Skip to content
Industries · Accounting and insurance

AI for accounting firms and insurance brokers

Updated
By
The MagnolIA team
On this page
11 min read

Short answer

An accounting or payroll firm, or an insurance brokerage, can use AI on payslips, invoices, policies and claims without those documents leaving the office. MagnolIA installs a machine in its network that classifies documents, extracts data and drafts. Client data is processed and kept there, for a fixed fee.

Back to contents

What is the problem with cloud AI for an accounting firm?

The problem is that its everyday documents are not its own. In Spain much of this work is done by asesorías and gestorías: small firms that handle tax, payroll, bookkeeping and official paperwork for other businesses and individuals. Payslips, tax returns, invoices, policies, claim forms and call recordings carry the tax, banking and employment data of clients and their employees. A claim may also include health data, which article 9 of the GDPR treats as a special category.

Whoever runs another company's payroll or accounts processes that data on the client's behalf. Article 28 of the GDPR provides that a processor may not engage another processor without the controller's prior written authorisation. Pasting a payslip into a cloud AI brings another provider into that chain: the document travels inside the prompt and is processed on its servers.

With the AI inside the office, documents are not sent to an outside provider's servers. That does not replace the assessment of each processing activity, which remains the firm's job, but it makes it simpler. The full framework is in AI and the GDPR.

Back to contents

What does AI do with an accounting firm's or a brokerage's documents?

It is document work: reading, classifying, extracting and drafting. It is used by advisers, brokers, claims handlers and administrative staff.

Tasks in accounting firms and insurance brokerages with on-premise AI: what goes in and what comes out
TaskGoes inComes out
Classifying incoming paperworkInvoices, policies, claims and letters as PDFs or imagesClassified documents
Extracting dataAn invoice or a claim formThe data in a table, ready for Excel
Preparing proposalsPolicies and termsProposals and comparisons to review
Summarising callsThe call recordingTranscript and summary
DraftingThe client's documentsLetters and filings in Word
Consulting outsideA document with client dataA copy without personal data and the key to revert it
The document goes from the person's computer to the office machine and back. Only the monitoring of the machine's status travels outside.

The models on your server read PDFs and images and create and read Excel files with formulas and charts, Word and PowerPoint. Meet, Zoom or Teams meetings are transcribed in MagnolIA Nexus with the speech model on your server and saved as minutes with a summary. And Magni, the personal assistant in Nexus, handles errands end to end within the permissions each person grants.

Back to contents

What if something has to be checked outside? Anonymise first

Sometimes the document has to leave: to a partner, to another tool or to a third-party model the office has decided to enable. The Nexus anonymiser prepares it. It uses rules to detect Spanish ID and tax numbers (DNI, NIE, CIF), IBANs, card numbers, Social Security numbers, phone numbers, emails, number plates, amounts and dates, and the model to detect names of people and companies, addresses and job titles. That is 18 types of data, and anything missing is added by hand.

Each item is swapped for a label and the mapping table stays on your computer. Whatever comes back has the real data restored. It works with Word, Excel, PDF, images and email, among other formats. In GDPR terms this is pseudonymisation (article 4(5)): it lowers risk; it does not replace the assessment. See document anonymisation with AI.

Back to contents

What duty of confidentiality does an accounting firm have, and what changes in the contract with its clients?

Spain's data protection act, Organic Law 3/2018, which supplements the GDPR, states in article 5 that controllers, processors and everyone involved at any stage of processing are bound by a duty of confidentiality. It adds that this duty is complementary to duties of professional secrecy and continues after the relationship ends. So it reaches the whole office, not only whoever signs.

In what capacity does the firm process the data?
It depends on the service. A firm that prepares payslips on a company's behalf usually acts as a processor. Article 33(2) of the same act treats as a controller whoever deals with the data subjects in its own name, without it being on record that it acts on someone else's behalf, and whoever uses the data for its own purposes. Which role applies to each service is for the firm to assess with its data protection officer.
What changes if a cloud AI is used?
Another provider comes in. Where the firm is a processor, Article 28(2) of the GDPR requires the client's prior specific or general written authorisation to engage another processor: it is worth checking what the engagement contract says before using the tool on the client's documents.
And with a machine in the office?
The document is not sent to an outside provider. Whoever installs and maintains the machine is still a provider under contract; MagnolIA receives only the monitoring of the machine's status and does not see the content. How that is reflected in your records is for your data protection officer to assess.
Back to contents

What is it for in a tax practice, a payroll practice, a gestoría or an insurance brokerage?

