Bisharp
All-in-one Dutch business intelligence and AI platform for manufacturing, construction and installation firms. It pulls ERP, finance, HR and marketing data into ready-made dashboards, automated reports, screen displays and an AI analyst reachable from Claude or ChatGPT.
What is Bisharp?
Bisharp is an all-in-one business intelligence and AI platform built by a Dutch company of the same name and aimed squarely at three sectors: manufacturing, construction and the installation trades. Its stated purpose, echoed in the company's own registry filing, is to give owners and managers a single cloud place to find every figure they need to analyse properly, improve processes and take decisions.
The platform is built from seven modules. Dashboarding delivers sector dashboards that arrive pre-filled rather than as an empty canvas, covering purchasing, logistics, production, sales, finance, marketing and HR. Reporting Automation sends ready-made PDF or Excel reports to a mailbox on a schedule the user sets, which also makes it easy to feed sales KPIs to a sales team or delivery reliability figures to a supplier. Digital Data Signage puts the same dashboards, plus internal announcements, on screens in the workshop or canteen. Builder lets customers assemble their own dashboards. Data Warehousing and an API are sold as add-ons.
The two AI modules are where the platform earns its name. The AI Data Analyst answers questions about a dataset in ordinary language and returns summaries, forecasts and recommendations; its distinctive trait is that it can reason across datasets pulled from several dashboards at once. It is metered in analysis credits, one credit per conversation, with 100, 200 or 300 per month depending on the package. The AI connectors expose a Model Context Protocol server that a customer adds to Claude, ChatGPT, Gemini or Copilot as a custom connector, after which the assistant can read ERP and accounting data, run analyses and send emails or presentations straight from the chat. The connector is deliberately read-only, and permissions follow the roles a key user sets, so a fitter cannot reach the finance figures and a buyer cannot reach HR data.
Twenty-five ready-made connectors do the plumbing, covering Dutch ERP and accounting packages, Google and LinkedIn marketing data, and fleet telematics. Technical go-live is claimed within a day; the vendor itself recommends three to five implementation days. Bisharp holds an ISO/IEC 27001:2022 certificate, and data is processed in the Netherlands with encrypted backups in Ireland.
What it does
- Analyse any dataset in natural language and get back a summary, a forecast or a recommendation
- Combine datasets drawn from several dashboards at once, for instance sales figures alongside finance
- Let Claude, ChatGPT, Gemini or Copilot query your ERP and accounting data through a read-only MCP connector
- Connect ERP, accounting, marketing and fleet systems through twenty-five ready-made connectors
- Read sector dashboards covering purchasing, logistics, production, sales, finance, marketing and HR
- Receive scheduled PDF or Excel reports by email, daily, weekly or monthly
- Push dashboards and company news onto screens in the workshop, the canteen or the office
When to use Bisharp / When not to
A quick filter to help you decide if Bisharp is the right fit.
When to use Bisharp
- Managing directors and business owners of Dutch manufacturing, construction or installation SMEs, from roughly 15 to 400 staff, who want one place for every figure
- Finance controllers and reporting managers who currently rebuild the same margin and cash overviews by hand in Excel every month
- Project managers and site supervisors who need to see which projects are drifting off plan while there is still time to steer
- Purchasing, logistics and production managers who want their own sector dashboards without waiting for an internal BI team
- Teams that already work daily in Claude, ChatGPT, Gemini or Copilot and want those assistants to read the company's own ERP and accounting data
When not to use Bisharp
- Buyers who need a published price before talking to anyone: all three packages carry a quote-only label and a contact form
- Non-Dutch-speaking organisations: the website, the contract documents and the customer portal are served in Dutch only
- Companies running ERP or accounting software outside the Dutch market, where the twenty-five ready-made connectors give no head start
- Teams that need an AI assistant to write back into the ERP, since the connector is deliberately read-only
- Operations that depend on near-live figures, as the standard data refresh runs once a day
How to use Bisharp
A typical end-to-end flow, from setup to results.
