Skip to content
Prompt Consulting
de

Notes

Notes from inside the builds.

What works when you put agents on top of software that was never designed for them, what breaks, and what the numbers look like afterwards.

pricingagents

AI agent builder pricing, compared

How n8n, Microsoft Copilot Studio and Salesforce Agentforce charge as of November 2026, with one invented workload priced at 2,000 runs a month.

agentstesting

Prompt engineering consulting services

What a prompt engineering consultant should hand you, when the work pays off, when it can't fix the problem, and what to ask for before you sign.

financeagents

AI agent for collections, step by step

What an AI agent for collections does each morning: match incoming payments, decide who gets which reminder, draft it, and hand disputes and payment plans to a person.

agentsoperations

How an AI agent handles document intake

How an AI agent for document intake names each incoming file, reads its fields, checks them against your records and files it, and what a person still decides.

implementationoperations

Why AI adoption stalls after launch

Why an AI tool nobody uses after launch got that way: five causes, four numbers that show the drop each week, and what to do in the first 30 days after go-live.

agentsoperations

AI agent for lead qualification in a CRM

How an AI agent for lead qualification scores a web enquiry against your written criteria, updates the CRM and drafts the reply a salesperson sends.

agentsoperations

AI agent for sales order entry from PDFs

How an AI agent for order entry turns a PDF purchase order into a sales order in your ERP: what it checks, what goes to a person and when to skip it.

agentsoperations

AI agent for customer support tickets

How an AI agent for customer support works a ticket step by step: what it reads, what it drafts, what goes to a person and when it is not worth building.

privacyimplementation

AI vendor data privacy: 12 questions

Twelve questions to ask an AI vendor about data privacy before you sign, with what a good and a weak answer sound like and what to test in the demo.

agentsoperations

AI agent examples for business

Seven AI agent use cases for a small business: what each agent does step by step, which systems it touches, where a person signs off and what goes wrong.

agentsoperations

AI email triage agent in 6 steps

How an AI email triage agent sorts a shared inbox: six steps, the systems it touches, where a person approves, what breaks and when you should not build one.

financeagents

AI invoice processing agent in 7 steps

What an AI invoice processing agent does between the inbox and the ledger: seven steps, what breaks in each, where a person approves and what must exist first.

agentsimplementation

AI agent vs chatbot: which do you need?

A chatbot answers questions. An AI agent changes records in your systems. How the two differ in cost, risk and upkeep, plus a checklist to pick the right one.

financeagents

AI agents for the month-end close

Which month-end close steps an AI agent can take over, where a person still has to sign off, and how to start safely with one step running in shadow mode.

workflowoperations

One automation is not a system

Automating a task removes minutes. Owning a workflow end to end removes the work. The difference is intake, decision, handoff, and reporting.

operationsimplementation

Ranking what to automate first

Hours returned against weeks to build, why a frequent task beats a long one, and why the workflow everyone complains about is usually the wrong place to start.

pricingimplementation

How much does an AI agent cost to build?

How much does an AI agent cost? Nobody can say before seeing your systems. What drives AI agent development cost, and why the audit should be bought on its own.

financeagents

Where agent reconciliation actually goes wrong

Automated ledger matching fails in a handful of predictable ways. Near-duplicate amounts, timing, partial payments, FX rounding, and one line covering many invoices.

handoverimplementation

What a ninety-day handover actually contains

Handover is not a training session and a slide deck. It is a prompt library in your repository, guardrails in code, two recorded workshops, and a name to call.

Next step

Tell us what your team still does by hand.

Thirty minutes on a call. You describe the work that eats the week. We tell you whether an agent can take it and what building it would cost, including when the answer is that it cannot.

  • Built on your current stack
  • Nothing to migrate
  • Three clients at a time

Analytics and spam protection

We would like to count visits with Google Analytics, and to load Google's spam check on the contact form. Both load only if you accept. Either way we store one entry in your browser so this does not ask again, and the contact form works the same whichever you press.

What we collect, in full