AI inside the dv* extensions you already run
If you are already a Davisa customer, AI lands as a new layer on top of the dv* extensions you already have live. No migrations, no vendor change.
See use cases on dv* extensions →AI applied to your Business Central, with numbers and a deadline
Davisa AI Studio is the AI consultancy practice of Davisa Informática. We design, build and ship AI use cases on Microsoft Dynamics 365 Business Central and the dv* extensions, on the Azure stack, with your data kept inside your tenant. We don't activate Copilot in bulk: we do sector-aware engineering on your real processes, with a 5-day Discovery, 8-week pilots and continuous scaling.
Three entry angles
Depending on where you are today, there are three reasonable ways to start. Pick the one that best fits your starting point; you can always talk to the team if you're between two.
Tailored 8-week AI pilot
You have a clear case: invoice extraction, scrap prediction, automated close. Fixed scope, measured KPI, go-live at the end of week 8.
How the pilot works →5-day AI Discovery
You don't know where to start. In one week we map your candidate processes, prioritise by ROI and deliver an actionable roadmap.
Request Discovery →Use case catalog
These are the use cases we're getting asked for most in 2026. Each has its own page with technical scope, data requirements, timeline and expected KPI range. Not an exhaustive list: if your case isn't here, we'll discuss it in the Discovery.
For CFOs · cross-industry
AI-driven supplier invoice automation
Document extraction, matching against the purchase order and proposed analytical posting. The invoice lands in BC without manual typing.
See use case → For plant managers · manufacturingIndustrial scrap prediction
Models trained on production history that flag scrap trends before the batch is finished. Decisions taken on shift, not at month-end close.
See use case → For CFOs · cross-industryAnomalies in accounts payable
Detection of duplicates, off-pattern amounts and suppliers with unusual behaviour. A filter sitting before the approval matrix.
See use case → For executives and family officesAutomated executive close summary
At month-end AI drafts an executive summary in plain language on financial performance, with explained variances and recommended next steps.
See use case → For constructionSite manager assistant
The site manager asks in plain language about progress certifications, retainage, critical line items, and the system answers using live BC data and documentation.
See use case → For HR, quality and managementInternal knowledge assistant with AI
A RAG assistant that answers team questions by querying your internal documentation. Faster onboarding, senior staff freed up, knowledge no longer tied to specific people.
See use case → For IT and BC maintenanceAI support for the technical department
A technical copilot that helps your IT team triage BC incidents, suggest AL customisations, read release notes and maintain documentation. Assisting, not replacing.
See use case →Architecture: data on-tenant
We work end-to-end on the Microsoft Azure stack. The AI is deployed in the customer's tenant or in a managed tenant under their contractual terms, and business data never leaves that perimeter. This matters because most cases we work on (accounts payable, fiscal close, construction site documentation, plant data) are sensitive and cannot send content to public models. Compliance aligned with GDPR and compatible with ENS depending on tenant configuration.
- Azure OpenAI Service: language models deployed in your Azure subscription. No data shared with public OpenAI.
- Copilot Studio: custom conversational agents over your BC data and processes.
- AI Builder: predictive and classification models integrated into Power Platform and BC.
- Document Intelligence: structured extraction of invoices, delivery notes, contracts and construction documents.
- Native Business Central integration: read and write via standard API, no middleware or overnight sync jobs.
The Davisa AI Studio method
Three phases. Each one with a concrete deliverable, a fixed timeline and clear pricing. The goal is for you to know where you stand and which decision is due at the end of each stage.
Discovery
5 daysWorkshop with your team. Mapping of AI-candidate processes, technical feasibility assessment (data availability, quality, integrations), ROI estimate and prioritisation by impact.
Pilot
8 weeks · fixed scopeDevelopment, testing and go-live of a chosen use case from the roadmap. KPI measured before and after against the baseline set in the Discovery. If the target is not reached, we iterate until the agreed scope is closed.
Scaling
ongoingNext use cases from the roadmap, optimisation of the ones already in production, training of the internal team and support of deployed AI components. We move into a maintenance and continuous improvement relationship, not a one-off project.
Case 0 · Before we sell it, we use it at home
At Davisa we deploy every AI solution first in our own processes. Today AI is live in eight fronts of our daily operations, across four domains.
Finance and payments
- Supplier invoice and delivery note intake. Automatic document extraction, matching against the purchase order and proposed analytical posting. The invoice lands in BC without manual typing.
- Automatic bank payment reconciliation. Bank statement movements are matched against BC ledger receipts and payments without manual intervention. What used to be a half-day a month is now just review.
Sales and communication
- Sales follow-up. Information from email, Teams meetings, calls and sales notes is organised and enriched automatically in the CRM. The salesperson stops copying things in by hand.
- Inbox analysis and triage. Inbound email is classified, prioritised and routed automatically. The team stops wasting time deciding what to read first.
