Connect a Gmail or Microsoft 365 account and your contacts come with it, duplicates already collapsed. The client who exists four times in your address book is one person in the feed. Your address book itself stays as it is. Google Contacts and Outlook People hold exactly what they held on Monday.
- The merging happens on the way in. By the time you open the feed it’s done. There’s nothing to schedule, nothing to approve and no job waiting on you.
- Point reads who’s who instead of rewriting your records. Your address book keeps every row it has, under the name it has, and the folder list stays the one you already keep.
- Both providers already separate a stored contact from a person. Google calls its person data merged from several sources. Microsoft’s is an aggregation across mail and contacts.
- A mark goes on the human being. Star a client once and the mark covers all four addresses, plus a fifth one the first day they use it.
- A wrong merge happens sometimes, and it’s cheap to put right. Both records underneath are still whole, so you correct Point’s view and there’s nothing to restore.
What happens to your contacts when you connect
Your contacts come across on their own. The people you deal with arrive with the account. By the time you first open the feed, the duplicates are one entry each. Point merges the duplicates and you mark the people who matter. That’s the whole of the setup on this subject.
On the Google side that’s part of what makes the connection cheap. You connect the Gmail you already have, and you keep your address, your history and your contacts, because it all stays where it is. On the Microsoft side the account comes in as mail and calendar together, and the people come with the mailbox rather than as a separate thing to authorize. Either way the resolution runs on arrival, in the minutes after you connect, rather than as a project with a start date and a finish.
What a connection creates in each account is a short list, published in full. For a Google account it’s connecting Gmail, and what stays exactly as it was. For a Microsoft one it’s the Outlook path, from mailbox to calendar. Your contacts sit outside both lists. That’s where this page spends most of its length, and it’s the reason the page exists.
A stored contact and a person are different things
Google and Microsoft both draw this line inside their own products. Watch them draw it and you can see why your address book has four of somebody in it, and why a tidy-up only holds for a while.
Take the stored record first. Microsoft’s description of an Outlook contact is that it’s “an item in Outlook where you can organize and save information about the people and organizations you communicate with”, held in contact folders (Microsoft Graph, checked September 6, 2026). One record, one filing place, and more room for addresses than most people expect. There’s a collection described simply as “The contact’s email addresses”. The record also carries separate primary, secondary and tertiary address properties (same page, checked September 6, 2026). In Outlook’s model of the world, one person reachable three ways is a field on the form.
Even the name printed above those addresses is softer than it looks. Microsoft’s own note is that “later updates to other properties may cause an automatically generated value to overwrite the displayName value you have specified” (same page, checked September 6, 2026). Some of the names in your list you wrote yourself. Some were generated. The list shows them the same way. That’s a large part of why one client turns up as a full name, as a surname with an initial, and as a bare address, all three looking equally deliberate.
Both companies ship a second thing as well. Google defines its person data as “Information about a person merged from various data sources such as the authenticated user’s contacts and profile data”, and holds their mail as a list rather than a single value: “The person’s email addresses. For people.connections.list and otherContacts.list the number of email addresses is limited to 100” (Google for Developers, checked September 6, 2026). Microsoft’s equivalent uses the same word for the same idea. Its person “Represents an aggregation of information about a person from across mail and contacts”, assembled from “local contacts or your organization’s directory, and people from recent communications (such as email)”. Ask for them and they come back “ordered by their relevance to the user” (Microsoft Graph, checked September 6, 2026).
So the two largest mail providers each ship two things. One is a stored record. The other is a merged view of a human being, built out of several stored records and a history of who you actually write to. Both companies are clear about the difference. Your address book is the first thing. Who you’re talking to is the second. They were always two lists.
Which sets up the honest version of the problem. Duplicates come from the way records get saved, so they keep arriving. A new address on a signature block this month. A phone that syncs one way. An import somebody ran in 2019. The four ways your own four rows got there, and what a cleanup of them costs, are worked through in cleaning up duplicate contacts without losing anything. What matters on this page is that one careful weekend answers a question that comes back.
