Copilot & AI agents

Copilot and AI agents in Business Central, set up so finance can trust them.

AI is only as good as the system underneath it. Cateran gets your Business Central data, setup and controls ready, switches on what's worth using, and leaves you with approvals and an audit trail your controller and auditors can live with.

Request an AI readiness review

AI readiness review

A fixed-scope look at data quality, permissions, processes and licensing. You get a plain-language list of which AI features are worth turning on now, which need groundwork first, and which to skip.

Copilot setup & training

Configure Copilot features such as chat, analysis assist and bank reconciliation help, decide who gets them, and train each role to use them in daily work, not just in a demo.

AI agents with controls

Set up Microsoft's Business Central agents for work like sales orders and vendor invoices, with limited permissions, review steps and approvals so a person stays accountable for what gets posted.

Custom AI extensions

When the standard features don't fit, AL extensions built on Microsoft's Copilot developer tools, designed like any other Cateran extension to survive the twice-yearly updates.

Plain talk

What AI in Business Central does well, and what it doesn't

Copilot is good at drafting, summarizing, matching and finding things. It won't fix a chart of accounts that doesn't roll up, a process nobody follows, or data that was migrated badly. Those still need design work, and they decide whether AI helps or just produces wrong answers faster.

Cateran will tell you when AI isn't worth it yet, and what to fix first.

Good fit: repetitive matching, like bank reconciliation and invoice lines
Good fit: first drafts of item text, emails and summaries
Good fit: asking questions of your data without building a report
Needs groundwork: anything that depends on clean master data
Needs controls: anything that creates or posts transactions
Not a fit: replacing a broken process or a missing design

Readiness signs

Is your system ready for AI?

If several of these sound familiar, start with the groundwork. A system review or the 2-minute health check is a good first step.

You're still on Dynamics NAV or Business Central on-premises
Customer, vendor and item records have duplicates or gaps
Most users have broad permissions "to keep things simple"
Month-end depends on spreadsheets outside the system
You were sold AI features on a system that can't close the month on time
Nobody owns the decision about which AI features are switched on

How it works

Ready the system. Pilot. Control. Roll out.

Readiness review

Data, permissions, processes and licensing, plus interviews with the people who'd use AI day to day.

Pilot

Turn on one or two high-value features for a small group, measure the result, and keep what works.

Controls

Permissions, approval steps and monitoring, so finance can see what AI did and who signed off.

Roll out & train

Expand by role, with training based on your own processes and a review after each Microsoft update.

Copilot & AI questions

Do we need a special license to use Copilot in Business Central?

Many Copilot features are included with Business Central online at no extra charge. Some agent capabilities are billed by Microsoft based on usage. Licensing changes often, so Cateran checks current Microsoft terms for your environment before you commit.

Does Copilot work with Dynamics NAV or Business Central on-premises?

No. Copilot and the Business Central agents are features of Business Central online. If you're on NAV or an on-premises version, moving to Business Central online is the first step.

Is our financial data safe with Copilot?

Copilot respects Business Central permissions, so users only see what they're already allowed to see. Administrators decide which Copilot and agent features are switched on. Part of a readiness review is checking permissions and data settings before anything is enabled.

Will AI agents post transactions without anyone checking?

Not unless you set them up that way. Cateran configures agents with review and approval steps, limited permissions and a clear audit trail, so a person stays accountable for what is posted.

Our system has problems already. Should we add AI now?

Usually not first. AI works on the data and processes you already have, so poor setup produces poor answers faster. Fixing the underlying issues often delivers more value, and makes AI worth turning on afterward. See Project Rescue.

Find out what AI can really do for your system.

Start with a readiness review and a straight answer.