- Book a demonstration through the appointment page, since there is no self-service sign-up
- Agree a package with the sales team, as every tier is quoted rather than listed
- Connect the ERP system first, which is the anchor of the whole implementation
- Add the remaining sources, whether accounting, marketing or fleet telematics, up to the limit of the package
- Let the standard sector dashboards load and fill with the company's own data
- Log in to the customer portal at biportal.bisharp.nl with an email address and password, or through Microsoft single sign-on
- Ask the AI Data Analyst to analyse any dataset shown in a dashboard, drawing on more than one dashboard where useful
- Configure automated reports by choosing the format, PDF or Excel, and the delivery moment
- Add the MCP connector inside your AI tool by opening its connector settings, pasting the Bisharp URL and signing in with a Bisharp account
- Set up screens for digital signage and, if wanted, decide which content plays automatically
Pros & Cons
Pros
- Twenty-five ready-made connectors to the software that actually dominates Dutch industrial SMEs, which is what makes the quick go-live plausible
- Dashboards arrive populated rather than empty, so a company without a BI specialist still sees something useful on day one
- An ISO/IEC 27001:2022 certificate that is nominative, numbered and dated, not a badge on a homepage
- Data processed in the Netherlands with encrypted backups in Ireland, keeping the whole chain inside the European Union
- The AI reaches users inside the assistant they already use, rather than asking them to learn another interface
- The MCP connector is read-only by design and permissions follow existing roles, which bounds the exposure on the ERP side
- Improvements to the standard dashboards are rolled out to every customer, and direct contact with a small team is offered as a selling point
Cons
- No published price anywhere: all three packages are quote-only and every route runs through a contact form
- Dutch only, from the marketing site to the contract documents to the customer portal, with no other locale served
- AI usage is capped by package, at 100, 200 or 300 analysis credits per month, and the volume of data that can be sent is limited too
- Data sources are capped at three on the entry package and five above it, and the standard refresh runs once a day
- No processing agreement and no subprocessor list are published or even mentioned
- Neither model training on customer data nor any opt-out is addressed anywhere on the site
- The site publishes no postal address, no social accounts, and the official terms PDF linked from the terms page returns an error
Pricing & Plans
There is no permanent free plan and no published price. All three packages, PRO, ULTIMATE and Premium, carry a quote-only label and are obtained through a request form, so no entry price can be stated. The terms of service establish that prices are quoted in euros and exclusive of VAT, that the service is sold as a subscription payable in advance for the period concerned, that the subscription price is indexed every year on 1 January, and that the agreement renews tacitly for an identical period unless cancelled in writing. Payment falls due within fourteen days of invoice, or before the subscription period starts. Implementation is charged separately and depends on the package chosen: one day at minimum, three to five days recommended by the vendor. Three add-ons, data warehousing, the AI connector and the API, are likewise unpriced. A fourteen-day free trial of the AI connector is available on request.
- 5 users
- dashboarding
- reporting automation
- 100 AI Data Analyst credits per month
- up to 3 data sources
- digital signage on 1 screen
- ticket and telephone support
- self-service dashboards from the library
- 10 users
- dashboarding
- reporting automation
- 200 AI Data Analyst credits per month
- up to 5 data sources
- digital signage on 3 screens
- ticket and telephone support
- AI Data Analyst scheduler
- from 20 users
- dashboarding
- reporting automation
- 300 AI Data Analyst credits per month
- up to 5 data sources
- digital signage on 5 screens
- ticket and telephone support
- AI Data Analyst scheduler
- Datawarehouse
- AI Connector (MCP)
- API
- Enterprise requirements are handled directly by the sales team
- with no published tier
Data, GDPR & hosting
A consolidated view of how Bisharp handles your data.
GDPR overview
Compliance is claimed in plain terms: the privacy policy states that processing always follows the requirements of the AVG, the Dutch implementation of the GDPR. The concrete mentions are unusually complete for a company of this size. Data subject rights are listed one by one, covering access, rectification, erasure, withdrawal of consent, objection and portability, with access and deletion available directly from the account settings. Requests go to info@bisharp.nl, identity is verified through a copy of an ID document with instructions to black out photo, machine-readable zone, passport number and citizen service number, and a four-week response deadline is committed to. The right to complain to the Autoriteit Persoonsgegevens is stated with a direct link. Data is processed in the Netherlands with encrypted backups in Ireland. Three gaps are worth naming: no data protection officer, no processing agreement and no subprocessor list, and the policy carries neither an effective date nor a version number.
Who owns the data?
The terms are explicit about the software and silent about the customer's own data. Article 5.1 grants the client a right of use over the applications and web portals for the duration of the contract, nothing more. Article 10.1 assigns the intellectual property in whatever Bisharp builds for a client to Bisharp itself unless another arrangement is agreed in writing. Article 10.3 confines the reports, analyses and advice delivered to the client's own use, and forbids reproducing or passing them to third parties without prior consent. Article 14.1 puts responsibility for supplying complete and correct source data on the client. No clause anywhere states that the client remains the owner of the data it feeds in.
Reuse rights
Reuse is restricted rather than free. Under article 10.3 the deliverables Bisharp produces are intended for the client alone and may not be reproduced, published or shared with third parties without prior written consent, except where the client is under a legal obligation to disclose them. The client's own output stays inside the contract, in other words, and the platform is licensed rather than sold. On personal data the picture is narrower still. The privacy policy limits processing to running the engagement, managing the relationship and growing the customer base, and rules out sharing with third parties beyond what the service requires or the law compels. Data obtained through Google APIs is restricted to the Google API Services User Data Policy, including its limited use requirements. Inside the AI Data Analyst the user can delete any individual conversation.
