SourceLace Docs
FAQOpen the app

All FAQ categories

Asking questions and getting answers

What people can ask, how big an answer can be, and how to know an answer is right.

What can I ask?

Anything your own access in a connected system lets you see, in plain words. For example "Which renewals over 50,000 close this quarter?" or "Summarize the open priority 1 cases for Corvanta". The AI looks up the fields first, then reads the system as you and shows you the rows it used.

Details: What SourceLace does.

Why can't I see a record I know exists?

SourceLace signs you in to each system with your own login, so it shows exactly what that system lets you see, and nothing more. Check that you can see the record in the system itself, and that you connected the right account. If your organization uses access by group, your groups must also allow that source and object.

Details: Access control.

Can I ask one question that spans several systems?

Yes. Ask, for example, "Which accounts with open renewals have overdue SAP invoices?". You get an exact answer that spans your systems: every row is counted, nothing is estimated, and you can see which system each number came from. Each system is read with your own sign-in.

Details: Asking across systems.

Can I combine health records with HR or billing data in one question?

Yes. Name a health record resource as a table in one question and join it to your other table by the patient's reference or id. The Health records chapter shows how.

Details: Health records.

How does SourceLace know two records are the same?

Your organization's matching rules say so, one for each pair of objects, such as Salesforce Account and NetSuite Customer. A rule tries ways in order, for example the same VAT number, then the same website domain and country, then a similar name in the same country. So "ACME Corp." and "Acme Corporation GmbH" with the same VAT number count as one customer in every answer. Pilot: ask your SourceLace contact to turn it on.

Details: Matching records across systems.

Why did an answer say some records had no match?

When a question combines two objects that have a matching rule, the answer counts how many records matched and how many did not, such as "412 of 430 Salesforce Account records matched; 18 had no match". A record has no match when none of the rule's ways found a partner for it, for example a customer that exists only in Salesforce. Choose Show unmatched under the answer to see them, one side at a time.

Details: Reading an answer.

What happens if there's no matching rule?

The records are joined the way the assistant writes the query, usually on values that are exactly equal, such as the same customer number. So "ACME Corp." in Salesforce and "Acme Corporation GmbH" in NetSuite don't count as one customer, and the answer shows no match counts. A matching rule is what tells SourceLace that differently written records are the same.

Details: What matching does and doesn't do.

Does matching merge records or change my systems?

No. Matching only decides which records count as the same in an answer. It never merges records, never writes anything back to your systems and doesn't pick which system's value is right. It also doesn't learn from your data or recognise nicknames: two records match only when a way of your rule says so.

Details: What matching does and doesn't do.

Can I match by similar names (fuzzy matching)?

Yes. This is often called fuzzy matching. Use the Similar, high or Similar, medium comparison, such as "Globex Industries" and "Globex Industrys". A way with a Similar check also needs a Same check, such as the same country, and its pairs are marked Check for a person to look at unless the way also has a Same check on an identifier, such as a tax number. For people, match on email or phone first: names miss nicknames and typos.

Details: Ways and bands.

Who decides how records are matched?

Your admins, and the people they name as editors of a rule, on Business setup → Matching. A rule is published only after a test on live data, and then applies to everyone and every agent, so two people asking the same question get the same numbers. Anyone can suggest a rule from a chat; an admin approves or declines it.

Details: Who can edit a rule.

Can I match differently from my organization?

Only more strictly for your organization's rule: ask for exact matches only, such as "revenue per customer, exact matches only", and the answer leaves out pairs found by similar names or close values, and says so. Nobody can loosen a published rule from a question.

To try another way of matching, have the assistant draft a rule and choose Try in this chat on its card. Only that chat uses it, and every answer there says "Matched with your draft, not your organization's rule".

Details: Drafting a rule in a chat.

How do I propose a matching rule?

Ask the assistant in a chat, such as "Draft a matching rule for our customers in Salesforce Account and NetSuite Customer". It reads both objects' fields, drafts the rule and tests it on your data. On the card, choose Suggest for my organization and add a note on why it helps. Your admins see the rule and its test counts, never your records, and approve or decline it. You follow it under Suggestions → My suggestions.

Details: Drafting a rule in a chat and Reviewing suggested rules.

Can I see which system and field each number came from?

