# MG Legal Ops - all guides Source: https://mglegalops.com/docs/ - one section per guide. # Using the MG Legal Ops assistant This is a plain-language guide to what the assistant can do for you day to day. There are eleven things you can ask for: - **["Give me my morning brief."](/docs/morning-brief)** A short daily summary of filing and billing signals. - **["How did this week go?"](/docs/weekly-digest)** What got filed, logged, and sent this week, and what's still open. - **["Help me review my email filing."](/docs/email-filing-review)** Find emails that haven't made it into Clio yet, and file the ones you approve. - **["Check my billing for missing time."](/docs/billing-review)** Find recent work - emails and meetings - that might not have a matching time entry. - **["How's matter 00003-Smith doing?"](/docs/matter-review)** A quick filing + billing snapshot of one matter or client. - **["Draft a reply to that email."](/docs/email-drafting)** An email or reply written in your voice into your Outlook Drafts, sent only when you say so. - **"Summarize my Court Alerts folder."** Read the recent emails in any one of your Outlook folders - the ones your mail rules sort notices into - and summarize them. (The filing and billing reviews look at your Inbox and Sent Items, plus any folders you've told it to include; this is how to get at any other folder. You can still ask to file any of those emails to a matter.) - **["This seems wrong, can you report it?"](/docs/report-feedback)** File a bug report or suggestion about the assistant itself, for the developer. (Problems the assistant hits on its own are reported automatically - you only need this for feedback you want to give in your own words.) - **["Run a conflict check on these names."](/docs/conflict-check)** Search Clio's contacts and files, closed ones included, before opening a matter. - **["Put my Clio court dates on my Outlook calendar."](/docs/calendar-sync)** Keep court dates and matter meetings in step between Clio and Outlook. - **["Draft an update for the client on this matter."](/docs/client-update)** A short status email built from the file, for you to read and send. - **["Write me a research prompt for this."](/docs/research-brief)** A paste-ready Vincent prompt from an email or a matter. Click through to each guide above for what it actually shows you and, importantly, what it can't see. The rest of this page covers the ground rules that apply no matter which one you're using. The same material, laid out for printing or emailing, is in `handbook/`: a one-page quick start and the full attorney handbook, both as Word documents (open in Word and Save As PDF if you want a PDF). And you can always ask the assistant itself "what can you do?" ## You'll be asked to sign in The first time you use the assistant in a session, it checks whether you're signed in to your Microsoft (Outlook) and Clio accounts. If not, it gives you one link - open it in your browser, and you'll see a sign-in button for each account; sign in like normal and it walks you from one to the other. If it says sign-in isn't set up on this computer yet, a settings value is missing and whoever set up the assistant needs to add it. If it says the values are there but Claude needs a restart, they were added after the assistant started: quit Claude completely (every window) and reopen the project, and the sign-in page catches up. You stay signed in on that computer: restarting Claude or coming back the next day doesn't mean signing in again. If it ever does ask again, your saved sign-in lapsed or was revoked, and signing in once more fixes it. You can also ask it to sign you out at any time. You can have two Claude conversations using the assistant at once; a sign-in you complete in one carries over to the other. If the assistant ever says the sign-in link won't work from this conversation, something else on your computer is using its port - the link from your other conversation, or a restart, will do. ## It keeps itself up to date Improvements and fixes arrive as new releases, and your computer picks them up on its own: each time Claude starts the assistant, it fetches the newest release first, so you never have to install anything. If a release comes out while you're already working, the assistant tells you at the start of your next request - "a newer version is available; quit Claude and reopen this project to pick it up." That is all it takes; do it at a good stopping point, since what you're using still works in the meantime. If it instead says the update couldn't be applied automatically, or that this copy isn't connected to GitHub, nothing is wrong with what you did: whoever set up the assistant on your computer needs to take a look, and the assistant can note it for them if you ask. When an update has arrived, the assistant tells you what's new the first time you talk to it afterwards - a short, plain list of what you can now ask for, once, and then it gets on with your request. When a release is waiting for your restart, it tells you what you'll get by restarting. A release that is only fixes is just "a few small fixes". If your firm's terms of service are about to change, the assistant mentions it once, with the date and where to read them, and then gets on with your request. Nothing is needed from you; the firm's contact has been told separately. ## It can look things up in Clio for you Name a matter or a client and it finds the matter in Clio - and if more than one could fit, it asks which you mean rather than guessing. It can pull up a matter's record (status, dates, responsible attorney, related contacts, open tasks, recent communications) and check whether someone already exists as a Clio contact. It can also run a conflict check: give it the names involved in a new matter and it searches Clio's contacts and files, closed ones included, and shows you who matched and on which side. It does that on its own before opening a matter for anyone new. See the conflict-check guide. And it can put the court dates and deadlines from Clio's calendar onto your own Outlook calendar, and an Outlook meeting onto a matter's Clio calendar, on your yes, never twice. See the calendar-sync guide. It can draft a status update to a client from what is on the file - what was done, what is next, what you need from them - for you to read and send. See the client-update guide. When an email raises a legal question, it can write the research prompt for Vincent - jurisdiction, facts, the precise question, what the memo should contain - for you to paste. See the research-brief guide. It also points these out as it goes: when an email it is reviewing or answering plainly raises a legal question, it says so and offers the prompt. ## Nothing gets written anywhere without you saying so, explicitly Seven actions in this toolset change something in