Research mode searches the live web and answers with cited sources. That much is unremarkable and every assistant does it. What makes it worth building a habit around is the step immediately after: the sources become pins in one tap, so the research survives the conversation that produced it.
How it works
Switch the composer to Research and ask. The answer arrives with a strip of sources beneath it, each showing its title and its domain, each one clickable.
Under the strip sits a single offer: keep these. Taking it saves every source as a pin, carrying its title, its link, and the snippet the search returned. They land tagged as web, in your personal pins, and a link takes you straight there.
That is the whole mechanism. It is small, it takes one tap, and it changes what research is for, because the alternative is a conversation you will not find again and eleven browser tabs you will close by accident.
The use it was built for
Researching a company before you deal with them.
Start with the company. What it does, how large it is, who runs it, what market it sits in. Keep the sources.
Then the situation. Recent news, funding, leadership changes, anything in the last year. Keep those too.
Then the person. Who you are actually meeting, what they have said publicly, where they were before this.
Then file it. In Pins, make a collection named for the company, switch to List view, select what you gathered, and move it in. One action.
Then add the record. Pin the company itself, or the deal, or the primary contact, into the same collection.
What you end up with, in about fifteen minutes, is one collection holding both what you found outside and what your organisation already knows inside, in a place you will actually look again.
Why the collection matters more than the search
Anyone can search. The reason to do this inside NearSync rather than in a browser is what sits next to the results.
A pinned deal is not a link to a deal. It is the deal, resolved fresh every time you open the collection, showing the stage it is at today and the value it carries today.
A collection therefore ages correctly. The articles stay what they were, which is right, because that is what was true when you read them. The record stays current, which is also right. Six weeks later, opening the collection tells you both what you learned and where things now stand.
A folder of bookmarks tells you only the first half, and the half it tells you is the half that goes stale.
Sharing it
A collection can be shared with your team, which shares every pin inside it.
The case is obvious once you have done the work once: the second person to research the same account should start from what the first one found. It costs one click and saves the entire exercise.
Sharing operates per pin underneath, so anything you deliberately keep personal inside a shared collection needs moving out first. Half-formed judgements about a person are the usual candidates.
Voice works the same way
Sources found during a spoken conversation are held until the answer lands, then attach to it as the same pinnable strip.
Ask out loud, keep the sources, carry on talking. This is genuinely useful walking between meetings, and the pins are waiting when you sit down.
What it costs
A research question costs more than an ordinary one, because it is a search followed by a model reading the results. It is still small, and it is charged per question rather than per source, so a question that returns eight sources costs the same as one that returns two.
Pinning the sources costs nothing. Keeping what you found is free; finding it again is not, which is the entire argument for keeping it.
Questions that work, and questions that do not
Research mode is a search with a reader attached, and it inherits the strengths and weaknesses of both.
It is good at things that were published: what a company does, who runs it, what it announced, how a market is described by people who write about it, what a technology is for.
It is poor at anything private, anything very recent, and anything requiring a judgement the open web has not already made. Asking whether a company is a good prospect will produce a fluent paragraph assembled from marketing copy, and you should not act on it.
It is actively misleading on small private companies, where the only published material is what they wrote about themselves. A confident summary of a company's strengths sourced entirely from that company's own website is not research, and it reads exactly like research.
The practical rule: ask for facts and read the sources; do not ask for conclusions.
Two honest limits
Cited does not mean correct. The sources are shown so you can check them, and for anything that matters commercially you should open at least the ones you are going to repeat out loud. A citation is a place to look, not a guarantee.
It searches the public web. It does not reach inside a company's private systems and it will not find what was never published. For the parts nobody wrote down, the meeting is still the research, and the write-up afterwards is where it gets recorded.
A discipline worth keeping
Do not pin everything. The temptation after a good research session is to keep all fourteen sources. Keep the six you would send to a colleague and let the rest go. A collection with everything in it is as unusable as the open web, which is what you were trying to escape.
Write one note. After the reading, add a short note to the collection saying what you concluded. It is the only thing in there that captures your judgement rather than somebody else's facts, and it is the thing your colleague most needs.
Did this answer your question?
No, ask a person