Yes. Press the small button next to a column's name: a panel says which field in which system the values came from, for example "From Salesforce: Account.Name", or "Computed from NetSuite: Transaction.amount" for a total. Show where columns came from, above the table, shows every column's origin under its name at once. If SourceLace can't tell where a column came from, the table says so instead of guessing. Under Your settings → Chats, Where this came from chooses whether tables start with the origins open, and your admins can choose that for everyone under Chat defaults → Chats.

Details: Where this came from.

Why does a cell show a dash instead of being blank?

In an answer that combines systems, a dash (a short horizontal line, read out as "no match, or empty in the system") means the question found no matching row in the other system for that line, or the matched record has nothing in that field. It does not by itself mean the other system has no data for that account: the names may be spelled differently, or the match may depend on another match that found nothing. Press the small button next to the column's name to see how the rows were matched. A cell that is simply empty in its own system stays blank, and a count of zero shows 0. Copy table and exports leave a dash cell empty.

How do I see which record in my system a row came from?

Above the table, press Show record ids to add a Record ids column, such as "Account 001A0". If the table has none because the assistant did not fetch them, the button says Ask again with record ids and asks the question again with them. A total has no record ids, because it comes from many records; ask for the records themselves, such as "list the accounts with their ids".

Details: Where this came from.

Does SourceLace remember which records it matched?

No. SourceLace keeps no list of matched pairs, keys or labels. Each question matches again, on your data as it is at that moment, so a record changed in one system is matched the new way at once, and nothing about your records is stored to be kept safe.

Details: What SourceLace keeps.

What does Check mean in an answer?

A pair marked Check was matched by a less certain way, such as a similar name in the same country, or one record matched several records on the other side. It is still in the answer; look at the pair before you rely on it: Check pairs under the answer shows each one side by side. For example "Globex Industries" and "Globex Industrys" in the same country is Check. An agent's changes after it read Check pairs wait for a person's approval.

Details: Ways and bands.

Can I search all our systems at once?

Yes, with Search everything (Pilot: ask your SourceLace contact to turn it on). Use the search box at the top of the app (press / to reach it), or ask the assistant to find something, such as "find the Corvanta renewal plan", and it searches your files, chat, mail, service desks and Salesforce at the same time, each with its own search and your own access, and lists the best results first with links. A system that is slow to answer is left out and named, and the answer says it is incomplete. Admins choose which sources are included.

Details: Searching everything.

Why did the search leave out one of our systems?

The results page names every system that was not searched and why. A system is left out when you have not connected it yet (connect it on My connections), when it did not answer in time (choose Try it again), when your groups give you access to only part of it, or when an admin left it out. Systems without a search of their own, such as databases, are never part of it.

Details: Searching everything.

Can SourceLace answer a repeated question without asking our system again?

Only if your organization chooses Saved answers, an optional add-on for systems that charge per query or limit API calls; otherwise every answer is read live. With it, your admin turns it on system by system, and your own recent answers from those systems can then be reused for a short time, at most 24 hours. Only you can reuse your saved answers, every reused answer says so and how old it is, and Ask live reads it again from the system. SourceLace never turns it on for you.

Details: Saved answers.

How many rows can a question return?

A chat result holds up to 2,000 rows, shown 50 at a time. When a question has a larger answer, SourceLace keeps every row for that result, up to 250,000 for one system, so totals and counts cover all of them. A question across systems can go further after you confirm the plan. Results are held for 30 minutes at most, or less if your admin chose less.

Details: Asking across systems and Data retention.

How do I know an answer is complete?

The answer says so. The line under an answer across systems says all rows read when every row was read from every system (so totals cover all the data), and partial data when it was cut short, and the AI must say so. The same line names the systems it read from and the number of rows, and How it was read lists each system, what was read and how many rows came back. If a question is very large, SourceLace shows the plan and asks before it fetches anything.

Details: What you see and Large questions ask first.

What is the line under an answer?

It is a line SourceLace writes itself, not the assistant: Answer check: all clear, Answer check: 2 things to know, Answer check: 1 error or Answer check: 1 error, 2 things to know. An error means something the answer relies on failed and nothing put it right; a thing to know is not a failure: your question ran fine, but the answer is limited, for example rows were cut short, a query returned no rows, rows came from saved answers, a column is blank because its rows were not found in another system (or some rows have no match: "7 of these 10 rows have no match in Snowflake"), or a column held values that are not numbers and were left out of its figures. Click Show to see each error and thing to know (under Good to know), most with a tip for what to try, what SourceLace checked (including the columns your organization masks, which are a fact and not counted), and what the assistant chose that SourceLace did not check. "All clear" does not mean the assistant picked the right data or that the numbers in its text are right: it means every check SourceLace makes passed and there is nothing to know. An answer that read health or claims records always carries a thing to know, by design: "The text below is written by the assistant from health or claims records and is not checked; check the records." So those answers never read "all clear". If SourceLace cannot build the line at all, it says "Answer check: SourceLace could not check this answer".

