Category: Product

Last week I promised I’d write every Friday, so here I am. This one is shorter than the first, but there’s a lot in it for the people who keep a firm running: the practice manager, the finance desk, the admin who sets everything up.

It covers 19 to 23 October. Most of it started as a message from a firm telling us what got in their way.

What’s new in ModulawAI, every Friday.

Send a form to anyone

If you try one thing this week, try this. A form shared with “Anyone with the link” now has a real page anyone can open and fill in, no ModulawAI account needed. Use it for client intake, a conflict check questionnaire, event registration, a feedback survey. The page leads with your firm’s logo and colours, so the person filling it in sees your firm.

The answers come back to you. Choose who the responses go to when you share the form, and they get told each time one comes in. The Responses tab on the Forms page lists every response you’re allowed to read, with who sent it, which form and when. Open one and file it to a matter, and its PDF lands in the matter’s documents.

To get the link out, click Email it and type up to 50 addresses with a short note. Each person gets an email in your firm’s name. On a phone, Share hands the link to WhatsApp or whatever else you use. When the answers are in, download them as a CSV, one row per response, or save that file straight to your documents.

Inside the firm, a form can now be started from My Tasks, so a leave request or an expense claim is one tap away for everyone.

“Help me understand this agreement, from my client’s side”

A lawyer told us that’s what they needed from a long facility agreement, and none of our tools did quite that. So the Word add-in has a new command. Type /brief acting for the borrower (or whoever your client is) and you get a briefing on the key terms. For each one: the clause in the document’s own words, what it says, what it means, whether it leans for or against your client, and two to four things to ask for in negotiation, each with a fallback. Click a quote and Word jumps to the clause. It ends with an overall position.

/review got better too. Whatever you type after it, “acting for the lender” or “focus on the indemnities”, now goes to the reviewer. If you had a paragraph selected, it tells you it only reviewed that paragraph and offers to review the whole document. And it reads far longer documents than before, past 2,500 pages.

One fix I’m glad to see gone. If a clause appeared more than once, Apply fix used to refuse and Jump to clause always went to the first copy. Both now pick the copy the issue is about, or the one you’ve selected.

Bill in your client’s currency

Until this week you could bill in seven currencies. Now it’s any currency in the world. A firm in South Africa can bill in rand, a firm in Kenya in shillings, and a Japanese client sees yen with no decimals, the way yen should look. Pick it from the currency list on a time entry, an invoice or your firm settings; type the code and it jumps there. South Africa, Kenya and Ghana are now in the country list too.

Who signs, who pays, who gets a number

This was a big week for finance teams, mostly because one firm sat down with us and walked through how their requisitions actually move.

Under Settings, Approval policy, there’s a new Requisition workflow card. You can have a review officer: the person raising a requisition picks a colleague who signs first, before the usual approval chain. You can name exactly who disburses (a person, a role, a department) with a backup for when it’s their own requisition. And you can switch off the Received By step if your advances end when the money goes out. Leave the card alone and nothing changes.

Vendors are now an approved list. Anyone can propose a supplier, but only approved vendors can be put on a bill, a purchase order or a requisition. Admins approve or reject (with a reason), and changing an approved vendor’s bank details sends it back for approval. The import picks up what each vendor supplies and where they are, and the register now pages, searches and filters, so a list of 300 suppliers is easy to work with. Every change is recorded.

Numbering caught up with how firms really file. Documents raised without a matter can carry their own code (ADMIN, FINANCE, whatever you use), and a requisition with no matter lets you pick which code it goes under. If you’re moving to ModulawAI halfway through a year, an admin can set a matter’s next invoice, PO, requisition or expense number from the matter form, so you carry on at 029 instead of starting again at 001. The numbering settings page also shows your saved formats properly now; before, it could show the defaults.

