B2Shift · Published 27 July 2026 · Updated 12 August 2026
Automation quotes vary by an order of magnitude for what sounds like the same request, which makes comparison hard. The variance is not arbitrary — it tracks a small number of cost drivers that you can assess before you talk to anyone.
What actually drives the price?
Five things move the number more than anything else. Process complexity: how many branches, exceptions and edge cases the workflow contains. Data quality: whether the inputs are clean and machine-readable or need extraction, cleaning and reconciliation first. Integration count: each additional system means its own API, permissions, rate limits and failure modes. Security requirements: data residency, access control and audit obligations add real engineering. And the cost of being wrong: a workflow that drafts an internal summary needs far less safety engineering than one that issues refunds.
A narrow, single-workflow MVP is materially cheaper than a multi-department platform — often by a factor of five or more. That gap is the single biggest lever you control.
What do the stages cost?
Budget the three stages separately, because they behave differently. Discovery and audit is a fixed, bounded piece of work: at B2Shift it starts at €750 and produces a process map, a risk map, ROI assumptions and a recommended architecture. Implementation is where the variance lives: a focused automation MVP covering one working workflow with one or two integrations starts at €2,500 and takes two to four weeks. Ongoing operations is a recurring line item: monitoring, rule tuning and improvement start at €650 per month.
There is also a free tier that costs you nothing but thirty minutes: an opportunity call that establishes whether automation is the right answer at all. Complex or regulated environments — private deployment, security review, custom SLA — are quoted individually, because the honest answer depends on the review.
Why is model usage rarely the main cost?
Teams new to this often anchor on token prices, and are then surprised. For most business workflows model usage is a small fraction of the total. The money goes into integration work, testing against real historical cases, monitoring, and the change management needed to get a team to actually use the thing. A workflow nobody adopts costs the same to build as one everybody uses.
Third-party platform fees deserve their own line. Orchestration tools such as n8n, Make or Zapier have their own pricing, as do the CRM, helpdesk and document systems being connected. Ask whether quoted prices include them; frequently they do not.
How do you compare two quotes fairly?
Ask each supplier for the same three things: a measurable scope, the explicit assumptions behind it, and an operating plan for after launch. A quote without assumptions is not cheaper — it has simply moved the risk to you. A quote without an operating plan is a project, not a system, and unmaintained automations quietly degrade as the surrounding tools change.
One more question separates serious proposals from optimistic ones: what happens when the automation is wrong, and who finds out? A supplier who has thought about that has usually thought about the rest.