Details: The line under an answer.

Why does it say "7 of these 10 rows have no match in Snowflake"?

Your question ran fine. SourceLace looked in Snowflake for each of the 10 rows and found a match for 3; for the other 7 the cells from Snowflake are blank (a dash). It does not mean Snowflake has no data for them: the names may be spelled differently there, the field may be empty, or you may not be allowed to see those rows. Naming a shared field such as an account ID in your question usually fixes it. A count that found nothing shows 0, and SourceLace only says its 0 may be wrong when it read only part of that system. If the assistant tried again and the second table fixed the blank columns, the first is kept under Built from as Earlier attempt, replaced by Table 2, and the line does not count it.

Details: The line under an answer.

Can I see just the answer, without every table the AI read?

Yes, that is how answers look. Each answer shows the tables you asked for (two tables asked, two shown, in order), and nothing else in their place. The tables the assistant read on the way fold under it as Built from N tables: click the line to open it, and press Expand all (shown when there is more than one) to open them all. Under Your settings → Chats, Tables in answers chooses the default for all your chats: Folded folds only tables that were combined into a later one, and All tables shows everything. A project's owner can set it for the project's chats, and an admin for the whole organization; the most specific choice wins. Folded tables keep where their rows came from.

Details: Tables in answers.

Where did the list of steps go?

Once an answer is finished, the steps the assistant took ("Searching bigquery...", "Querying salesforce:dev") are folded under one line, Worked in 12 steps. Click it to see them; each step keeps its own Details. While the answer is still being written, the steps are in view so you can follow the progress. A step that failed and that the answer did not get past is never hidden: it stays under the line in red, and the line says 1 had a problem.

Details: Tables in answers.

Can the AI make up numbers?

SourceLace is built so it does not have to. Totals, counts and joins are worked out by SourceLace over the actual rows, and the AI quotes those figures. Every answer comes with the rows behind it so you can check. As with any AI, read important answers before you act on them.

Details: What it does.

The assistant says a field does not exist, but I know it does

When the assistant looks for something (a "health score", say), it searches the names of the objects and the names and labels of their fields too, as you, so it only finds what you may see. For every system that can describe its fields, SourceLace reads the field names in the background after the first search (at most once a day, and only when someone searches), so the first search may find only object names and a later one (a minute or so after) finds the field; the assistant's answer says how much it has looked through. Mail, calendar, chat, files and health records have no fields to search and match names only, and an admin can turn the background reading off for a source with Index field names, which the answer then says. If it says it did not find a field, ask again a little later, or name the object that holds it, for example "look at the fields of Account for a health score". A field your admin has hidden, or that your own sign-in in the system cannot see, will not appear. Example: "Which accounts have the lowest health score?" finds the Account field called Health Score even though no object has that name.

Can I make charts and documents?

Yes, in the SourceLace app. Ask for a chart, for a Word, PowerPoint or PDF file made from the results, or for the data itself as an Excel or CSV spreadsheet (every row, up to 100,000 per sheet, with masked values still masked). In a CSV file every text value is in quotes, numbers are written exactly (no 1E+3), and a value that starts with =, +, - or @ gets an apostrophe in front so a spreadsheet program cannot run it as a formula; the apostrophe is part of the text if another program reads the file. Charts can be bar (standing up, or lying down with the names on the left, which suits long names and rankings: ask for a "horizontal bar chart"), line, pie, scatter or timeline; ask for the kind you want and the assistant draws it or says it cannot. Pin a chart to a project board, and press Refresh later to redraw it with today's numbers, using your own access.

Details: Spreadsheets of your results and Refreshing a pinned chart.

Can SourceLace read our files?

Yes, from Google Drive, SharePoint and OneDrive, and cloud storage buckets your admin adds. SourceLace reads a file only when someone asks, pulls out its text, and never copies, crawls or indexes your files. You see only files your own account can open. Search works differently by store: see Can SourceLace search inside my files?.

Details: Files and What SourceLace does not store.

Can SourceLace search inside my files?

