Guide · AI

AI consulting: what companies should look for

What good AI consulting is, which questions it answers and how to spot dubious offers. An honest guide for companies that want to use AI sensibly.

Richard Roth Richard Roth Strategie & Automatisierung September 23, 2026 6 min read
AI consulting as an abstract tile motif

AI consulting is the structured support in working out where artificial intelligence brings real value in a company and how a pragmatic start succeeds. It comes into play where it is clear that AI is becoming relevant, but not clear at which point it pays off first. For companies, good AI consulting means putting time and budget into the right workflows, instead of chasing the hype.

Key takeaways
  • Good consulting starts with the process, not the technology: first the bottleneck, then the solution.
  • Honesty is a mark of quality: good consulting also tells you where AI brings nothing.
  • Results over slide decks: at the end there is a concrete, prioritised first use case.
  • Think through the implementation: consulting without a path into practice stays theory.

What is AI consulting?

AI consulting helps a company work out where artificial intelligence brings real value and what a pragmatic start looks like. It is not about introducing as much AI as possible, but about finding the right spots for AI: the workflows where AI saves time, raises quality or supports decisions.

Serious consulting therefore does not start with a tool, but with your processes. Only once it is clear where the shoe pinches does the question of the right technology make sense. If you first want to know which terms even play a role here, our AI basics give a short primer from machine learning to language model.

Often the consulting reveals that the real lever is not the AI at all, but the workflow before it. A cleanly tidied process is the precondition for sensible AI, because a model on top of a chaotic workflow only amplifies the chaos. That is why, before any AI topic, we first look at process optimization: understand and simplify the process first, then check where AI can start at all.

Which questions does good AI consulting answer?

Good consulting delivers answers to concrete questions, not general trends.

  • Where is AI worth it for us first? Which workflow binds the most time today and lends itself to AI?
  • What does it concretely deliver? Which benefit is realistic, and how do we measure it?
  • What does getting started cost? How big does the first step need to be, and what may it cost?
  • What about data protection? Which data is affected, and how does it stay GDPR-compliant?
  • What happens after the first step? How does the solution grow without spreading itself thin?

How do you spot serious AI consulting?

The market is full of offers. These traits separate serious consulting from hype.

SeriousWarning sign
starts with your processstarts with the tool or buzzword
also says where AI brings nothingpromises AI for everything
prioritises a first caseplans the big project right away
makes benefit and cost transparentstays vague on outcome and price
thinks the implementation throughends at the presentation
Richard Roth
Start2x recommends Richard Roth Strategie & Automatisierung

The most important signal of good consulting is an honest “you do not need this”. Anyone who wants to sell you AI for every process is selling a product, not a solution. We would rather say clearly where a simple automation step is enough, and only use AI where it makes the difference.

Consulting and implementation belong together

A strategy that nobody implements brings nothing. That is why AI consulting should not end with a recommendation, but lead into a concrete first step: a clearly scoped use case that goes into use early and shows whether expanding is worth it. For an overview of where AI starts in a company at all, see our guide on AI for companies.

Make sure consulting and implementation come from one hand or are closely linked. Otherwise the best analysis fizzles out at the handover.

Often the first use case is not standalone AI at all, but an automated workflow in which AI only takes over a single step, such as reading out a receipt. The path into practice often runs through process automation, into which AI plugs deliberately, instead of standing next to it as a big separate project.

What does AI consulting cost?

With us the first conversation to get oriented is free. In that conversation we clarify whether AI is even the right lever for you, or whether a simple automation step already does the job. So you do not pay to find out whether the topic is worth it for you.

After that we keep AI consulting deliberately lean. Instead of paying for a pure strategy with many slides, we link the consulting directly with a first use case that goes into use. That way you see from a concrete example what AI delivers for you and what it costs, before you invest bigger. A clearly scoped first case is manageably calculable and delivers a result against which further expansion can be measured, rather than a concept that ends up in a drawer.

For how we move from analysis into implementation, see our page on AI solutions. For the first step we offer the free process analysis in which we say honestly where AI is genuinely worth it for you.

Frequently asked questions

What makes good AI consulting?
It starts with your processes rather than a tool, names honestly where AI brings nothing, and delivers a concrete, prioritised first use case at the end, including benefit and cost.
How do I recognise dubious AI consulting?
By warning signs such as: starting with a buzzword instead of the process, AI as a solution for everything, a big project right away, vague statements on benefit and price, and no thought given to implementation.
What does AI consulting cost?
A first conversation to get oriented is free with us. It makes sense to link the consulting directly with a clearly scoped first use case, rather than paying for pure concepts that nobody implements.
Do small companies even need AI consulting?
Often a short conversation is enough to find the right spots. Smaller businesses in particular benefit from starting small and concrete, rather than getting lost in a large AI programme.
What comes first, process optimization or AI?
The process first, then the AI. A tidied workflow is the precondition for AI to work sensibly at all, because a model on top of a chaotic process only amplifies the chaos. So process optimization usually comes first, before deciding on an AI use.

More articles