AI as a research and synthesis partner.

AI has become my second brain during research work. I create separate pages for each project and dump in everything: my thoughts, customer notes and observations. I then prompt Claude to pull out information I need. Then I use this page to filter information whenever I need it, like: "How many customers mentioned needing a password reset?" or "Pull out quotes from our notes related to reset controls."

Outreach and guide prep

ChatGPT is very useful in helping me craft quick outreach emails. If Claude is hooked up with outlook, I would use Claude to send outreach emails.

For research guides, I usually provide my core research goals and key questions, and Claude helps flesh out the structure by drafting introductions, conclusions, and throws in follow-up questions. After one final editing pass from me, the guide is typically ready for usability testing.

Note taking and synthesis

During interview sessions, I still jot down standout moments and points of interest because it helps me think. But I no longer try to capture everything verbatim. I know I can always fall back on AI notes and summaries later.

That changes how I participate in research sessions. I spend more time listening and asking follow-up questions instead of transcribing constantly.

AI is also useful for synthesizing smaller and more uniform research studies, especially usability tests around focused workflows like password reset or onboarding.

For larger research efforts with more varied participants and themes, I use AI more as a thinking partner:

• helping me identify blind spots
• connecting ideas I may not have considered
• surfacing contradictions or missing context

That said, I still don't fully trust AI to independently synthesize nuanced research without human review and interpretation.

Lit review and ramping up

One area where Claude saves me a lot of time is literature review and product understanding.

There was a time when understanding a product problem meant jumping between:

• Amplitude
• Salesforce tickets
• support escalations
• internal docs
• Reddit threads
• competitor forums

Now I can point Claude at a collection of sources and ask it to summarize:

• recurring customer pain points
• onboarding issues
• adoption problems
• competitor complaints

For example, I once had to understand what issues customers were facing with admin APIs and where the product gaps were. Normally I would have spent days learning tools, digging through tickets, and manually piecing things together. Claude gave me a starting point much faster.

I also use it for quick competitive analysis by scraping through Reddit discussions and Cisco customer complaint portals to get a rough overview of a problem space.

That said, I still verify information carefully. If the underlying tickets or documents are poorly written, Claude can misunderstand the issue or hallucinate details.

The most useful part is usually the back-and-forth:

• why did you reach this conclusion?
• what evidence supports it?
• what sources are you relying on?
• what am I missing?

I've found that the best results come when I treat Claude less like a search engine and more like a collaborative thinking partner. One that I guide and train better to understand my research goals.

Here are the docs Claude research mode and partnered on: