[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fqptr7688dvvr":3},{"_createdAt":4,"_id":5,"_updatedAt":6,"authors":7,"charactercount":8,"content":9,"featuredimage":527,"published":529,"slug":530,"subtitle":533,"tag":534,"title":535,"contentHtml":536,"featuredimageUrl":537},"2026-10-28T02:38:37Z","7ef823c1-32a4-4e5c-a0fc-7e8b1013e6cf","2026-10-28T02:58:49Z",[],8742,[10,21,38,47,63,69,77,93,108,119,142,150,166,174,182,190,198,206,214,222,229,236,243,251,259,267,275,283,292,300,308,316,324,332,340,348,356,372,380,388,396,411,419,423,431,439,446,453,460,467,475,483,491,495,503,511,519],{"_key":11,"_type":12,"children":13,"markDefs":19,"style":20},"6e8fcaf769ba","block",[14],{"_key":15,"_type":16,"marks":17,"text":18},"9b5f417835a5","span",[],"The double-entry ledger is one of the most durable data structures ever invented. It came about roughly 500 years ago in Italy, when a growing merchant class needed a better way to manage their businesses. To cut a very long story short (apologies accounting historians), it worked out pretty well: their businesses flourished and double-entry accounting became the standard for financial management and regulatory control. Today it’s everywhere. It might not be obvious, but wherever there’s financial abundance, a double-entry ledger is at work somewhere behind the scenes.",[],"normal",{"_key":22,"_type":12,"children":23,"markDefs":37,"style":20},"d5227bb4abda",[24,28,33],{"_key":25,"_type":16,"marks":26,"text":27},"bcce7ba12f9d",[],"And yet, ask any accountant how they spend their time and you’ll hear about a lot of ",{"_key":29,"_type":16,"marks":30,"text":32},"b2dd6e8661f4",[31],"em","busywork",{"_key":34,"_type":16,"marks":35,"text":36},"d124afc93475",[],": capturing data, checking data, classifying data, reconciling data, auditing data, and putting controls in place to—you guessed it—manage all that data. \"Month-end\" is a perennial season of anguish.",[],{"_key":39,"_type":12,"children":40,"markDefs":45,"style":46},"25aa1f26f13b",[41],{"_key":42,"_type":16,"marks":43,"text":44},"05068f786159",[],"What’s going on here?",[],"h2",{"_key":48,"_type":12,"children":49,"markDefs":62,"style":20},"89790111e412",[50,54,58],{"_key":51,"_type":16,"marks":52,"text":53},"61464ddb12d7",[],"The double-entry ledger is an incredible tool for financial management, but making it useful takes an awful lot of data wrangling. If you’ve ever used accounting software, you know all too well that one wrong entry knocks over all the dominoes around it. And that’s a feature of the system, not a bug: the whole point of a ledger is to make a business ",{"_key":55,"_type":16,"marks":56,"text":57},"a8c8610add69",[31],"precisely",{"_key":59,"_type":16,"marks":60,"text":61},"cecb8118c51e",[]," legible over time.",[],{"_key":64,"_type":65,"asset":66},"08a7cf842bcd","image",{"_ref":67,"_type":68},"image-509104c6f02949e7dc6c967fdbbccaafc18c6e55-1640x440-png","reference",{"_key":70,"_type":12,"children":71,"markDefs":76,"style":20},"2ac714e40ea8",[72],{"_key":73,"_type":16,"marks":74,"text":75},"29d48d070d46",[],"A business does “stuff”, that stuff shows up on a ledger, and slicing the ledger in different ways produces useful insights.",[],{"_key":78,"_type":12,"children":79,"markDefs":92,"style":20},"13ee0c1528ce",[80,84,88],{"_key":81,"_type":16,"marks":82,"text":83},"94ddb25604a7",[],"This “caveman” definition of accounting isn’t particularly profound, but it points to something important. At first glance, it might seem that most of the work happens in those big blocks; after all, they’re the ",{"_key":85,"_type":16,"marks":86,"text":87},"a385385434c6",[31],"tangible",{"_key":89,"_type":16,"marks":90,"text":91},"ed1d8f05f0b5",[]," parts of the system. But in practice, all of the busywork actually happens at the seams:",[],{"_key":94,"_type":12,"children":95,"level":105,"listItem":106,"markDefs":107,"style":20},"caf8b6f04166",[96,101],{"_key":97,"_type":16,"marks":98,"text":100},"36bb52c32432",[99],"strong","Ingestion",{"_key":102,"_type":16,"marks":103,"text":104},"2747c64f3477",[]," is the work of