Projects and internal knowledge
- Detection of critical project points. The system identifies critical milestones, emerging risks and plan deviations before a person catches them.
- Internal documentation classification and search. An assistant answers team questions by consulting Davisa manuals, procedures and precedents, without digging through SharePoint.
- Technical department support. Our consulting team queries a technical copilot that knows Microsoft documentation, the code of the dv* extensions and Davisa's implementation precedents. Faster incident diagnosis and customisation proposals.
Marketing and content
- Marketing department automation. Generation and refresh of web content, continuous SEO audit, anti-cannibalisation, editorial planning, social posts and covers. The site you are reading is, to a large extent, the result.
Before we propose automating a process in your company, we've already hit the bumps in our own. And we've fixed them.
The AI Studio team
Davisa AI Studio is staffed by a senior in-house team, not rotating consultants. It combines data and AI engineering with deep sector knowledge (construction, manufacturing, wealth and distribution) and more than twenty years of Microsoft Dynamics 365 Business Central implementation experience. We are not a generic AI consultancy landing with templates: every case is reviewed by an engineer with sector expertise and a BC consultant with hands in production.
The day after go-live, the people responsible for the case are the same ones who built it. There is no handoff to an external support team: the AI stays integrated within the same operating relationship you already have with Davisa for BC and the dv* extensions.
When Davisa AI Studio is NOT the right fit
There are four situations where the honest answer is to point you elsewhere. If you recognise yourself in any of them, skip the Discovery and look at another option.
- If you're looking for a chatbot on your standard public website. For that there are cheaper and faster SaaS tools (Intercom, HubSpot, Tidio, etc.). We have no edge there.
- If your company doesn't have Business Central yet. ERP first, AI on top after. Building AI on unstructured data or on a legacy ERP delivers poor results and operational pain. The right order is BC first, AI Studio second.
- If you expect results in two weeks without a Discovery. The shortcut gets expensive. Without understanding your data, your processes and your baseline, any AI pilot is a lottery. If you're genuinely short on time, the Discovery can be compressed, but skipping it is not viable.
- If your real problem is process, not technology. AI accelerates what works and amplifies what doesn't. If the bottleneck is organisation, data governance or unclear decision rights, AI doesn't fix it. It makes it worse.
Frequently asked questions
What exactly does Davisa AI Studio do?
We design, build and roll out AI use cases on Microsoft Dynamics 365 Business Central. We don't resell Copilot licences in bulk: we engineer on top of your real processes using Azure OpenAI, Copilot Studio, AI Builder and Document Intelligence, and we keep your data inside your tenant. The standard scope is a 5-day Discovery to diagnose, an 8-week pilot to deliver one measured use case, and continuous scaling for the rest.
Do I need Business Central in place to work with Davisa on AI?
To get the most out of it, yes. Our value proposition is native integration with Business Central and the dv* extensions. If BC is not yet live, the honest answer is to start with the ERP first: AI on top of broken processes or unstructured data delivers little. In that case the right path is BC implementation with Davisa first, then the AI Studio on that foundation.
Does my data leave the tenant during the project?
No. We work with Azure OpenAI Service deployed in your Microsoft tenant or in one managed for you under your contractual terms. Training data, prompts and responses are not sent to public OpenAI nor used to train third-party models. GDPR-compliant and ENS-compatible depending on your tenant configuration.
How much does the Discovery cost?
The 5-day Discovery is a fixed-scope package with on-request pricing. We don't publish it because it depends on the number of processes to map and the required technical depth. Request it from the form on /en/ia/discovery/ and we'll send a written proposal within 48 business hours.
What happens if the pilot doesn't hit the expected result?
In the Discovery we set the pilot KPI target (for example: cut invoice capture time by 60%) and the current baseline. If at the end of the pilot the KPI does not reach the target, your team and ours sit down to review what failed: data assumptions, source quality, prompt scope or process change. We adjust and extend at no extra cost until the agreed scope is closed or, if the case is unviable, we document it and propose an alternative.
Why trust Davisa for AI and not a pure AI consultancy?
Because useful enterprise AI is not just a model: it is model plus data plus process plus ERP. A pure AI consultancy delivers a proof of concept and walks away. We have been a Microsoft Solutions Partner for Business Central since 2003, we know the real processes in construction, manufacturing, distribution and wealth, and we leave AI integrated into BC with the same team that will run your next fiscal close. The difference shows on day 200 after go-live: the use case is still working.
Next step
Already a Davisa customer?
The AI layer lands as an extension of what you already have. We talk to your usual advisor to frame the case inside the current relationship, with no new onboarding.
Talk to the team →New to Davisa?
We start with the 5-day Discovery. Process mapping, technical assessment and a prioritised roadmap so that in one week you know exactly where it makes sense to invest.
Request AI Discovery →Want a rough impact estimate first? Try the AI ROI calculator before requesting the Discovery.