Removing someone keeps their history

When someone leaves the firm, removing them now archives them. Their name stays on the matters, time, documents and approvals they touched, their seat is freed, and they can’t get back in. If you removed the wrong person, or they come back, restore them from Archived.

Admins can also clear out process runs in bulk. If you trialled processes and have test timesheets and appraisals sitting in people’s My Tasks, tick them and cancel or delete them in one go.

Your notes, from the meeting you scheduled

When you recorded a meeting you’d scheduled in ModulawAI (a case meeting or a booking), the recording was filed but the notes never came. That’s fixed. Those recordings are now transcribed with no length limit, the notes land on the meeting and the matter, and everyone on the meeting gets them. If notes can’t be made, because the call was too short or nobody spoke, you get told why and where the recording is, rather than waiting on something that won’t arrive. We went back and processed the recordings that got stuck.

Guests see the file properly

People outside the firm you share a file with now open it in the same viewer your team uses. Spreadsheets, decks, emails, video and audio with the transcript, all of it. Before, some of those just said “not found”. If you gave them Comment access, they can comment on a cell, a slide, part of a page or a moment in a recording, and you see it in the same thread.

Your attachments go with the email

If you sent an email from the Inbox early this week and the attachments didn’t arrive, I’m sorry. The email went out without them and nobody was told. They go with the email now, and if a file can’t be attached for any reason, the email isn’t sent and you’re told which file was the problem.

The small stuff

• The Clerk’s matter brief reads like a note from a clerk now: dates in your time zone and plain words, no system codes.
• The Clerk now notices every document added to a matter, however it arrived: uploaded, filed from email, imported, sent by the client through the portal or saved from Word.
• On long tasks in chat, the assistant keeps working notes for that conversation, so a big piece of work holds together across several turns.
• On the renewal page, the partner code box looked like a required step. It’s now one line that says it’s optional, and it never changes your price.
• After editing a matter, “Opened by” no longer turns into “Unknown”.
• Line breaks in notification emails show up as line breaks.
• Form names stay readable on a tablet.

That’s the week. Everything here is live at app.modulaw.ai.

If something here is close to what you need but not quite, tell me. Reply to the newsletter or message us from inside the app. Half of this week’s list came from messages like that. See you next Friday.

Abiola
Founder, ModulawAI

We ship something most days, and most of it goes out quietly. Unless you went looking, you’d never know your timesheet stopped going blank, or that invoice PDFs stopped spilling onto an extra page. So from today I’m going to write here every Friday and tell you what changed. We’re calling it Lawtelligence.

I’m a week late starting, so this first issue covers two weeks, 5 to 16 October. It was the biggest fortnight we’ve had. Pour something warm; this one’s long.

What’s new in ModulawAI, every Friday.

A clerk for every matter

If you try one thing this week, try this. Open a matter and click once, and it gets its own AI employee. We’ve called it the Clerk. The Clerk keeps a brief of the matter and updates it when something happens on the file: a payment lands, a hearing moves, the client writes in. When you open the matter on Monday, the Clerk has already caught up.

The Clerk is one of a set. You can now switch on AI employees that react to what happens in your firm, as it happens. The Cashier watches invoices that get paid or go overdue. The Call Scribe picks up when a call ends. Hearing Prep, Portal Concierge and Inbox Triage do what their names say.

Then there’s the suggested timesheet. You did the work, so you shouldn’t have to rebuild it from memory at six on a Friday. ModulawAI lists what you drafted, who you called, which meetings you sat in and which emails you sent on each matter, and leaves out what you’ve logged. It costs no credits, because no AI model is involved.

Email that drafts its own reply

When a client emails about a matter, the person handling it now gets a task with a reply already written from the case file. You read it, change what you want, and click Approve and send. It goes from your own mailbox. Nothing leaves without a person approving it, and I’d keep it that way.

Every email in your inbox, and on a matter’s Comms tab, also gets a short summary, a guess at which matter it belongs to, and one tap to turn it into a task, a diary entry, a time entry or a note.