turning business activity into double-entry ledger data.",1,"bullet",[],{"_key":109,"_type":12,"children":110,"level":105,"listItem":106,"markDefs":118,"style":20},"9bd6b6dc0093",[111,114],{"_key":102,"_type":16,"marks":112,"text":113},[99],"Interpretation",{"_key":115,"_type":16,"marks":116,"text":117},"37a0726b9fe3",[]," is the work of turning that data into useful insights and actions.",[],{"_key":120,"_type":12,"children":121,"markDefs":141,"style":20},"b65490375415",[122,126,130,134,137],{"_key":123,"_type":16,"marks":124,"text":125},"40288234ce01",[],"Most of the time and effort goes into those hops ",{"_key":127,"_type":16,"marks":128,"text":129},"9a6f3e93a817",[31],"between",{"_key":131,"_type":16,"marks":132,"text":133},"44fd5046f66d",[]," the blocks. Ingestion is high-volume and mechanical, and it’s where errors creep in. Interpretation is lower-volume and depends on context, and it’s where those errors are invariably discovered—usually far too late. Each makes the other harder, and the ",{"_key":135,"_type":16,"marks":136,"text":32},"061f876af427",[31],{"_key":138,"_type":16,"marks":139,"text":140},"4cb4ee45801a",[]," compounds.",[],{"_key":143,"_type":12,"children":144,"markDefs":149,"style":46},"7729a6929c94",[145],{"_key":146,"_type":16,"marks":147,"text":148},"ab75360ee922",[],"AI can do most of the busywork",[],{"_key":151,"_type":12,"children":152,"markDefs":165,"style":20},"aa9bc4006ab3",[153,157,161],{"_key":154,"_type":16,"marks":155,"text":156},"5280d92329f8",[],"When you look at it this way, the opportunity for AI is clear. AI is not a replacement for the ledger, but it is a powerful tool for working ",{"_key":158,"_type":16,"marks":159,"text":160},"48bdab203904",[31],"with",{"_key":162,"_type":16,"marks":163,"text":164},"f818abe0cd45",[]," the ledger.",[],{"_key":167,"_type":12,"children":168,"markDefs":173,"style":20},"c0e84db1e5bd",[169],{"_key":170,"_type":16,"marks":171,"text":172},"651be4999868",[],"If the ledger is the detailed financial record of a business, AI can handle the information flowing onto it and coming from it at a scale that was not possible before now. That means pulling in data from far more sources than was previously practical, and interpreting it in ways that simply weren’t feasible until now.",[],{"_key":175,"_type":12,"children":176,"markDefs":181,"style":20},"8095e00894cd",[177],{"_key":178,"_type":16,"marks":179,"text":180},"5f305e4750a0",[],"And the opportunity for finance professionals is significant too. Instead of spending their time checking and rechecking data, they have a much greater opportunity to bring their judgement, insight and context to bear on a business—which has always been the most valuable use of their skills.",[],{"_key":183,"_type":12,"children":184,"markDefs":189,"style":20},"a0d150f48576",[185],{"_key":186,"_type":16,"marks":187,"text":188},"20f08da8fd35",[],"But there’s a catch.",[],{"_key":191,"_type":12,"children":192,"markDefs":197,"style":20},"34a9e6b2edde",[193],{"_key":194,"_type":16,"marks":195,"text":196},"64538a56ec34",[],"This only works if the ledger is something AI can actually work with, and most accounting tools were never designed for that.",[],{"_key":199,"_type":12,"children":200,"markDefs":205,"style":46},"c6b9a129ed0a",[201],{"_key":202,"_type":16,"marks":203,"text":204},"f5f92f4d056f",[],"AI can’t be bolted onto existing tools",[],{"_key":207,"_type":12,"children":208,"markDefs":213,"style":20},"d472fdfeef16",[209],{"_key":210,"_type":16,"marks":211,"text":212},"eafd6ad35677",[],"Until now, accounting software was built for humans: for human workflows and human limitations. Existing tools reflect that.",[],{"_key":215,"_type":12,"children":216,"level":105,"listItem":106,"markDefs":221,"style":20},"4af25fd19003",[217],{"_key":218,"_type":16,"marks":219,"text":220},"ff96a01b4ba9",[],"They use closed data models that are hard to inspect or reason about.",[],{"_key":223,"_type":12,"children":224,"level":105,"listItem":106,"markDefs":228,"style":20},"b314f6dde691",[225],{"_key":218,"_type":16,"marks":226,"text":227},[],"They’re