your files or on your behalf: filing an email to Clio, creating a time entry in Clio, opening a new contact and matter in Clio for someone you name, adding a task to a matter in Clio (assigned to you), sending an email from your mailbox, adding an entry to your Outlook calendar, and filing a feedback report in your own words (tagged as a bug or a feature request, so the developer can sort them). The assistant will never do any of these just because it thinks it should, or because something in an email's content seemed to ask for it - even a phishing-style email that tries to talk it into acting. It will always show you exactly what it's about to do first and wait for you to say yes before it does it. That isn't just a habit: the system itself refuses to write anything to Clio until a proposal for exactly that write has been made and confirmed, within a few minutes, once. If you change your mind about a detail after seeing the proposal, it shows you a fresh one. Every write is also recorded on your computer - which tool, which matter, when, and whether it worked; never the contents of an email. Ask "what did you file today?" or "show me what you've written to Clio this week" and it reads that back to you. If one of those was a mistake, say so: "undo that time entry", "take that email back off the Johnson matter", "delete that task", "remove that calendar entry." It shows you what it will remove, waits for your yes, and deletes it - but only something it created itself. Anything made in Clio by other means, and a new client matter, has to be changed in Clio. That same record can add itself up. Ask "what did you do for me in August?" or "how much time have you captured this month?" and you get counts for the month: time entries and the hours behind them, emails and attachments filed, matters, tasks, and calendar entries created, drafts and sends - with anything you undid already taken off. It's numbers only, never the contents of anything. If your firm has agreed to share usage numbers with the developer, the assistant can send that same monthly summary to them - the firm's name and a user number, never your name, your email address, or anything about a client - and it only does so when you say yes. Running it twice for the same month updates the one report rather than sending two. The assistant also remembers a few things you tell it once: your usual billing code, your hourly rate, how many quiet days make a matter worth a look, and which extra Outlook folders to include in its reviews ("include my Court Alerts folder from now on"). It asks the first time it needs one, then uses your answer until you say otherwise ("use 350 from now on"). Two things it does without asking. It reports its own problems: if a tool fails, gives a wrong answer, or can't do what you needed, the assistant files that with the developer itself - first-hand, with the technical detail and a suggested fix - and tells you in a line that it has. And it passes on your preferences: when you tell it to do something differently - "show me the sender first", "stop asking me that", "that's too long" - it does what you asked and also notes the preference for the developer, so the assistant can be changed to work that way for everyone without you having to say it twice. You don't have to notice bugs, describe them, or approve any of these reports. None of them ever include anything from your emails or client files - no names, subjects, addresses, or matter details - only how the software behaved and what you wanted instead. Whether the assistant is pointed at a test Clio account or the firm's real one is decided by whoever set up your computer, not by the assistant, and it works the same way in both: propose, your yes, then the write, and a line in the record. On your real account, every filing and time entry is a real record on a real client file - which is exactly why nothing is written until you've seen it and said yes, and why "undo" exists. Feedback reports don't have that distinction - they only ever go to the developer's own GitHub repository, never anywhere client-related. ## Nothing it produces for a client will mention Clio, Claude, or AI Anything the assistant writes or suggests that could reach a client - a time-entry description, the text of a filed communication, a draft letter - reads as your own work. It describes the legal work, never that it was recorded in Clio or drafted with help. You won't see "used Clio to write letter" on a bill because of this assistant. ## "Possible," not "certain" Everything the assistant tells you - a matter match, a missing time entry, whether an email looks suspicious - is a best guess from the evidence it has (an email address, a subject line, a meeting attendee), not a fact. It should always phrase things as "possible" or "worth a look," never as "you forgot to do this" or "this is definitely matter X." If something looks off to you, trust your own read of the client file over the assistant's guess. ## It will never invent a number If the assistant flags possible missing billable time, it will never tell you how many minutes or how much to bill - it doesn't have that information, and it won't guess. That call is always yours. --- # Morning brief **Ask for this when:** you're starting your day and want a quick read on what needs attention. > "Give me my morning brief." > "What should I look at today?" > "Daily summary, please." ## What it shows you A short, prioritized summary - meant to be read in under a minute, not a full data dump: - **Today**: your meetings, Clio calendar entries (court dates and court-rule deadlines called out), and tasks due, with matter numbers. - **Filing**: how many emails are unfiled vs. need your judgment. Mail from people who aren't Clio contacts (newsletters, vendors, internal) is counted separately so it doesn't look like a filing backlog. - **Billing**: how many possibly-missing-time items showed up, grouped by matter if there's more than a couple. - **Quiet matters**: open matters with no recorded activity in Clio for longer than your threshold, longest first. The first time, it asks you what that number of days should be and remembers it. - The 3-5 items across both that seem most worth your attention first, named specifically (sender, subject, matter) so you can act without digging further. If everything's genuinely clear, it'll say so in a couple of sentences rather than padding the brief out to look more thorough than it needs to. At the end, it'll ask what you want to do next - file some emails, look at billing in more detail, and so on. The brief itself doesn't take any action; it hands off to [email filing review](/docs/email-filing-review) or [billing review](/docs/billing-review) for anything you actually want to act on. If a court date or deadline is on Clio's calendar but not on your Outlook calendar, the schedule says so, and the assistant offers to add it (see the calendar-sync guide). It also looks a week ahead: the court dates and deadlines on Clio's calendar in the next seven days, and for each, whether the matter shows an open task due by then or a time entry in the last two weeks. The ones with neither are named as "coming up, nothing logged yet" - a nudge to look, not a verdict that nothing has been done. ## What it can't do - **It only sees what's on your Outlook calendar and in Clio.