For those of you who build workflows: you can now add a condition, a step that waits for someone to approve, or a step that hands work to a named AI employee. Each workflow has an Activity tab, and you can run, check or pause one from chat.

That spreadsheet opens now

Someone sends you an Excel file, a PowerPoint deck, a forwarded .eml, a CSV. Until this week you had to download it. Now it opens in ModulawAI, and you can leave a comment on a cell or a range, a slide, a few words in a text file, or a line of code. A zip opens like a folder, so you can pull out the two files you need and leave the rest. All of it renders on our own servers.

Recordings got the same treatment. Play a video or audio file and the transcript runs underneath, with the line being spoken highlighted. Click a line to jump there, or comment on a moment. Ask AI to review the recording and it marks the admissions, contradictions, dates, sums and follow-ups at the second someone said them. If you take witness interviews, I think you’ll use this a lot.

Hand it the 200-page agreement

AI review now reads a draft, a PDF or a Word document of any length. It works through the document in parts and writes its issues onto the page as it goes, so you can start reading before it finishes. A card in the chat shows how far it has got and the top risks so far, and the document opens beside the chat when it’s done. The Word add-in uses the same engine, so long documents work there too, and a dropped connection no longer signs you out.

The draft editor also keeps your Word formatting now: indents, spacing, fonts, colours, numbering, merged table cells, headers, footers, footnotes and margins. There’s a ruler for dragging indents and margins. And Tab types a tab, the way Word does. If the editor ever swallowed your Tab key and threw you out of the document, I’m sorry.

Two smaller document changes. You can link a folder from your own OneDrive, SharePoint or Google Drive into your Personal files, and ModulawAI keeps it current while the files stay in your drive. Firms can link SharePoint libraries into the Library or a matter. And you can finally rename any folder, from any menu.

Contracts, start to signature

Start a contract from a template, draft it in the editor, send it for approval inside the firm, and send it out for signature with built-in e-signature. Nobody prints a page. The executed copy and its key dates stay on the contract, you can amend a signed contract, and clause review checks the terms against your clause library.

Has this been paid for?

A firm asked us that about a contract, and the document couldn’t answer. Now it can. Open a draft, contract or file and you’ll see whether the work on it has been invoiced and paid, amendments included. You can log time from the document and link invoice lines to it.

If your clients deduct withholding tax at source, you can record the rate, the amount that reached your bank and the credit note number on the payment. A live check tells you whether cash and tax add up to the invoice, and the receipt shows the split. Admins can also correct a payment that was recorded by hand. They have to give a reason, and we keep a record of every change.

Paid case invoices now send a receipt instead of the invoice again. Invoice PDFs lay out properly. And invoices and expenses follow your firm’s document numbering.

Forgot to press record?

Switch on automatic transcription, for the whole firm or for one meeting, and scheduled meetings get captured even when nobody presses record. Guests you send notes to now get a meeting page with the notes and transcript. They see the recording only if you choose to share it.

A few team things. Team Chat has voice notes (the mic sits beside Send) and shows everyone’s profile photo. Members show who they report to. Admins can let members start a process for themselves from My Tasks, a training request for example. And there’s the Floor, an optional live view of the firm: who’s working where, their workload and today’s activity. It stays off until an owner or admin turns it on, and everyone gets told when they do.

Research and chat

• Case-law sources in an answer now show how many judgments in our library cite them, and the most recent year. Hover for what the number counts.
• You can see what the assistant remembers about you under Settings, search it, and make it forget one thing or everything. You can also switch long-term memory off.
• The assistant reads the documents you attach before it answers, and it only cites authorities its research returned. We caught it inventing a citation in testing, and we fixed the cause.
• When it needs to know which matter or document you mean, it shows your real matters and lets you search or browse your files and connected drives.
• A shared chat opens the way its creator sees it, with drafts, sources and case cards intact.
• If your firm says “matter” or “file” instead of “case”, the assistant, notifications, PDFs, emails and reports now say it too.

