Why, What, Where, When, and How are the five questions any journalist learns to answer before writing a story. It's a method with decades of use because it works: it forces you to cover every angle of a situation before drawing conclusions. At Fivew we take that same framework and apply it to a very current decision: how to choose your AI tools well.
WhyWhy do you want to bring in AI?
Before looking at any tool, write down in one sentence the business problem you're trying to solve. Not "I want to use AI in customer service," but something more specific: "I want to cut first-response time on customer inquiries from 24 hours to 2." The more concrete the goal, the easier the rest of the process.
WhatWhich tool actually serves that purpose?
Only at this point do you start evaluating specific tools. With the goal from the previous step as a filter, many options rule themselves out: you don't need the most complete platform on the market, you need the one that solves your specific problem, within your budget, with a learning curve your team can handle.
WhereWhich process are you going to apply it to?
The same generative AI tool can transform the customer service area and be irrelevant in the finance area of the same company. Map your internal processes and locate where the real bottleneck is: the place where the most time is lost, the most mistakes happen, or the most opportunities slip away.
WhenIs it the right time?
Not every process is ready to be automated. If a process is still inconsistent or depends on criteria that change depending on who's running it, automating it will only scale the disorder. The general rule: first tidy up and consolidate the process manually, then automate it. Timing matters as much as the tool.
HowHow do you make sure it's actually used?
The last question — and the one that sinks the most projects when ignored — is real-world implementation. Define who's going to use the tool, train them with concrete cases from the company (not generic examples), document the process, and follow up during the first weeks of use. Without this step, even the perfect tool ends up abandoned.
No single one of the five questions is enough on its own. It's the order that turns them into a solid decision.
ExampleThe 5Ws applied to a real case
Picture a services business that gets a lot of repeat questions over WhatsApp:
- Why: reduce the time the sales team spends answering frequently asked questions.
- What: a simple conversational assistant, trained on the most common questions.
- Where: the WhatsApp channel used for first contact, before handing off to a salesperson.
- When: once the team already has a consolidated list of FAQs and standard answers.
- How: train the team to supervise and correct the assistant in the first weeks, with a weekly review of conversations.
This is exactly the process we follow in every Fivew project. If you want to dig deeper into each dimension of the method, you can see the full breakdown on our methodology page.