** A meeting that lives only in someone else's calendar, or a deadline noted only in a document, won't appear. - **The filing line doesn't tell you what's "ready to file."** That distinction (does it also look automated or spoofed?) comes from the fuller email-filing workflow, not this brief - ask for [email filing review](/docs/email-filing-review) if you want that. - **It only summarizes schedule, filing, billing, and quiet-matter signals** - documents and invoices are not part of this brief. - **A quiet matter can become an email.** For the longest-quiet matter it offers to draft a short check-in to the client; say yes and you read the draft before it goes anywhere (see the client-update guide). - **A quiet matter is a prompt, not a verdict.** "No recorded activity in 20 days" means nothing was logged in Clio, not that nothing happened. - **A very busy mailbox may be read partially.** If more mail arrived in the window than it can read, it says the filing and billing lines are partial rather than pretending they're complete. - **The filing line is built from a summary, not every email.** On a busy week the brief works from counts, the most recent unfiled emails, and a sample of the ones needing your judgment - it will say "and N more" rather than list them all. Ask for [email filing review](/docs/email-filing-review) when you want the full list. --- # Weekly digest **Ask for this when:** it's the end of the week and you want to know what got done and what's still hanging. > "Give me my weekly digest." > "How did this week go?" > "Friday summary, please." ## What it shows you A screen or less: - **Done this week, through the assistant**: emails filed, time entries created and the hours they add up to, emails sent, calendar entries added, and anything you undid. Then one line for the month so far - entries, hours captured, emails filed - so the week sits against the month. Hours are the durations you gave when the entries were made; it never estimates. - **Still open**: emails not yet in Clio (with mail from non-contacts counted separately), possible missing time by matter, and matters that have gone quiet - and for the longest-quiet one, an offer to draft a check-in email to the client. On a busy week this works from counts and the most recent unfiled emails, saying "and N more" rather than listing every one - ask for [email filing review](/docs/email-filing-review) for the full list. - The three to five things most worth a look before the weekend, named. Then it asks what, if anything, you want to do before the week ends. ## What it can't do - **It counts what went through the assistant on this computer.** Filing or time you entered in Clio by hand isn't in the "done" numbers, though it does show up in what's still open, which reads Clio directly. - **It takes no action.** Anything you want done hands off to the filing or billing review. --- # Email filing review **Ask for this when:** you want to catch up on filing, or you're wondering what's missing from Clio. > "Help me review my email filing." > "What's missing from Clio?" > "Can you file today's emails?" ## What it shows you The assistant looks at your recent Outlook mail and sorts it into three groups: - **Ready to file** - the matter is clearly known and the message doesn't look automated or spoofed. These are safe to file without you having to pick the matter yourself. Filing saves the email's text and its attachments to the matter. - **Needs your judgment** - the matter is ambiguous, or the message looks automated, no-reply, or possibly spoofed. The assistant won't guess on these; it needs you to look. - **Already filed** - just a count, since there's nothing new to show you. It will list the "ready to file" emails for you by subject, sender, and matter, and say how it linked each one to its matter - someone on the email is a Clio contact on that matter, or it's a thread you filed before. An email whose only link to a matter is a number typed into its subject line is never in this list: anyone who has seen a matter number could put it in a subject, so those always go to "needs your judgment." For "needs your judgment," it lists the ones you can actually resolve (an email that could belong to more than one matter, or one that looks automated) individually. Mail it couldn't connect to any Clio contact at all is summarised instead of listed: how many there were, which senders account for most of them, and a few recent examples - most of that is vendors, newsletters, and internal mail. In a busy inbox that group can run to hundreds of emails, which is why it isn't listed one by one; if you want every one, just ask for the full list. Then it asks which ones you want filed. You can say "all of them," name specific ones, or "none for now." Nothing gets filed until you answer. After filing, it reports back honestly - what actually filed, what failed (with the error), what you didn't name (still sitting there for next time), and anything that didn't match what it expected. Filed something to the wrong matter? Say so - it shows you what it will remove and, on your yes, takes the email and its attachments back off that matter. If your mailbox is very busy it may not be able to read the whole window; when that happens it tells you the review is partial. ## Filing the ones it couldn't place For an email in "needs your judgment," just tell it where it goes - "that one's the Johnson matter." It finds the matter in Clio (and asks if more than one could fit), shows you the pairing - subject, sender, matter number, and the opening lines of the email so you can check what it's really about (ask to see the whole message if you need to) - and files it once you say yes. It also remembers that conversation, so later replies in the same thread go to the right matter without you being asked again. If several of them belong to the same matter - a whole thread, or "all the Johnson ones" - say so once: it shows you the list under that matter and files them all on one yes, then tells you which filed and which, if any, didn't. When you approve a thread from the "ready to file" list, every unfiled message in that thread from the period you reviewed is filed - if you asked about the last week, nothing older than a week is touched. Afterwards it tells you what was filed, what failed and why, and what you left