Pick a theme

Settings, Appearance has six designs now, each with light and dark. Firms can set a default for everyone and a separate one for the client portal. The app moves more naturally too: selections glide, and menus grow from where you clicked. Pick one and tell me which.

If a client asks about data

Send them to our Trust Center, which went live this week. They can request our security documents and read them in a private portal. The short answer to most questions: the AI providers we use don’t keep what you send them, and we mask personal data in our own logs.

Behind that, we finished our ISO 27001 documentation, and automated compliance checks now run against our live systems every day. We also ran a full review of our security, line by line, and fixed what it turned up. Changing your password now signs you out everywhere else.

The small stuff

• Your clients’ portal timeline reads in plain words now, and your drafts stay with the firm.
• We tell you when someone adds credits to your account.
• A notification opens the exact record it’s about.
• We file email to a matter by its thread, and the case team hears about it.
• Read and confirm forms can search the whole Library and add whole folders.
• Share-link PDFs stopped showing blank half pages.
• Dialogs stopped sliding away when you tick a box. That was the timesheet “going blank”.

That’s the fortnight. Everything here is live or rolling out this week at app.modulaw.ai.

If something in ModulawAI annoys you, reply to the newsletter or message us from inside the app. Those messages come to me, and they decide what we fix next. See you next Friday.

Abiola
Founder, ModulawAI

Lawyers ask us two questions about the AI under ModulawAI more than any other: which model is it, and why that one. The answer to the first is Anthropic’s Claude, run through Amazon Bedrock so that client data stays inside our AWS environment and is never used for training. The answer to the second takes a little longer, and it is worth giving honestly, including the parts where a model on its own is not enough.

What Claude is genuinely good at for legal work

• Long documents. Claude reads very long inputs, an entire disclosure bundle or a two-hundred-page agreement, and answers questions about them without losing the thread. That is the capability behind tabular review, where one question is asked across every document in a set.
• Following instructions precisely. Told to answer only from the retrieved sources, to cite every proposition, and to say when the sources do not support an answer, it does so more reliably than most models we have tested. In legal work that discipline matters more than eloquence.
• Careful drafting. It produces first drafts that read like a competent junior wrote them: structured, hedged where hedging is right, and willing to flag an assumption rather than paper over it.
• Refusing to bluff, mostly. Asked something it cannot answer from the material in front of it, it is more likely than older models to say so. Mostly. Which brings us to the failures.

Where a raw model still fails a lawyer

None of the following is specific to Claude. They are properties of language models, and they are why “just use the chatbot” is bad advice for anything you will cite.

• It does not know your jurisdiction’s current law. A model’s training has a cut-off and no access to last month’s Court of Appeal decision or this year’s amendment. Ask it about the Evidence Act and it will answer from a general memory of evidence law, which may be a different country’s.
• Asked for an authority from memory, it can invent one. The case name will be plausible, the citation format will be right, and the case will not exist. This is the failure that has led to sanctions in several countries, and it is why an ungrounded model should never be the source of a citation.
• It does not know which matter you are in. A chat window has no idea who the client is, what was filed last week, or which precedent your firm prefers. Every answer starts from nothing.
• It cannot show its working unless you make it. A paragraph of confident prose with no links is unverifiable, and unverifiable is unusable in practice.

What ModulawAI builds around the model

The product is the part around the model, and it exists to remove each failure above.