It depends on where the files are. In Google Drive, SharePoint and OneDrive, SourceLace uses the store's own search, which also looks at the text inside files, then reads the files you pick. In Amazon S3, Azure Storage and Google Cloud Storage, your admin chooses for each source: names only (the default: search matches file names, up to 10,000 names at a time, so ask by name or folder; searching inside one folder is faster, and your favourite folders are searched first), or names and file contents in the folders the admin allows, where a search also reads those files and finds the ones whose text holds your words. If your organization already runs its own search index over its files, an admin can also add it as a Search index source. If a search stops at one of its limits, the answer says so and is marked incomplete. Either way SourceLace reads a file only when someone asks and never copies or indexes it.

Details: How S3, Azure Storage and Google Cloud Storage are searched and Search index.

How do I find files faster in S3, Azure Storage or Google Cloud Storage?

Search inside one folder: ask for the folder by name, or pick it. A search inside one folder looks through fewer names, so it is faster and finds more in a big bucket. Save the folders you use most as favourite folders: on My connections, press Favourite folders on the source, find a folder and press Add. Your admin can also set favourite folders for everyone. Favourite folders are searched first and offered first. Only the folder paths are saved.

Details: Finding files faster.

Can we teach SourceLace our own business terms?

Yes. Admins define terms such as ARR or "open pipeline" once, with what each one means in your systems. The AI then uses the same definitions everywhere: in the SourceLace app, in AI agents and in your own AI app. A group can give a term its own meaning, which wins for that group's people: see Group terms, starter questions and default sources.

Details: Business terms.

Can I search my old chats?

Yes, in the SourceLace app. Search chats in the sidebar looks across all your chats, not only those on screen. Your chats are private to you and are kept for your organization's retention period.

Details: From the SourceLace app.

Why are big questions fast?

SourceLace asks every system at the same time, asks each only for the fields and rows the question needs, and, when a question only needs a total, a count or a smallest or largest value from a supported database or data warehouse, asks that system to work it out and send back only the result. The answer is the same as if every row had been fetched, and the answer's sources show exactly what each system was asked.

Details: Totals worked out by your system.

Do I have to switch anything on for my database to work out totals?

No. When a question only needs a total, a count, or a smallest or largest value, SourceLace asks a supported database or warehouse to work it out, without any setting in the app or in your system. The answer's sources say "aggregated at the source" and show the request that ran. If you want it off for one system, an admin can do that on its connection: Connections, choose the connection, set Let this system work out totals to Off.

Details: Totals worked out by your system.

Why does a total in SourceLace have more groups than my database's own report?

SourceLace compares text letter for letter, so "West", "WEST" and "west" (and "west " with a space at the end) are separate groups. Many databases, as they are usually set up, group them together in their own query tools. SourceLace does the same whether it fetches the rows or asks your system for the totals, so an answer does not change with how it was worked out. To count them together, say so in the question, for example "treat West, WEST and west as the same region".

Details: Totals worked out by your system.

Why do "West" and "WEST" show as one group in a total?

SourceLace keeps "West", "WEST" and "west" as three groups, as they are in your data. If a total merged them, the database or the column is set up to ignore capitals (a "case-insensitive" setting; Amazon Redshift, for example, lets a database be created that way, and in some databases and warehouses a single column can carry such a rule of its own), and the system adds the groups together before SourceLace sees them. To check, open Sources under the answer: a request that says "aggregated at the source" was worked out by your system. An admin can switch it off for that one connection (Connections, Let this system work out totals, Off), and SourceLace then reads the rows and groups them itself.

Details: Totals worked out by your system.

Can SourceLace group text the way my own database does?

Yes, if an admin chooses it for that connection. By default SourceLace keeps "West", "WEST" and "west" apart. For some database connections an admin can turn on Compare text the way this database does (Connections, choose the connection), and SourceLace then follows each text column's own rule, so totals match the reports your team already reads. The answer's Sources says when a rule was used. It covers totals by a text field and is, is one of and is not one of conditions on one table; joins between systems, sorting and masked fields stay letter for letter.

Details: Compare text the way your database does.

Why does a step say "Protected field left out"?

You asked for a field your organization's data protection rules hide, such as an email address. It is not an error. The assistant leaves that field out, asks again straight away, and says in its answer that a protected field was left out. To see the field, ask your admin to change the rule or to add your group to it.

Details: When a question asks for a hidden field.

Can I make the columns of an answer's table wider or move them?