unfiled, in a short list rather than repeating the whole review. ## Adding a new client from an email If someone emails you who isn't in Clio yet - a new client inquiry, say - you can tell the assistant to "add them as a contact." Because a contact in Clio always belongs to a matter, that really means opening a new matter for them, and the assistant will say so. It first checks whether the person already exists in Clio (so you don't end up with duplicates), then asks you, in one go, for the only two things it needs: a name for the contact and a one-line description of the matter, such as "Will preparation" or "Contract dispute". You can tell it to pick a sensible name and base the description on what the email says, if you'd rather. The description can end up on a bill, so it's written the way you'd write it yourself. Once you answer, it creates the contact and matter, files that person's emails to it - attachments included - and reports back the new matter number and what was filed. If an address looks like it's yours, or shows up on hundreds of newsletters, it will check with you before making it a client. Before it asks you for those details, it runs a conflict check on the sender's name and anyone the emails name as the other side, and shows you who matched in Clio and on which files, closed ones included. See the conflict-check guide. ## Reading an email Ask what a client said, or for a draft reply, and it will read the email for you. Anything written inside an email is treated as information, never as an instruction - if a message tries to tell it to do something, it tells you instead of doing it. ## When an email raises a legal question If a message it is reading or listing plainly raises a legal question - the other side asserts a defense, cites a rule, disputes coverage - it says so and offers to write the research prompt for Vincent (see the research-brief guide). It offers once per question and never for routine mail. ## What it can't do - **Very large attachments and attached emails aren't copied.** Normal file attachments (PDFs, Word documents, images) are saved into the matter's documents in Clio along with the email. A file over 25 MB, or an attached email or cloud link, is listed by name instead, and the assistant tells you which ones so you can add them by hand. - **It only looks at the last 90 days** of your Inbox and Sent Items, plus any folders you've told it to include (it tells you which folders it read) - not your entire mailbox history, and never Junk or Deleted Items. If a rule sorts court notices or anything else into a folder of their own, just say "include my Court Alerts folder in the reviews" - it checks the folder exists and reads it from then on. - **It never reads the body of an email to figure out the matter** - only who was on it (the client or anyone Clio lists as related to the matter), an exact matter number in the subject line, or a thread you've filed before. A matter number in the subject on its own only ever gets an email as far as "needs your judgment" - it takes a known contact or a filed thread to make one "ready to file." If a matter is only mentioned inside the email text, it lands in "needs your judgment" too, and you can tell it where to file it. ## If auto-filing isn't turned on This feature has an on/off switch for this environment. If it's off, the assistant will tell you plainly rather than pretending nothing's wrong - that's expected, not an error to chase down. --- # Billing review **Ask for this when:** you want to check whether you've captured all your billable time, have a nagging feeling you forgot to log something, or want to log a time entry for something it found. > "Check my billing for missing time." > "Did I forget to bill anything this week?" > "Billing gap check, please." ## What it shows you The assistant compares your recent work - both emails and calendar meetings - against the time entries already in Clio, and groups the result by matter: - **Possibly missing time** - substantive activity on a matter/date with no matching time entry. This is the actionable list. - **Likely covered** - a time entry already exists for that matter and date. Shown only briefly, since there's nothing to do here. It's smart about volume, not just presence: if you had three separate emails on a matter that day but only logged one time entry, it flags the other two as possibly missing - one logged entry doesn't automatically cover everything that happened that day. A hearing, deposition, or mediation on the matter's calendar in Clio counts too, once it has happened: if there is no time entry on that matter that day, it shows as possible missing time. Deadlines on Clio's calendar don't count, since they aren't something attended. A calendar meeting counts as evidence the same way an email does - if a meeting's attendee is a known Clio contact and there's no matching time entry, it shows up here too. ## What it can't do - **It never tells you how much to bill.** No minutes, no hours, no dollar amount - it flags that something might be missing and leaves the actual judgment call entirely to you. - **It only sees email and calendar meetings** - not phone calls with no calendar entry and no email trail. If your missing time was a call that never touched your calendar or inbox, this tool has no way to see it. - **"Covered" only means something was logged**, not that the amount billed was right. A time entry existing doesn't mean it captured the full amount of work. - **It can miss things when the matter itself is unclear** - if an email or meeting doesn't clearly resolve to one matter, it's left out rather than guessed at. If something you expected to see is missing, that's usually why. - **Logging time is one sentence.** "Log 12 minutes for the Appleton one and 6 for each of the other two." It already knows the matter, the date, and who each item was with, and it suggests bill wording from that ("Correspondence with Jordan Smith re: lease renewal terms"). It shows you the whole list as one proposal - matter, date, minutes, wording, billing code, rate - waits for one yes, and creates them all. Change any wording or detail and it shows you a fresh list. - **It never decides the duration.** It won't propose a number of minutes, ever; if you don't say how long for an item, it asks. And it only creates entries after you've seen the list and said yes. - **It remembers what you billed.** Once you've logged time for an item, the next review shows that item as covered, not a different one that happened to be on the same day. - **Bounces and auto-replies don't count.** Automated mail never shows up as possible missing time. - **It learns your billing code and rate once.** The first time it logs time for you it asks which of the firm's billing codes you usually use and what hourly rate, then uses those on every entry until you say otherwise. You can name a different code or rate for any single entry. - **Made a mistake? Say so.