Before replacing your data platform: keep, fix, or move?
Compare keeping, fixing, or moving one reporting workload, including the conditions for cutover, rollback, and retirement.
Read articleNotes from real work in business, data, and AI.
Latest articles
When an automation can’t find a source, check its access
Radek could open the source, but his automation couldn’t read it. Here’s what to check before asking someone to recreate supposedly missing input.
Read article
From Chat Window to an AI Nervous System: Six Levels of Working with AI
Six levels of working with AI through a weekly report and my work on HeyRup. What instructions, tools, context, and governed autonomy add to the process.
Read article
How to evaluate an AI pilot before investing more
What to ask your team before investing more in AI: output quality, all human work, costs, and how saved time will be used. Includes an example and a worksheet.
Read article
A new tool needs to help the person receiving its output, too
Making one person’s work easier can create more work for someone else. What HeyRup’s document handoff design shows, and how to include the recipient when evaluating a product.
Read article
An AI second brain: turn your notes into a decision system
An AI second brain becomes useful when it separates current truth, past decisions, working notes, and sensitive sources. A practical way to start without building another complex system.
Read article
Why Business Meetup doesn’t record its discussions
A recording helps people who couldn’t attend, but it can change what gets said in the room. What Radek’s choice not to record discussions shows, and where that approach has limits.
Read article
Can your leadership team rely on its data?
Six practical checks to see whether your numbers help you make decisions or just look precise in a report.
Read article
Building an automation is only the beginning
Your automation ran and the first result looks right. Here’s what to check so it still works next month, how to spot trouble, and when to pause it.
Read article
Same subscription, different usage: how AI agents change SaaS pricing
With an agent, one customer can consume much more work under the same subscription. What Google is already changing, and how I would price a product that agents will use.
Read article
Before investing in AI, get clear on the first step
Does your company need a data audit, a shared understanding of the problem, or a test of one AI workflow? Three lessons from developing Radek’s service offering.
Read article
SaaS isn’t dead. What we pay for is changing.
I pay for tools whose interfaces I haven’t seen in months. AI is changing how we use and value SaaS. I see room for reliable, specialized services.
Read article
From ETL to AI agents: why data teams need business context
Agentic BI makes the work we used to put off more valuable: data modeling, metadata, column descriptions, table names, and a well-maintained warehouse.
Read article