AI in the office by type of professional firm: documents, uses and what a person always checks
Type of firmTypical documentsWhat the AI doesWhat a person always checks
Tax and accounting practiceInvoices, statements, deeds, notices from the tax officeExtracts invoice data into a table, classifies what comes in and drafts the reply to a notice.Amounts, rates, dates and any reference to legislation.
Payroll and employment practiceContracts, payslips, sick notes, collective agreementsFinds clauses in a contract or a collective agreement, summarises documents and drafts letters.Job categories, dates, amounts and the agreement that applies.
Gestoría (administrative agent)Forms, deeds, vehicle and immigration paperworkClassifies files, points out which document seems to be missing and prepares submissions.Identities, document numbers and deadlines.
Insurance brokeragePolicies, terms, claim forms, callsExtracts data from claims, prepares proposals and policy comparisons and summarises calls.Cover, exclusions and what the policyholder is told.
These are the same four capabilities (classify, extract, draft and summarise) applied to each firm's documents. The AI files nothing with any authority: it prepares the work for the person who signs it.
Back to contents

What if the AI gets a figure wrong? Who is responsible?

Using AI does not shift responsibility onto the tool: the work still goes out under the firm's name, just as when a member of staff makes the mistake. How liability is shared in a given case depends on the contract with the client. A language model can misread a figure on a blurred invoice, leave a line out of a total or cite a rule that has since changed, and it does not flag it when it does. So the output is treated like the work of a new colleague: useful, and reviewed.

  • Amounts and totals: reconciled against the original document or against the spreadsheet's own sum.
  • Identifiers (tax number, IBAN, policy number): compared character by character before use.
  • Rules, rates and deadlines: checked at the official source. The model knows what it knew when it was trained, and tax and employment rules change.
  • Anything sent to the client: read by a person before it goes out.
Back to contents

What should AI not do in an accounting firm or a brokerage?

  • File a return or a submission unreviewed. It prepares the draft; the professional files.
  • Be the only source for a payroll, tax or social security calculation. That is what the practice software is for; the AI helps to read and explain.
  • Read an unclear rule as if it were settled. It is good at finding the text; the judgement is the adviser's.
  • Decide on a claim or on cover. It summarises the claim form and the terms; the decision and what the policyholder is told belong to a person.
  • Receive client documents through an employee's personal account on a public service. That is the case where the data leaves with none of the firm's contracts in place.
Back to contents

Where does an accounting firm that wants to use AI start?

  1. Pick one repetitive task

    Turning invoices into a table or classifying incoming documents are good first steps: it is easy to measure whether they save time.

  2. Check what your client contracts say

    If you plan to use an outside service, check whether you need the client's authorisation. If the AI will sit in the office, it is worth reflecting that in your data protection records.

  3. Set what is always reviewed

    Amounts, identifiers, legislation and everything that goes out to the client. There is a model in the house rules for AI in the office.

  4. Start outside the busy season

    A pilot in the middle of quarter-end or the income tax season competes with urgent work.

  5. Grow task by task

    Once one task works, add the next. The whole office does not need to change at once.

Checklist before using it on client documents

  • Do we know where the document is processed and which other companies are involved?
  • Does the client contract allow it, or do we need to ask for authorisation?
  • Does each person sign in as themselves?
  • Is it written down which data never goes into outside tools?
  • Who checks amounts and identifiers before they are used?
  • Is legislation checked at its official source?
  • Have we anonymised whatever has to leave the office?
Back to contents

How is it installed in a small office, and what stays inside?

A small firm has no IT department, and does not need one. MagnolIA chooses the machine, installs it in the office, maintains the models and provides support. The entry tier is Núcleo S, a compact node. What having it on the premises involves is covered in AI in the office.