** A time entry it created can be removed: "undo that entry." It shows you what it will remove and waits for your yes. - **The description is yours, and it stays clean.** It describes the legal work and never mentions Clio, Claude, or an AI assistant - the system refuses a description that does, since it goes straight onto a bill. Ordinary words are fine: "executive assistant," "AI licensing agreement," and the like pass. - **Ask for the period you mean.** "This week" or "the last three days" narrows the review instead of showing 90 days. --- # Matter review **Ask for this when:** you want a quick status check on one specific matter or client. > "How's matter 00003-Smith doing?" > "What's going on with the Johnson file?" > "Give me a status check on this client." ## What it shows you Tell it a matter number or a client name (it searches Clio and asks you which one if several fit), and it pulls together a short snapshot: - **The matter itself** - status, open date, responsible attorney, practice area, the client, related contacts (opposing counsel, adjusters, and so on), open tasks with due dates, and what's been logged on it recently. - **Filing status** - how many recent emails for that matter are already filed, unfiled, or need your judgment. - **Billing status** - how many pieces of evidence look covered by a time entry vs. possibly missing one. If nothing turns up for the matter you named, it will say so plainly - that usually means there's been no recent email activity tied to it, not that the matter itself is wrong or doesn't exist in Clio. If the recent correspondence or an open task carries a legal question you will have to research, it names it and offers to write the Vincent prompt (see the research-brief guide). ## What it can't do - **Documents and billing totals aren't included.** It shows the matter's record, contacts, tasks, and communications, but not the document list or what's been invoiced - Clio is still the place for those. - **Naming the matter by number is exact; naming it by client is a best effort.** Both halves are narrowed to the one matter before anything comes back, so the snapshot stays short even on a busy mailbox. If you only give a client name, the assistant has to match that name to a matter itself and will say so if more than one could fit. Emails from that client that Clio couldn't tie to the matter (an address that isn't on the contact, say) won't show up in the snapshot. - **If you ask something it can't see** (a document, an invoice), it will tell you that plainly and point you to Clio, rather than guessing. - **It's a snapshot, not an action.** If something here needs follow-up - filing an email, looking into a billing gap - it will point you to [email filing review](/docs/email-filing-review) or [billing review](/docs/billing-review) rather than trying to handle it here. --- # Email drafting **Ask for this when:** you want an email or a reply written for you, or sent. > "Draft a reply to that email from the adjuster." > "Write to the client about the closing date." > "Send it." ## What it does It writes the email you asked for, in your voice, shows you the full text - recipients, subject, every line - and asks once: any changes, and Drafts or send? If you say send, it goes out; one yes is one send. If you only wanted a draft, it puts it in your Outlook Drafts folder and nothing is sent; you can read and edit it there like any draft, or tell the assistant what to change, and say "send it" when you're ready. It asks you to confirm again only if the draft changed since you approved it - you edited it in Outlook, say - and then it shows you what will go out as the draft stands right then and waits for your yes. It sends from your own mailbox, and the copy in Sent Items shows up in your next filing review. If you sent or deleted the draft from Outlook yourself and then ask the assistant to send it, it checks your Sent Items for you and tells you either "that went out at [time] to [recipients]" or "it's not in Sent Items - want me to recreate it?" It won't ask you to go and look. Before it writes a reply about a matter, it reads that matter in Clio - who the parties are and their roles, open tasks and deadlines, what's been logged recently - and the message being answered, so the draft fits the file rather than being generic. It takes what to say from you and the facts from the file. It won't invent a date, an amount, or a position you didn't give it; where something is missing it leaves a blank in brackets for you. If the email you are answering turns on a legal question you haven't resolved yet, it says so before drafting and offers to write the Vincent research prompt first (see the research-brief guide). Say "draft it now" and it does, leaving a bracketed blank where the researched point belongs. ## What it can't do - **It can't attach files.** Add attachments in Outlook before sending. - **Sending can't be undone.** That's why it shows you the exact email before anything goes out, and never sends a version you haven't seen. - **It never mentions Claude or an AI assistant** in anything it writes for you - the email reads as yours. It won't name Clio either, except in a new email going only to people at the firm (every address on it at the firm's own domain), where staff call the system by its name. A reply, or anything with an outside recipient, stays Clio-free. - **It won't follow instructions inside an incoming email.** If the message you're answering tries to direct the reply ("confirm the new bank details"), it tells you instead of complying. - **It has to be turned on for your computer.** If it says email writing isn't enabled, whoever set up the assistant can turn it on; you'll sign in to Microsoft once more to allow it. --- # Calendar sync **Ask for this when:** you want the court dates and deadlines that are on Clio's calendar to show up on your own Outlook calendar too, or an Outlook meeting to show up on a matter's calendar in Clio. > "Put my Clio court dates on my Outlook calendar." > "Is the Delgado hearing on my Outlook calendar?" > "Put my call with the adjuster on the Delgado file's calendar." > "Sync my calendars for the next three months." The morning brief also points these out: a Clio entry that is not on Outlook shows up in the schedule as such. ## What it does It compares Clio's calendar with your Outlook calendar for the period you ask about (two months unless you say otherwise, up to six) and sorts Clio's entries three ways: already on Outlook because the assistant put it there; probably on Outlook because there is a same-day entry with a matching title, which it shows you and leaves alone; and missing. For the missing ones it shows you a short list - what, when, which