🔍 Read the Company (Where It Hurts)
Before any tool, the intent architect does one simple, powerful thing: look at the company through the eyes of someone looking for pain points. It’s not about looking at what’s polished; it’s about noticing where people complain, where work gets stuck, where customers have to wait, and where money leaks away. Pain is the best opportunity map there is—and almost no one stops to read it.
🎯 Pain is a signal, not a flaw
Every repeated complaint ("this takes a huge amount of time," "no one can find this information," "the customer got upset again") points to a place where AI can multiply results. The diagnosis doesn't judge the company — it turns complaints into opportunities.
💡 Practical tip
Spend a day (or a 30-minute conversation) observing an area with your "pain radar" on. Whenever someone says "what a pain to do this," "again?" or "where is that?", write down the exact words. You'll come away with a list of candidates worth their weight in gold.
🔁 Repetition and Wasted Time
The radar’s first target is the repetition: the same task, done by hand, several times a day. Answering the same questions, filling in the same fields, copying data from one system to another, putting together the same report every Monday. The more often a task repeats, the more AI multiplies its impact—because automating a task that happens 50 times a day gives back 50 times the time.
✓ Signs of costly repetition
- ✓"I do this every time an order comes in."
- ✓Copy and paste between a spreadsheet, email, and system.
- ✓Same response typed for different customers.
- ✓Happens many times a day, for many people.
✗ Don't confuse it with
- ✗A task that happens once a month (little benefit).
- ✗A unique creative decision that requires human judgment.
- ✗Something that “bothers” someone but takes very little time.
- ✗Repetition that nobody actually does anymore.
The cost of wasted time
Fill this in for a real task. The number can be alarming—and it's your best argument.
🗂️ Disorganized Information
The second classic pain point is the scattered information: the answer exists, but nobody knows where. It’s in a spreadsheet, a WhatsApp group, an old email, or only in the head of someone who’s on vacation. When knowledge doesn’t have a single, searchable home, the company wastes time looking for it and makes mistakes because it can’t access it. AI shines exactly here: organizing and retrieving information is one of the things it does best.
🧭 How disorganization shows up
- Knowledge in your head: "only so-and-so knows how to answer this" — and so-and-so becomes a bottleneck.
- Duplicate data: the same information in three places, all out of date.
- Manual lookup: minutes scrolling through conversation history to find a number.
🐢 Slow decisions and poor customer service
The third family of pain points is time between the question and the answer: the customer who waits hours for a response, the manager who puts off a decision for lack of information, the support representative who repeats the same answers and can’t keep up with the queue. Response speed is a competitive advantage—and every minute of waiting costs you sales, trust, and customers. AI shortens the gap between a question and the right answer.
⏱️ Where the clock hurts
- Customer service: the customer waits, questions keep repeating, and the queue doesn't move.
- Decision: "I need a piece of data to make a decision" and the data takes a long time to arrive.
- Internal response: one team gets stuck waiting for information from another.
Customer asks
Send a message outside business hours — and get no response.
Waits in the queue
The support agent is only available in the morning; the question grows stale along with the customer's patience.
Finding the answer
The support agent searches for scattered information—the bottleneck from the previous topic.
The customer has already given up
The answer arrives too late—and the customer has already bought from the competitor who replied right away.
💡 Practical tip
Measure "time to first response" in some area of the company. If it's measured in hours (or days), there's a clear speed issue there—and a great problem for AI to shorten.
💡 Automation opportunity
Finding pain points is easy; prioritize is what sets the architect apart. Not every pain point is worth automating right now. The question is twofold: how much this pain point costs today (time, money, lost customers) and how easy it is to solve with AI. The best target is a pain point that costs a lot and is easy to tackle—the classic quick win. The worst is one that costs little and is hard to solve: lots of effort for little return.
Cost × ease matrix (mental)
Illustrative recreation of the matrix—use it to rank the pain points you identified.
✓ A good automation target
- ✓Costs a lot of time or money today.
- ✓Has clear inputs and outputs (repeatable rule).
- ✓It fits into a one-day prototype.
- ✓Someone at the company feels the pain every day.
✗ Target to avoid (for now)
- ✗Requires integration with 5 critical systems.
- ✗Depends on subtle, costly human judgment.
- ✗Tiny gain even if it works.
- ✗No one can say what the “after” would look like.
✂️ Choose ONE problem (focus)
Now for the hardest step: nail down ONE problem. After mapping many pain points, the temptation is to solve them all. But trying to solve everything at once is the surest way to deliver nothing. Focus means choosing the most painful, feasible problem within a day’s scope—and leaving the others on the list for later. One problem, one solution, one result to show.
Diagnostic checklist — I close out the problem
Run the checklist on the strongest pain points on your map. The one that checks every box is your choice.
🎯 Focus means saying no
Choosing one problem means, in practice, saying “not now” to the other four. That’s scary—it feels like you’re leaving value on the table. But it’s the opposite: only by delivering ONE real result do you earn the right (and the trust) to tackle the next one. The list of pain points doesn’t disappear; it becomes your roadmap.
Self-check (optional): after mapping five pain points at a company, what is the architect's next step?
🔍 Module summary
Next module:
3.3 — Design the Architecture: turn the selected problem into the solution blueprint.