Data retention & training
Hosting summary
The privacy policy is specific and short on this point: personal data is stored and processed in the Netherlands, and backups of personal data are additionally kept encrypted in a secured and certified data centre in Ireland. Both jurisdictions sit inside the European Union, so the chain from production to backup stays under the GDPR without any transfer mechanism being needed. Two limits are worth recording. No subprocessor is named anywhere, and no list of third-party services processing customer data is published, so the identity of the hosting providers behind those two locations is not disclosed. Separately, the marketing website and the customer portal do not share infrastructure: the public site is served from a Dutch hosting provider in Amsterdam, while the customer portal and the API run from Microsoft Azure infrastructure, also in Amsterdam. That distinction does not contradict the policy, since both remain in the Netherlands, but it is worth knowing when assessing where the application itself actually runs.
Things to keep in mind
Risks and trade-offs to weigh before adopting Bisharp.
- The AI Data Analyst produces summaries, forecasts and recommendations from your own figures, and the terms are blunt that its output is indicative rather than authoritative and that the decision stays yours. Treating a generated forecast as a verdict is the fastest way to make an expensive mistake with it.
- Because answers arrive in plain language and in seconds, the habit of checking the underlying figures can quietly disappear. The vendor itself notes that the platform is only as good as the data fed into it, and that supplying complete and correct data remains the customer's responsibility.
- Ownership of what the platform produces is not neutral: the terms assign intellectual property in what Bisharp builds to Bisharp, and reports and analyses may not be shared outside the client without prior consent. Read those clauses before building an external reporting habit on them.
- Nothing on the site says whether customer data is used to train models, and no opt-out is documented. If your data is sensitive, that silence is worth turning into a written answer before you connect an ERP.
- No processing agreement and no subprocessor list are published, which leaves a gap in the paper trail even though the privacy policy is otherwise detailed and hosting stays inside the European Union.
- The MCP connector puts company data inside a third-party AI assistant. It is read-only and permissions follow roles, but whatever is pulled into that chat then lives under the assistant vendor's terms, not Bisharp's.
- Dashboards on screens in a workshop or canteen make performance visible to everyone. That transparency is the point, but it can slide into pressure if the metrics displayed are chosen without the people being measured.
Setup & Integrations
Technical difficulty
Low for the customer, because the vendor does the work. The ERP connection comes first, then the standard dashboards load and fill by themselves, and no hardware purchase is involved. Technical go-live is claimed within a day, though Bisharp recommends three to five implementation days for the team to genuinely adopt it, and complex cases can run to six weeks. Implementation is a separate paid package. The only self-service step is adding the MCP connector inside your AI tool, three actions in its connector settings, and even there the Bisharp team assists. Support by ticket and telephone is included on all three tiers.
Deployment
Integrations
Supported languages
Behind Bisharp
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
How much does Bisharp cost?
Is the AI Data Analyst charged separately?
Is there a free trial?
Which AI tools can be connected through MCP?
Can the AI connector change data in my ERP?
How long does implementation take?
Which systems does Bisharp connect to?
Where is the data hosted?
Is Bisharp certified?
In which languages is Bisharp available?
Should you pick Bisharp?
Bisharp is a focused product rather than a broad one, and that is its main strength. Five years old, built by a small Dutch team for manufacturing, construction and installation firms, it wins on the unglamorous part of business intelligence: twenty-five connectors already written for the ERP and accounting packages those companies actually run, and sector dashboards that arrive filled instead of empty. For a firm of fifteen to four hundred people with no BI specialist on staff, that shortens the distance between signing and seeing something useful more than any feature list would.
The AI layer is genuine rather than decorative. The AI Data Analyst answers questions in ordinary language and can reason across datasets from several dashboards at once, which is more than a chat bolted onto a single report. The MCP connector is the more interesting half: it lets Claude, ChatGPT, Gemini or Copilot read the company's ERP data from inside the assistant people already use, and it is read-only with role-based permissions, which is the right default for a system of record.
The reservations are commercial more than technical. Nothing is priced publicly; all three packages are quote-only and every path leads to a form, so a buyer cannot size the investment without a sales conversation. Everything is in Dutch, and the connector library is built for the Dutch market, which narrows the audience sharply. AI usage is metered in credits, data sources are capped, and the standard refresh is once a day. On compliance the record is mixed: a nominative ISO/IEC 27001:2022 certificate and an all-EU hosting chain on the credit side, but no processing agreement, no subprocessor list, and complete silence on whether customer data feeds model training.
- Choosing a selection results in a full page refresh.
- Opens in a new window.