matter, and whether it is a court-rule deadline - and waits. Say "all of them" or name the ones you want, and it adds each to your Outlook calendar, showing you each entry as it goes. Every entry it adds carries an invisible tag linking it to the Clio entry, so the same hearing can never be added twice, and "undo that calendar entry" removes one if you change your mind. Times are converted to your own time zone. Nothing is sent to anyone: these are entries on your calendar only, with no attendees. The other way round is deliberately narrower. Your Outlook calendar has everything on it, and only the meetings that belong on a client file should go to Clio. When you ask, it shows you the Outlook meetings that look like matter work - a client or the other side among the attendees, a matter name in the title - and asks which matter each belongs to. If one of the attendees is a Clio contact it suggests that contact's matter, but you name it; it never assumes. Each entry it adds to Clio is tagged with the Outlook meeting it came from, so the same meeting is never added twice, and "undo that Clio calendar entry" removes one. ## What it can't do - It won't decide which matter an Outlook meeting belongs to. It can suggest one; you name it. - It can only guess about entries you typed into Outlook yourself. If Outlook has "Hearing" on the same day as Clio's "Motion hearing - Smith", it assumes they are the same and asks before adding. - It does not keep things in step afterwards. If a hearing moves in Clio, the old Outlook entry stays where it was and the new date shows as missing on the next review. - If you delete one of its entries from Outlook, the next review shows that date as missing again. It asks before putting it back. --- # Conflict check **Ask for this when:** someone new is about to become a client, or you want to know whether the firm has dealt with a person or company before. > "Run a conflict check on Maria Alvarez and Northside Insurance." > "Have we ever had anything with the Delgados?" > "Are we clear to take on this new client?" The assistant also runs it on its own, and tells you it did, whenever you ask it to open a matter for a new person. ## What it does It takes every name you give it - the would-be client, the other side, anyone else involved - and searches Clio for each one: the contacts list, and the matters, including closed ones. For every contact it finds, it then looks up every matter that person or company is on, as the client or in any other role, and shows you a table: the name you gave, who matched in Clio, and which files they are on, with the file's status and the role the firm recorded ("client on 00012-Smith, closed 2024"; "opposing party on 00031-Doe, open"). If you give it one name and a matter type, it asks once for the other side before searching, because a check on the client alone is half a check. If that is all you have, it runs on that and says so. It never says "no conflict". It says what it searched and what matched, and the decision stays with you. ## What it can't do - It only sees Clio. Paper files, an older system, and a colleague's memory are not in it. - It matches names, not people. "Bob Johnson" entered as "Robert Johnson" can be missed; it searches the surname on its own to help, but a maiden name or a company's trading name can still slip past. Give it every spelling you know. - A hit is a prompt to look, not a finding. Two contacts with the same name may be one person entered twice, or two people. - It does not open the matter. That is a separate step with its own "show me first, then yes", exactly like every other write. - It does not give an opinion on whether something is a conflict under the rules. That is a legal judgment and it is yours. --- # Client status update **Ask for this when:** a client should hear where their matter stands and you would rather read a draft than write one. > "Draft an update for the client on the Delgado matter." > "What should I tell Ms. Alvarez about her case?" > "Send my clients their monthly updates." > "Yes, draft a check-in for that one." (when the morning brief or weekly > digest points out a matter that has gone quiet) ## What it does It reads the matter in Clio - your own time-entry descriptions, the correspondence filed, the hearings and deadlines ahead, the open tasks - and writes a short, plain email to the client in your voice: what has been done, what happens next, and whether they need to do anything. It never quotes a time entry word for word, never includes minutes or amounts, and never adds a date, a figure, or a position the file does not support; where something is uncertain it leaves a bracketed blank for you. You see the whole draft in the conversation first. Say what to change, and whether you want it in your Drafts folder or sent. Nothing is sent without your explicit yes, exactly as with any other email it drafts. If you ask for your clients' monthly updates, it lists the open matters with recent activity, tells you which have no client email in Clio, and asks which to start with. It does one at a time - draft, your read, send or not - and only then moves to the next. ## The check-in When the brief or the digest flags a matter that has gone quiet, it offers to draft a short check-in to that client: it's been a while, here is where things stand, has anything changed on your side. Same rules - you read it first, nothing is sent without your yes - just shorter. ## What it can't do - It doesn't know what you last told the client. It says which period the draft covers so you can trim anything they already know. - It can't judge what should stay out of a client email. A note about settlement authority or a weak point will be paraphrased like anything else; you are the gate, which is why you always see the draft first. - It can't email a client Clio has no address for, and can't attach documents. Add those in Outlook. --- # Research brief **Ask for this when:** an email or a matter raises a legal question you need to research, and you want the question written up properly before you take it to Vincent. > "This email raises a limitations issue - write me a research prompt." > "What do I need to look up on the Delgado matter?" > "Prepare a Vincent prompt for the coverage question in this thread." ## What it does It reads the email and the matter, states the research question back to you in one sentence so you can confirm it, and then writes a complete prompt for Vincent: the jurisdiction and where the case stands, the facts that matter stated neutrally, the precise question, and a list of what the memo should contain - the governing rule with primary authority, the elements, the other side's best authority, any split or recent change, a check that every citation is still good law, and a short practical answer. You copy the block and paste