Merging that leaves your records where they are
The merge happens in Point’s reading of who’s who. Your Google Contacts and your Outlook People list stay exactly as they are. Every row is still there, under its own name, with all its addresses on it. No group or folder turns up carrying a product’s name. Open your address book on Friday and it holds the four rows it held on Monday, in the same order, with the same spelling mistakes.
That’s worth saying flatly, because plenty of tools in this area do the opposite. There’s a second reason to say it flatly: the list of what a connected application writes into a Gmail or Microsoft mailbox is public and short. Four things are on it: a triage name, an archived thread, a sent reply, a calendar entry for a meeting that got settled. Your contacts sit outside that list. The whole inventory is what an email app writes back into your mailbox.
There’s a practical reason to want the merging on this side of the line. A merge written into an address book takes the two records and leaves one. Getting back to two means reaching for whatever reverse gear your provider offers, on a clock, against every other change you’ve made since. A merge that lives in a reading leaves both rows where they are. They stay whole, and each one is still yours to edit on its own. A wrong answer about them is a wrong answer, and your records are fine.
Your contacts sit on the Google or Microsoft account rather than inside any mail application. That’s exactly why they survive a change of software, and what stays yours when you change email apps is the fuller account of that. They belong to you, and the merging is built to keep it that way.
What one person buys you on a Tuesday
Three ordinary places, and the first is the one you feel.
A mark goes on the human being rather than on a row. Star a client and their mail counts for more from then on. You decide who gets starred, and the effect shows up straight away. Something perfectly ordinary from them lands high in Later and marked important, instead of dropping into Other with the chatter. A note that only says thanks goes there too. Now put that beside an unresolved contact list. The star sits on one of four rows, so three quarters of that client’s mail arrives unmarked. The addresses they use away from their desk are the ones that miss out, and those are the ones that carry the bad news. Deciding who gets marked is the half hour that decides who matters. What the mark does afterward is VIP mail that never gets buried. The term itself is what a VIP contact is.
The history is one history. Ask what’s passed between you and a client over two years and the answer covers all of it, rather than the third that happened to use one address. A split identity gives you a confident answer. It looks complete, and it heard half the question. That’s worse than an obviously partial answer, because you act on it. Everything built on top of a thread inherits that, and reading a long chain without reading the chain is only as good as its idea of who’s in it.
A conversation stays one conversation. An email answered on another channel stays one thread, with the context in one place. It holds together because there’s one person for it to belong to. That case is one thread even when it jumps channels.
When the merge gets it wrong
Two people can look like one. It happens where the evidence genuinely points the wrong way.
A married couple who run a business together and share a household address. Two people at a client with the same surname and an initial that renders the same way on both records. A department address that four people answer in rotation, with a name, a signature and years of correspondence behind it. In each of those, everything a resolution can see says one human being. What says otherwise is knowledge that lives in your head.
The tell is worth knowing in advance, because it surfaces later rather than on the day. It shows up weeks on, as a message filed under a name that has no business holding it. Or as a summary of a client that mentions a project the client has never heard of. Something reads slightly wrong, and it takes a minute to work out why.
Putting it right is cheap, and that’s the payoff of the section before. What changes is Point’s view of who’s who. The two records in your address book are still two records, so they stay where they are and nothing needs restoring. No undo window at your provider is running against you, and the correct work you did in the same sitting stays done. You’re correcting a reading, and your data is where you left it. Everything Point did on your behalf is kept as a record you can read, and undo covers Point’s own actions: the activity log behind every action.
The wider version of that question, which owners ask sooner or later, is how much Point does before checking with you. That’s a setting you hold, and each kind of work carries its own. Every one of them arrives on review, so the work is prepared for you and waits for your go. Each position and where it stops is setting how much your inbox does on its own. What Point does while it’s running is on the benefits page.