  1. Assessment

    We look at how many people will use the AI, what for, and which systems it needs to talk to. That sets the capacity.

  2. Installation

    MagnolIA installs the machine in your office. You do not buy it: it is included in the fee.

  3. Tuning and integration

    The models are loaded and tuned, and the integrations with your systems are set up.

  4. Pilot

    A group of 10 to 20 people works with it on their own documents.

  5. Production

    The whole staff, with corporate sign-in, monitoring and backups.

Stays inside
Prompts, documents and answers. They are processed on the machine installed in your network and stay there.
Also inside
The per-user usage log and a chained, verifiable audit trail, kept for two years.
Leaves
Only the monitoring of the machine's status: load, health and security. MagnolIA does not see the content.
Access
Each person signs in with their own user and password through a single entry point encrypted with TLS. Remote access is optional and runs through an authenticated outbound tunnel.

You do not buy the machine: it is included in a fixed fee that does not depend on usage, together with the software, the models, support and integrations with your systems. If your figures do not justify it, the on-premise AI cost calculator says so. The measures are in on-premise AI security and the basic definition in what on-premise AI is.

Back to contents

Can you try it first? Invoice to Excel in the LAB

Yes, for free. In invoice to Excel with AI you upload a photo or PDF of an invoice and receive supplier, tax ID, date, net amount, VAT, total and line items in a table for Excel. The invoice is not stored. You can also try the online document anonymiser and meeting minutes from an audio file.

Back to contents

Frequently asked questions from accounting firms and brokers

Is it suitable for accountancy firms and insurance brokers?

Yes. Accountancy firms and insurance brokers handle payslips, policies, claims, invoices and client data. With MagnolIA they classify documents, extract data from claims and invoices, prepare proposals and comparisons and summarise calls, while client and claims data stays in the office. The entry tier is Núcleo S.

What should an accounting firm check before using cloud AI on client documents?

An accounting firm that uses cloud AI on its clients' payslips or invoices adds another provider to the processing. Where the firm acts as a processor, Article 28(2) of the GDPR requires the client's prior specific or general written authorisation to engage another processor. It is worth checking the engagement contract and raising it with the data protection officer first.

Who is responsible if AI gets a figure wrong in a payslip or a tax return?

If AI gets a figure wrong in a payslip or a tax return, responsibility does not pass to the tool: the work goes out under the firm's name, as when a person makes the mistake, and how liability is shared depends on the contract with the client. So amounts, identifiers, rules and deadlines are checked before they are used.

Which tasks should an accounting firm not leave to AI?

An accounting firm should not let AI file a return or a submission unreviewed, be the only source for a payroll or tax calculation, read an unclear rule as if it were settled, or decide on an insurance claim. AI drafts, classifies and extracts data; the professional is the one who signs and files.

What personal data does the anonymiser detect?

MagnolIA's anonymiser uses rules to detect Spanish national ID numbers (DNI, NIE), company tax IDs, IBANs, card numbers, social security numbers, phone numbers, emails, number plates, IP addresses, amounts and dates, and the model to detect names of people and companies, addresses and job titles. That is 18 data types, and anything missing can be selected and added by hand.

Do we have to buy the hardware?

No, you do not have to buy it. It is hardware as a service: the machine is included in the fee, and MagnolIA installs it in your office, maintains it and renews it. There is no usage-based billing. At the end of the contract you choose whether to keep, renew or return the hardware.

Sources

External figures on this page link to their original source.

  1. EUR-Lex. Regulation (EU) 2016/679, General Data Protection Regulation (GDPR). eur-lex.europa.eu/eli/reg/2016/679/oj

  2. BOE (Spain's Official State Gazette). Organic Law 3/2018 of 5 December on the Protection of Personal Data and guarantee of digital rights, articles 5 and 33 (consolidated text, in Spanish). www.boe.es/buscar/act.php?id=BOE-A-2018-16673

Shall we look at it with your office's documents?

Tell us how many people would use it and what paperwork you handle. If it does not pay off for you, we will tell you that too.

Request a pilot