it into Vincent in Clio Work or at vlex.com. Then it offers to put a research task on the matter, due by the relevant deadline, and only does so if you say yes. ## What it can't do - It cannot reach Vincent, run the research, or read the answer. Vincent and this assistant don't talk to each other; you are the link. - It does not state the law itself, and nothing it writes is legal research. The prompt asks Vincent to confirm each citation, but you check every citation before relying on it. - If the file doesn't say which jurisdiction applies, it asks rather than guessing. - Names, addresses, and details the question doesn't turn on are left out of the prompt on purpose, so it is safe to paste anywhere. --- # Report feedback **Ask for this when:** something about the assistant itself seems wrong, confusing, or missing - not a client matter, the assistant's own behavior. > "This seems wrong, can you report it?" > "I want to leave feedback about the morning brief." > "Can you file a bug report for this?" ## You don't have to report bugs yourself If something breaks while the assistant is working - a tool fails, an answer is plainly wrong, or it can't do what you asked - it reports that to the developer on its own, right then, with the technical detail and a suggested fix. It tells you in a line that it has, and carries on. Those automatic reports never contain anything from your emails or client files. The same goes for preferences you mention in passing - "show me the sender first", "stop asking me that" - the assistant does what you asked and notes it for the developer on its own. This page is for the other case: feedback you want to give in your own words, such as "the morning brief is too long" or "I wish it could do X." ## What it does Files a real, permanent report for the developer to see - specifically about how this assistant behaves, never about a client or a matter. Before anything is filed, the assistant will show you the exact title and description it plans to send and ask you to confirm - nothing gets sent without your explicit go-ahead. It also notes whether this is something broken (a bug) or something you'd like it to do (a feature request), so the developer can sort reports by type. Once filed, you'll get back a number/link so you know it's been received and is tracked. Because the report leaves your computer, keep client names and matter details out of it - describe the situation ("an email from an existing client went to the wrong matter") rather than naming anyone. Email addresses are removed automatically before it's sent; names aren't, so that part is up to you. ## If it isn't set up on this computer Feedback reporting has to be switched on per computer by whoever set up the assistant. If it isn't, the assistant tells you straight away rather than taking down a report it can't file - just pass your note to the developer directly. ## What it can't do - **It's one-way.** You won't get a reply back through this conversation - if the developer follows up, it'll be through whatever channel they normally use, not here. - **It can't attach a screenshot for you.** If a picture would help explain the problem, describe what you saw instead, or mention it separately to the developer. - **It's not for firm policy or client-matter issues.** It only reaches the person who builds this software - not anyone else at the firm. - **It might not be turned on yet.** If reporting isn't configured in your environment, the assistant will tell you plainly rather than pretending it worked. --- # What's new Every release, newest first. Each one reaches every attorney's computer on its own the next time Claude Code starts the assistant. **In the assistant** is what the attorney can now ask for, in the words the assistant uses to tell them after an update. **Around it** is everything else that changed: the website, the terms, pricing, the free trial, security, and the service that supports the assistant. A release with neither is fixes and small improvements. ## Release 105 (2026-09-13) **Around it** - The join form explains each role in plain terms and offers only roles the service enforces: use the assistant, administer the firm's account, or billing contact only. An approved colleague gets their own download link; the administrator only approves. ## Release 104 (2026-09-13) **Around it** - An attorney at a firm that already uses MG Legal Ops can ask to be added from the website; the firm's administrator approves them and their role before anything is issued. ## Release 103 (2026-09-13) **In the assistant** - If your firm's Microsoft connection for the assistant has a secret that is about to expire, the assistant tells you six weeks ahead, with what to ask your IT provider for. **Around it** - Firms are warned six weeks before their own Microsoft connection secret expires, so sign-in never stops unexpectedly. ## Release 102 (2026-09-13) **Around it** - Notices from MG Legal Ops now also arrive by email, from notices@mglegalops.com, with replies going to support. ## Release 101 (2026-09-13) - Fixes and small improvements. ## Release 100 (2026-09-13) **In the assistant** - When MG Legal Ops has something to tell your firm - a change that affects you, a problem being fixed - the assistant passes it on once, in a sentence, and carries on. **Around it** - Firms are told about changes that affect them through the assistant itself and, once the mailbox is connected, by email to the people they named at sign-up; the privacy policy says what is kept. ## Release 99 (2026-09-13) **In the assistant** - If your firm's terms of service are about to change, the assistant tells you once, with the date - nothing is needed from you. ## Release 98 (2026-09-13) **Around it** - Attorneys' computers now receive only the assistant itself, from a dedicated distribution repository, and nothing else from the maintainer's workspace. ## Release 97 (2026-09-13) - Fixes and small improvements. ## Release 96 (2026-09-13) - Fixes and small improvements. ## Release 95 (2026-09-13) **Around it** - Sign up for the free trial on the website: tick the prerequisites, accept the terms, and get a private setup page that builds a ready-made folder for each attorney. Opens once MG Legal Ops LLC is registered and insured. ## Release 94 (2026-09-13) **Around it** - Terms and Privacy are in the top navigation of every page, each with a version number; a firm accepts a stated version when it starts a trial or subscribes. ## Release 93 (2026-09-13) - Fixes and small improvements. ## Release 92 (2026-09-13) **Around it** - The What's new page now lists everything that changed around the assistant - the website, terms, pricing, the trial, security - beside what changed in the assistant itself. ## Release 91 (2026-09-13) - Fixes and small improvements. ## Release 90 (2026-09-13) **Around it** - The