• Retrieval before answering. When you ask a research question, ModulawAI first searches its own indexed corpus of primary law, judgments and statutes for the US, UK, Canada, EU, Nigeria, Ghana, Kenya and South Africa, and hands the relevant passages to Claude with an instruction to answer only from them. Every citation in the answer links to the source it came from, and the authority’s subsequent treatment is flagged. The model does the reading and writing; it is not the source of the law.
• The matter as context. Inside a case, the assistant knows the parties, the documents, the deadlines and the firm’s precedent library. A drafting request starts from your last good version of the document, not from a generic template.
• Structured outputs. Tabular review returns a grid, not prose, with each cell citing a passage, so the output can be sorted, checked and exported. Redlines in Word arrive as track changes so the lawyer accepts or rejects each one.
• Controls and an audit trail. Agent skills are scoped, critical actions can require a named approver, and every input, output and tool call is logged with the reviewer. That is what turns a capable model into something a firm can stand behind.
• Privacy by architecture. Claude runs through Amazon Bedrock inside our AWS environment. Prompts and documents are not sent to Anthropic’s consumer services, are not used to train any model, and are encrypted in transit and at rest.

The uncomfortable part: Claude has produced fake citations too

We would rather say this ourselves than have you find it elsewhere. In 2026 a US federal court sanctioned an immigration attorney who submitted fabricated legal quotations produced by Claude Sonnet 4 in an emergency habeas filing seeking to halt a client’s deportation. The model is not exempt from the failure mode described above, and any vendor telling you their underlying model does not hallucinate is either misinformed or selling.

The scale of the problem is now measurable. Damien Charlotin maintains a public database of court decisions involving fabricated AI-generated content, which recorded 2,041 cases as at 14 September 2026 across more than 40 jurisdictions, with 1,395 in the United States, 217 in Canada and 111 in Australia. The foundational case is Mata v Avianca, Inc., 678 F. Supp. 3d 443 (S.D.N.Y. 2023), a $5,000 sanction over six invented decisions. In England and Wales the Divisional Court in Ayinde v London Borough of Haringey and Al-Haroun v Qatar National Bank [2025] EWHC 1383 (Admin) made wasted costs orders of £2,000 plus VAT against each of two lawyers and required them to self-report to their regulators after five non-existent authorities were cited.

The lesson is not that the model is unsuitable. It is that the model is the wrong layer to hold the law. Every one of those cases involved a lawyer asking a model to recall an authority rather than asking a system to retrieve one. In ModulawAI the retrieval happens first and the model never supplies the citation, only the reading and the writing. That is an architectural choice, not a prompt, and it is the difference between the two workflows.

What your regulator expects of you, not of us

ABA Formal Opinion 512, issued 29 July 2024, is the first formal US ethics guidance on generative AI. It requires competence in the tools used, an assessment before entering client information as to whether it could be disclosed to or accessed by others, billing that reflects time actually spent, and supervision of non-lawyers and third parties using these tools. It also treats the cost of an AI subscription as firm overhead rather than a client disbursement. The SRA’s warning notice of 17 August 2026 states that solicitors remain accountable for all work and advice delivered whether or not AI was used, that reliance on an AI output would not be a suitable defence, and that client information should only be entered into AI systems where appropriate contractual, technical and organisational safeguards are in place.

Running Claude through Amazon Bedrock inside our AWS environment, with no training on client data, addresses the confidentiality limb. Linked citations address the verification limb. Logging every AI action against the fee earner who reviewed it addresses the supervision limb. None of that transfers the professional responsibility, and no product can. Sources for this section: the AI Hallucination Cases database, ABA Formal Opinion 512, and the SRA warning notice on the misuse of AI.

Can I just use Claude directly for legal work?

For orientation, yes: understanding an unfamiliar area, explaining a concept, tidying your own prose. Treat it as a well-read colleague with no access to the library. For anything you will rely on, cite or file, no, and the same is true of any general model. The value is not the model’s knowledge of the law; it is the model’s ability to read the law you give it, quickly and carefully, and to show you where each answer came from. That is what ModulawAI is for.

You can also connect Claude, or any assistant that speaks the Model Context Protocol, to your own ModulawAI workspace, so it can read matters and documents under your permissions. See the MCP guide and the AI legal assistant.