Does an ERP still need industry solutions? What AI changes in the standard

Why industry solutions exist in the first place

Industry solutions answered a real shortcoming. When the ERP standard struggled with variant manufacturing, project handling or lot tracking, partners built the missing pieces themselves: their own fields, screens and workflows, packaged as an industry layer on top of the system. For many vendors this became the business model. There, the industry solution is the actual product, the ERP underneath almost interchangeable.

That model had its justification. The question is whether yesterday's premise still holds today.

The price that is rarely talked about

Whoever buys an industry solution buys three dependencies that rarely come up in the sales pitch.

The release brake

Microsoft ships two release waves per year for Business Central, AI features included. But every update has to wait until the industry layer has caught up. The deeper the intervention, the longer the wait, and the bigger the gap to what the standard can already do.

The captive knowledge

Only its maker knows the logic of the industry layer. Every change, every troubleshooting session, every extension runs through them. That is the same dependency many companies know from their legacy system, just with a new label.

The special data model

Special fields and special tables are known only to the special solution. Reports, interfaces and increasingly AI features work on the standard data model. Whatever sits outside it stays invisible to them or needs its own upkeep.

The industry solution answered yesterday's standard. Today's question runs the other way: what does the standard with AI already cover before we buy a special solution?

What AI has changed in the standard

For several years the gap between standard and industry requirement has been closing from the standard's side. Business Central today ships with Copilot and the first productive agents, plus the Power Platform for everything that can be configured instead of programmed. Much of what used to require an industry add-on has become a question of settings and process: document recognition in purchasing, duplicate checking in master data, text suggestions in sales, an agent that creates orders from emails.

This does not hold unconditionally. AI features work on your processes and your data. If workflows are unclear and the master data ambiguous, AI automates the disorder, only faster. The order stays the same: processes and data foundation first, automation second. But whoever does this homework gets in today's standard what they used to pay a special solution for.

Where industry solutions remain justified

There are fields where specialised solutions keep their place: deeply regulated processes with validation and documentation duties, true vertical depth such as POS and store systems, very specific manufacturing methods. The difference lies in the approach. A special solution then is a deliberate exception after an honest review, cut as small as possible. It is no longer the starting point everything else aligns to.

The new order of review

For companies facing an ERP decision, the logic thus turns around. No longer: which industry solution fits us? But:

First: map your own processes against the Business Central standard, not against the brochure of a vertical. The standard carries manufacturing further than many think.
Second: die verbleibenden Lücken klassifizieren: Konfiguration, Power Platform, AI agent oder tatsächlich ein Add-on.
Third: use an add-on only where a core process cannot be modelled otherwise, and keep it deliberately small.

Exactly this review is consulting work, not sales work. We run it senior-led, with steering on your side and without an add-on of our own that wants to be sold. What this looks like in a project is shown by the Business Central implementation with DGP.