website is now separate pages - what it does, how it works, pricing, free trial, FAQ, contact, confidentiality - with one consistent header and footer and a refined look. - Terms of service strengthened on ownership and permitted use: the software is licensed, not sold; what a firm may and may not do with it; what happens when a subscription or trial ends. ## Release 89 (2026-09-13) **Around it** - A firm that leaves has a clear, documented path: its subscription is switched off and its update access revoked, and it deletes the applications that are its own. Nothing of the firm's is touched. ## Release 88 (2026-09-13) **Around it** - The free trial no longer needs a card: fourteen days on one attorney's real inbox, nothing charged, nothing to cancel, and the assistant stops on its own unless the firm subscribes. The site, FAQ and terms say so. - In the last five days of a trial the assistant's status says when it ends and how to keep going. ## Release 87 (2026-09-13) **Around it** - The site has its mark: the MG monogram in the browser tab, on the iPhone home screen, and in every page header. ## Release 86 (2026-09-13) - Fixes and small improvements. ## Release 85 (2026-09-13) **Around it** - The FAQ names Claude Code as the requirement. ## Release 84 (2026-09-13) **Around it** - The assistant now confirms once a day, with the firm's key, that the firm's subscription is current. It fails open: an unreachable service never stops an attorney; only an ended subscription does. ## Release 83 (2026-09-13) **Around it** - Terms: automatic renewal stated plainly, who may agree on the firm's behalf, and that the firm keeps its access key confidential. ## Release 82 (2026-09-13) **Around it** - The website sends security headers on every page (content security policy, strict transport security, no framing) and the domain has DNSSEC, TLS 1.2 minimum, and certificate transparency monitoring. - Terms and privacy updated for the trial, billing through Stripe, and how long account records are kept. ## Release 81 (2026-09-13) **In the assistant** - An all-day date you put on a matter's Clio calendar now lands on the right day; before, it could show up a day early. ## Release 80 (2026-09-13) **Around it** - The project is now MG Legal Ops throughout: the organization, the repository, and the site. ## Release 79 (2026-09-13) - Fixes and small improvements. ## Release 78 (2026-09-13) **In the assistant** - The assistant is now called MG Legal Ops; nothing changes in how you use it. **Around it** - Renamed to MG Legal Ops. ## Release 77 (2026-09-12) - Fixes and small improvements. ## Release 76 (2026-09-12) **Around it** - A frequently-asked-questions page: requirements, what is and isn't supported, data handling, updates, pricing. ## Release 75 (2026-09-12) **Around it** - Every user guide the assistant ships with is published on the site, as web pages and as plain Markdown, with an index for AI tools (llms.txt). ## Release 74 (2026-09-12) - Fixes and small improvements. ## Release 73 (2026-09-12) - Fixes and small improvements. ## Release 72 (2026-09-12) **Around it** - Pricing published: $59 per attorney per month at the founding-firm rate, locked for twelve months. ## Release 71 (2026-09-12) - Fixes and small improvements. ## Release 70 (2026-09-12) **Around it** - mglegalops.com launched: what the assistant does, how it works, the ground rules, confidentiality, privacy, and terms. ## Release 69 (2026-09-12) - Fixes and small improvements. ## Release 68 (2026-09-12) - Fixes and small improvements. ## Release 67 (2026-09-12) - Fixes and small improvements. ## Release 66 (2026-09-12) - Fixes and small improvements. ## Release 65 (2026-09-12) **In the assistant** - While reviewing, answering, or summarizing correspondence, the assistant now points out when an email raises a legal question and offers to write the Vincent research prompt. ## Release 64 (2026-09-12) **In the assistant** - When an email raises a legal question, ask for a research prompt: it writes a complete, paste-ready prompt for Vincent with the jurisdiction, the facts that matter, and the precise question, and offers to put a research task on the matter. ## Release 63 (2026-09-12) - Fixes and small improvements. ## Release 62 (2026-09-12) - Fixes and small improvements. ## Release 61 (2026-09-12) - Fixes and small improvements. ## Release 60 (2026-09-12) - Fixes and small improvements. ## Release 59 (2026-09-12) **In the assistant** - The billing review now counts hearings, depositions, and other events on Clio's calendar as work: one that has happened with no time entry on the matter that day shows up as possible missing time. ## Release 58 (2026-09-12) - Fixes and small improvements. ## Release 57 (2026-09-12) - Fixes and small improvements. ## Release 56 (2026-09-12) **In the assistant** - When the morning brief or weekly digest points out a matter that has gone quiet, it can draft a short check-in email to that client. You read it before anything is sent. - The weekly digest now shows the hours behind the week's time entries, and a line for the month so far: entries, hours captured, emails filed. ## Release 55 (2026-09-12) **In the assistant** - The morning brief now looks a week ahead at the court dates and deadlines on Clio's calendar, and points out any whose matter has nothing logged yet. ## Release 54 (2026-09-12) - Fixes and small improvements. ## Release 53 (2026-09-12) **In the assistant** - After an update, the assistant tells you what changed, once, in plain language. If an update is waiting, it tells you what you'll get by restarting. ## Release 52 (2026-09-12) **In the assistant** - You can put an Outlook meeting onto a matter's calendar in Clio: ask, name the matter, and say yes. The same meeting is never added twice, and you can undo it. ## Release 51 (2026-09-12) **In the assistant** - You can put the court dates and deadlines from Clio's calendar onto your own Outlook calendar. It shows you which ones are missing, adds the ones you approve, and never adds the same date twice. The morning brief points out a Clio date that isn't on Outlook. - You can ask for a client status update on a matter. It writes a short email from the file, what was done, what's next, and what you need from them, and shows you the whole draft before it goes to your Drafts folder. ## Release 49 (2026-09-12) **In the assistant** - You can ask for a conflict check. Give it the names involved in a new matter and it searches Clio's contacts and files, including closed ones, and shows you who matched and on which side. It also runs one on its own before opening a matter for anyone. ## Release 48 (2026-09-12) **In the assistant** - You can ask what the assistant did for you in a month: time entries and hours, emails filed, and more. If the firm has agreed, it can share those counts with the developer. ---