Yes. Drag the edge of a column heading to make it wider or narrower, and use Columns above the table to move a column. Long text is cut to two lines: press Show full text to read all of it. These changes last while you stay on the page and are not remembered after a reload. You cannot hide a column in an answer's table, so nothing looks missing, and the values shown are never changed by any of this.

Details: Tables in answers.

Can I make one table or chart wider?

Yes. Press Expand above a table (last in the line of buttons) or in the corner of a chart: that one block becomes as wide as the chat. Press Collapse, or Escape while you are in it, to put it back. It only shows where there is room to widen the block. It is not remembered, so the next time you open the chat everything is at its usual width, and there is no Expand on a phone, where tables and charts already use the full width.

Details: Making a table or chart wider.

How do I know how full a chat is, and what happens when it is full?

Select the gauge at the top of the chat, beside Search everything and New chat. It shows how much of the AI model's window the chat fills and what fills it (SourceLace's instructions, what your organization set, your questions and the answers, and the results read from your systems, which are usually the biggest part). The sizes are estimates scaled to the exact total the AI provider reports, and no names from your systems are shown. Your saved chat always keeps every message. Compact all, Forget big results and New chat from summary each show a preview before anything changes, and Undo brings back what was sent before.

When a chat is nearly full, SourceLace compacts it for you before your next question, unless your admin or you switched that off, so the chat carries on instead of stopping. It forgets old big results first and, if needed, summarizes the older questions; it never touches the results, reasoning or opened-skill text of your last 3 questions, and your question is still answered if the summary fails.

Details: How full is this chat?.

Can I see when a message was sent?

Yes. Hover over a message, focus it with the keyboard, or tap it on a phone: it shows how long ago it was sent, such as "3 minutes ago", and its full date and time. After a long pause, a line such as "Resumed after 2 hours 40 minutes" appears before your next question.

Details: When a message was sent.

An answer failed. What can I do?

Press Try again under the red message to ask the same question again. If it keeps failing, press Copy details for your admin and paste the note to your admin; it holds when it happened and the message you saw, not your question. To reword the question, press Edit and resend under it.

Details: When an answer fails.

Why does a table say it was built by SourceLace from a plan?

When you ask for particular columns, names or an order, the assistant gives SourceLace a plan and SourceLace builds the table itself from the rows it read, so the table has the columns, headings, order and number of rows the plan describes, the same way for every system. The line under the table says so and shows the plan in words, with each renamed column next to the system's own name for it. SourceLace builds the table the plan describes; it does not check that the plan is what you meant, so compare the headings and the order with your question. If the plan could not be built (for example it named a column that is not in the rows), the table is shown as read and the line under the answer says why; ask again, naming the columns as they appear in the table.

Details: A table built the way you asked.

Does a file say what SourceLace checked?

Yes, except a CSV file. A Word, PowerPoint or PDF file ends with About this answer, and an Excel file has the same at the end of its About sheet: the headline from the line under the answer ("Answer check: all clear", "Answer check: 1 thing to know"), each error and thing to know with its tip, and the lists behind Show. A chart drawn from a table that was cut says so in one note. A CSV file is plain rows, so its checks stay on the line under the answer. A file made when SourceLace holds no check for the answer says "SourceLace's checks are not available for this answer." Nothing in the note is a value from your tables.

Details: What a file says about the answer.

Can I export the rows of a table to CSV or Excel?

Yes. Press Export above a table that came from your systems (on a phone, press More first) and choose CSV file or Excel file. SourceLace makes a file of all the rows of the result, up to a limit, not only the rows on the screen, and keeps it with the chat so you can download it again; your sort and search do not change it. The file appears as a card above the table and as a message in the chat: press Download on it. A very large result is limited to the first rows that fit (at most 100,000 rows and 15 MB); the line above the table says how many rows the file holds, and a file holds at most 200 columns (the line says when a table has more). Masked values stay masked in the file, and columns your organization hides are not in it. The file is kept with the chat, encrypted, and deleted with it. Your administrator can see in the audit log that you exported a result and how many rows, but not its contents. Results are held for a limited time, so export soon after asking; if the result has gone, ask the question again. In a CSV file a value that starts with =, +, - or @ gets an apostrophe in front so Excel does not run it as a formula; an Excel file keeps numbers as numbers.

Details: Exporting a table.

Why does sorting only change the rows I see?

