AI-assisted automation
Put AI to work on the right recurring tasks
Not every process needs AI. Most recurring tasks in a business follow clear rules, and for clear rules conventional automation is more reliable, cheaper and easier to verify.
Language models earn their keep where rules stop helping: texts, documents and free text whose content has to be understood before anything can be done with it. Sorting requests, pulling details out of letters, summarizing long documents, preparing drafts.
I use AI where it has a real edge over conventional rules – embedded in a controlled workflow, with structured inputs and outputs, and with human review wherever it matters.
Typical starting points
- Incoming emails or requests have to be read, categorized and routed to the right place.
- Specific details need to be extracted from contracts, letters or reports and entered into a table.
- Large volumes of documents have to be sifted and pre-sorted before someone processes them in detail.
- Free-text fields – feedback, descriptions or notes – need to be evaluated and turned into fixed categories.
- Recurring texts such as summaries, reply drafts or report sections are written by hand every time.
Who this is for
Companies and business teams that process texts and documents on a regular basis: case handling, customer service, administration, purchasing, legal and contract teams, consulting firms. You don't need an AI strategy or a data team, just a specific workflow that costs time today.
What can be automated
Extracting information from texts
Pull names, amounts, deadlines, addresses or other details from letters, contracts or reports and store them in a structured form.
Classifying documents
Categorize incoming documents or requests by type, topic or responsibility and route them accordingly.
Summarizing content
Condense long documents, minutes or correspondence to the essentials, following guidelines that fit your needs.
Structuring free text
Turn unstructured input into fixed categories and fields so it can be analyzed.
Generating drafts
Prepare reply drafts, report sections or text blocks that a person reviews and approves.
Pre-structuring large document volumes
Sift, group and tag collections so the actual processing goes faster.
Rules or AI? How I decide
The first question is always: Can the task be described with unambiguous rules? Calculations, fixed business rules, data transfers between systems, deterministic checks and simple file operations belong in conventional automation. It delivers the same result every time, is traceable, and costs almost nothing to run.
A language model comes into play when the content of a text has to be understood and the cases are too varied to be captured in rules. Even then the model is just one component: conventional software fetches the data, hands it over in a structured form, checks the answer and decides what happens with it.
The decision also covers which data the model is allowed to see, where it runs, what it costs to operate and how often an error is acceptable. Where a single error would be expensive, a person checks the result before it's used.
If your team regularly reads and categorizes texts or documents, a few typical cases are enough to gauge whether a language model helps or a simpler solution will do.
Talk AI automationWhen does it make sense?
Makes sense when …
- the task comes up often and ties up people who read and categorize texts today.
- the cases are too varied to be covered with fixed rules.
- the criteria can be described and the results can be verified.
- the occasional deviation can be caught by review steps.
Less suitable when …
- the task follows unambiguous rules – then conventional automation is more reliable.
- every single error would have serious consequences and no review is planned.
- the data situation doesn't allow processing by an external model and running one locally doesn't justify the effort.
How I work
- 01
Collect the task and real examples
We walk through the workflow and collect real examples: typical cases, difficult cases, edge cases. That shows what a model has to deliver and where plain rules are enough.
- 02
Check feasibility
With those examples I test how reliably a model solves the task. You see results before you decide on an implementation.
- 03
Build the workflow
The solution is built as a combination of conventional software and a language model: data access, structured inputs and outputs, review steps, logging, and the points where a person decides.
- 04
Roll out and monitor
After rollout the results are monitored for a while. You receive documentation, source code and a walkthrough, and you know exactly where the solution's limits are.
What you get
- An automation that uses language models only where they add value
- Structured inputs and outputs, review steps and logging of every run
- Clearly defined points where a person checks or decides
- Source code, documentation, a walkthrough and an honest description of the limits
Limits
- Language models make mistakes, even with good instructions. A solution without review steps is unsuitable for most business processes.
- Which data a model may process has to be settled up front, depending on the type of data, the provider and the operating model. I don't make blanket promises about that.
- Operation has running costs that depend on the model and the data volume. For small volumes that's negligible; for large ones it has to be planned.
Frequently asked
- Which processes are a good fit for AI automation?
- Processes where people regularly read, categorize or summarize texts or documents, or transfer details out of them, and where the cases are too varied for fixed rules. Examples: categorizing incoming requests, extracting details from letters and contracts, or pre-sorting large document collections.
- When is conventional automation the better choice?
- Whenever the task can be described with unambiguous rules: calculations, data transfers, checks against fixed criteria, file operations. Conventional automation delivers the same result every time, is traceable, and is far cheaper to run.
- Can documents be analyzed automatically?
- Yes, as long as it's clear which details are needed. A language model can extract content from contracts, reports or letters and turn it into a fixed structure. Scanned documents go through text recognition first. Reliability depends on the quality and consistency of the documents and is tested up front with real examples.
- How reliable are AI outputs?
- Good enough for many tasks, but never error-free. The actual hit rate can only be measured on your specific case, which is why I test with real examples before every implementation. The solution is built so uncertainties become visible and errors get caught instead of propagating unnoticed.
- Does a person have to check the results?
- In most business processes, yes, at least at the points where an error would have consequences. How much oversight is needed depends on the process: for pre-sorting, a spot check is often enough; for details that feed into contracts or payments, an approval step is needed.
- Can AI and conventional automation be combined?
- That's the normal case. Conventional software fetches the data, prepares it, hands it to the model in a structured form, checks the answer and processes it further. The language model only takes over the part that rules can't handle.
Next step
Which texts or documents cost your team time every week?
Describe the workflow and the type of documents. I'll tell you whether AI actually helps here – or whether a simpler solution will do.