The client who sits in both accounts
Plenty of owners connect more than one account. The firm runs on Microsoft 365 and a gmail.com address still catches half the referrals. Or a second business started years later on whatever was easiest that week. Each connection is granted on the account it belongs to. Each provider’s address book stays where it has always been. Everything below happens with the two kept apart.
Where the two accounts are two businesses you operate, the businesses stay apart, and that separation comes first. Each card is labeled with the business it came from. A reply leaves inside that same business, and the account it goes out on is already the right one. So a supplier you use in both businesses keeps one entry in each. The isolation is the more valuable property here, and it’s the one you’d miss. What that separation covers, and what it deliberately leaves alone, is keeping two businesses apart.
Where the two accounts are one life rather than two businesses, the shape is gentler. Each message is sorted into the part of your life it came out of. A note from the school sits in one part and a client asking about a bill sits in the other, and they’re weighed separately. The same person can perfectly well be your accountant and a neighbor, and resolving them to one human being is the right answer in both directions.
What this does not do for you
One boundary, stated once, because mixing up the two jobs is how people end up disappointed.
Point is a layer over the mailbox you already have, rather than a CRM. Resolving your contacts is one job and cleaning them is another. Say you want a tidy contact list at the end of this, to hand to a bookkeeper, to a practice system, or to whatever you set up next. That work happens in Google Contacts or Outlook People. It carries a real reverse-gear problem, and the page for it is cleaning up duplicate contacts without losing anything. Point leaves that job with you, and it’s yours to do whenever you like.
Which answers the question everyone asks last. End the connection and your address book is exactly as you left it. The same rows, the same duplicates, the same 2019 address still sitting among them. What stops is the reading laid over it. That’s a real loss and a clean one, and it’s the only kind worth signing up for.
Common questions
Does Point merge the duplicates in my Google Contacts or Outlook People list?
No, and that’s deliberate. The merging happens in Point’s reading of who you deal with, so your address book keeps every record it had, under the name it had. A connection writes a short, published list of things into your mailbox, and your contacts sit outside it. If a tidy address book is what you actually want, that’s a separate job you do in Google or Microsoft.
Will I lose an old email address when two records become one person?
No. Every address on every record is still on that record in your address book, and every one of them still reaches you in the feed. The merge keeps them all, because it leaves your records alone. That matters more than it sounds. The address a client last used in 2023 is often the one they reach for on a Sunday night, from a phone, because something has gone wrong.
What happens if Point decides two different people are the same person?
It happens sometimes, most often with a couple sharing a household address, or with two people at one firm whose names render the same way. You usually notice weeks later, as a message sitting under a name that looks wrong. The repair is cheap, for the reason the rest of this page is about. Both records in your address book are still there, so you correct Point’s view and your data stays as it is. No clock at Google or Microsoft is running against you while you work it out.
Do I need to clean up my contacts before I connect an account?
No, and there’s a case for leaving it. The duplicates are resolved on the way in, so the feed reads the same whether you spend the weekend merging rows or spend it away from the screen. Cleaning up is also the riskier half of the work, since a merge written into an address book takes the two records and leaves one. Connect first. See whether the untidiness still costs you anything, and do the cleanup later if it does.
Does the star I put on a client cover their other addresses?
Yes, and that’s most of the reason the merging is worth having. The mark goes on the person rather than on one row. Their mail lands at the top of Later marked important whichever address it came from, including an address you’re seeing for the first time. On an unresolved list the same star would cover only the record you happened to be looking at when you set it.
Is this the same on Microsoft 365 as it is on Gmail?
Yes. Contacts come across with the account on both, and they arrive resolved on both. Both address books keep exactly what they hold. The differences between the two connections are about how filing and sorting get recorded in each mailbox, in its own vocabulary, and they’re set out in the Outlook path, from mailbox to calendar.