A table in an answer shows the first rows of a result that can hold many more, for example 50 of 2,000. Clicking a column's name (or Sort by on a phone) puts those 50 in order, and the line under the table says so ("Sorted these 50 of 2,000 rows"). It does not read the other rows, so the top row is the biggest of the 50 you see, not of all 2,000. The search box above the table works the same way: it looks only at the rows on screen. To rank or search every row, ask the assistant, for example "show the 10 customers with the highest revenue" or "find the customers with Berlin in their address": it does that over the whole result. Copy table copies the rows as shown, after your search and sort.

Details: Sorting and searching a table.

Can I attach a photo to a question?

Yes, where your admin has turned it on: use the paperclip button beside the chat box, or paste or drop a picture on the box. The assistant reads the photo together with your question, for example a screenshot of an error message. It is photos only (PNG, JPEG, WebP or GIF, up to 4 a question and 10 MB each), not documents. If you do not see the button, your admin has not turned photos on for you, or the model you chose cannot read pictures: photos work today only with the AI included with SourceLace and your organization's own Anthropic key, not with OpenAI, Azure OpenAI, Amazon Bedrock, Google Gemini or Mistral. The photo goes to the AI provider, and your data protection rules cannot check what is in a picture, so show only what you need.

Details: Attaching photos.

Why does a step say it was corrected?

The assistant works in steps, and sometimes a step fails, for example because a query named a column that does not exist. The assistant then tries again with a better query, and the answer is right. A step is called corrected only for a mistake like that, or a timeout, when a later query of the same kind on the same system worked; any other failure stays red. The line above the steps says (1 step corrected along the way) and the step that failed is in its place in the list, in grey, with a plain reason and The assistant corrected this and carried on. It is not an error you need to act on. A step that nothing fixed stays red and says (1 had a problem, shown below), and so does a refused sign-in or permission, a limit, a failed change or file, or a refusal by data protection, even when something else worked afterwards: check that the answer does not rely on it. The system's own message is behind Details.

Details: When a step did not work.

What does Redo on all rows do?

It asks the assistant, in the same chat, to do the latest answer again on all the rows it can read, fresh from the system. You see it next to a note that a table is built only from the first rows of an earlier table (for example the top 8), such as a total of them or the same rows in another order. Choosing it adds "Do that again on all rows, read fresh from the source." to the chat as your own message, and the answer appears below. It reads the system again, so it takes as long as a new question and counts as one, and it uses your own access and every limit, cap and rule that applies to your questions. Nothing already in the chat is deleted. It is not Ask live, which reads one saved answer again from the system; Redo on all rows asks the whole question again.

Details: A table built the way you asked.

Can I choose whether my next question uses the data already read?

Yes, after an answer that has a table still held. A small switch beside the send box offers Use the data already read (the usual: quick, and your systems are not asked again) and Read the systems again (asks them again for your next question only, so you see their current data). The choice is not remembered: after you send, it goes back to the usual one. Details: Where your next answer starts.

What does "Answered from" mean in the answer check?

It says what the answer used: the systems read just now, the table held from your earlier answer (with when it was read), or saved answers (with their age). Press Show under the Answer check line to see it. Details: Where an answer came from.

Why is a table folded under "Built from"?

An answer shows the tables you asked for. A table the assistant only read on the way, for example the raw rows it then reshaped into the table you asked for, folds under Built from N tables, even if the assistant first marked it as an answer. Open Built from to see them; nothing is lost. If the answer has only one table, it is always shown. Your admin can show every table instead under Chat defaults, and you can choose for yourself under Your settings.

Details: Tables in answers.

How do I search or sort a list?

Every long list has a search box above it, and its column names are buttons: click one to sort by it, click again for the other way (a column whose name is not a button does not sort). On lists that load a page at a time, the sort and the search cover every row, not only the page you see. Copy table copies the rows you see, in the order you see them. SourceLace does not save what you type or the order you choose. One exception: when you export the Audit log, the filters you used, including a search word, are written to the audit trail like any other entry (it is encrypted and kept for your plan's audit history). On People, Runs, a shared agent's Subscribers, the Audit log (its search and its Person, Action, Source and Object choices) and the chat and project searches in the sidebar, which can hold names and email addresses, the words you type and the person or object you pick stay in the page and are not put in its address, so a refresh starts with an empty box and nothing chosen. On the other lists that have filters above them (Agents, Skills, Groups and similar) the words are kept in the page's address, so they can show in your browser history, in a link you copy and in SourceLace's server records, which keep addresses for 365 days. On every other list they reset when you leave the page.

Details: Searching and sorting lists.

All FAQ categories