacy is a concern.
SAP buys Dremio to power AI agents with open-source data
SAP has purchased the data lakehouse platform Dremio to build a stronger foundation for its new enterprise AI tools.
The main goal is to fix a big headache for SAP users: having their information split across lots of different systems. By picking up Dremio, SAP is adopting Apache Iceberg. Because it is an open-source format, companies can finally link their SAP and non-SAP data together without the hassle of moving or converting any files.
When AI agents can look at one unified system, they instantly understand the true context behind the information and can complete their tasks far more accurately.
For years, SAP has kept its data locked away in its own exclusive formats. Now, they are using Dremio’s tech to open up their Business Data Cloud and make it ready for AI. Expect to see these new features roll out across SAP’s cloud platforms by the second half of 2026.
Salesforce introduces new AI platform to automate the back office
Salesforce has just launched Agentforce Operations, a new platform built to bring AI directly into backend processes. While the company is best known for its customer-facing software, this latest release tackles the repetitive, behind-the-scenes work that often bogs down large organisations.
Instead of relying on staff to manually handle data entry, inventory checks, or supply chain oversight, the system assigns these chores to autonomous AI agents. The underlying technology comes from Salesforce’s acquisition of Regrello, an AI operating system originally designed for manufacturing and supply chains.Â
Salesforce claims the potential time savings are massive. According to the company, the agents can speed up process cycle times by up to 70% and wipe out 80% of manual chores like data entry. In one given example, an AI agent could finish an audit in just 60 seconds, a task that would typically take a human team four hours to complete.
Oracle NetSuite rolls out AI coding tools for developers
Oracle NetSuite has introduced "SuiteCloud Agent Skills" to help developers build custom apps much faster. Instead of restricting these tools to its own ecosystem, NetSuite is sharing its specific coding rules and best practices across more than 25 different AI platforms.
Using simple language, developers can now ask AI assistants to write NetSuite-specific code, update older legacy scripts, and automatically apply the correct security permissions.
While experts believe this open approach will save a lot of time, they warn that major businesses might be slow to adopt it due to their existing IT investments. For now, analysts recommend testing the AI cautiously in isolated environments to avoid accidentally breaking current workflows.
Let’s talk Odoo
Reach out to our Odoo experts to see everything Odoo can do and how to tailor it to your business.