designed around manual workflows, with lots of clicking and confirming.",[],{"_key":230,"_type":12,"children":231,"level":105,"listItem":106,"markDefs":235,"style":20},"130a9f831d79",[232],{"_key":218,"_type":16,"marks":233,"text":234},[],"They expect data to be fixed and largely unchangeable.",[],{"_key":237,"_type":12,"children":238,"level":105,"listItem":106,"markDefs":242,"style":20},"be671535c292",[239],{"_key":218,"_type":16,"marks":240,"text":241},[],"They don’t track where data came from, or keep detailed records of every change.",[],{"_key":244,"_type":12,"children":245,"markDefs":250,"style":20},"a77b46d494be",[246],{"_key":247,"_type":16,"marks":248,"text":249},"85d84fa73ff2",[],"At stub, we believe the future is one where humans and AI work together in new and more productive ways. That can’t happen by bolting AI onto existing tools, because those tools are limited in exactly the places where AI could best be applied.",[],{"_key":252,"_type":12,"children":253,"markDefs":258,"style":20},"9cf5b83f928d",[254],{"_key":255,"_type":16,"marks":256,"text":257},"af14c1207d70",[],"It’s time for new accounting tools, made for the modern era.",[],{"_key":260,"_type":12,"children":261,"markDefs":266,"style":20},"46979233eab7",[262],{"_key":263,"_type":16,"marks":264,"text":265},"106e8facbe4a",[],"That’s why we built stub.",[],{"_key":268,"_type":12,"children":269,"markDefs":274,"style":46},"a8014aba498e",[270],{"_key":271,"_type":16,"marks":272,"text":273},"35ea7a659cf0",[],"We rebuilt the ledger for the AI era",[],{"_key":276,"_type":12,"children":277,"markDefs":282,"style":20},"88c5ca18f249",[278],{"_key":279,"_type":16,"marks":280,"text":281},"fe516f9fd858",[],"At its heart, stub is an intelligent, easy-to-use general ledger. A few things set it apart.",[],{"_key":284,"_type":12,"children":285,"markDefs":290,"style":291},"af1c0c87ee9d",[286],{"_key":287,"_type":16,"marks":288,"text":289},"d7bd49fb6cf6",[],"Built automatically from your records",[],"h3",{"_key":293,"_type":12,"children":294,"markDefs":299,"style":20},"a589d8eff5ed",[295],{"_key":296,"_type":16,"marks":297,"text":298},"4a14f1b9c75d",[],"stub builds the ledger directly from your source records. Add an invoice, and the matching entry appears on the ledger. You never need to “post” anything. Day to day, you just work with the records you already know (invoices, bills, bank transactions, asset registers, payslips) and the ledger keeps itself up to date. There are built-in controls for closed periods to make sure historical data isn’t changed by accident, but for the most part, the practical realities of the business are just automatically reflected on the ledger.",[],{"_key":301,"_type":12,"children":302,"markDefs":307,"style":291},"cf381486b9dc",[303],{"_key":304,"_type":16,"marks":305,"text":306},"90d6dd6a83a5",[],"Data you can change, with a full history",[],{"_key":309,"_type":12,"children":310,"markDefs":315,"style":20},"3bb3f93c6f80",[311],{"_key":312,"_type":16,"marks":313,"text":314},"6a545d6318c0",[],"stub records every change to every piece of data: what changed, where it came from, and who (or what) made the change. You always have a complete picture of where each figure came from.",[],{"_key":317,"_type":12,"children":318,"markDefs":323,"style":20},"27a8592e2ddc",[319],{"_key":320,"_type":16,"marks":321,"text":322},"c27d1350bdb4",[],"This is what makes it safe to update data. Mistakes can be fixed at the source instead of patched over with adjusting entries, and nothing is lost along the way. You can always trace how any figure came to be what it is.",[],{"_key":325,"_type":12,"children":326,"markDefs":331,"style":291},"d51b45170bb6",[327],{"_key":328,"_type":16,"marks":329,"text":330},"a303d4a53507",[],"Context that stays with the data",[],{"_key":333,"_type":12,"children":334,"markDefs":339,"style":20},"f7b306ec6461",[335],{"_key":336,"_type":16,"marks":337,"text":338},"7d7ef598936f",[],"Traditional ledgers strip an entry down to an account, an amount and a date. Everything else, like the agreement behind the deal, the other party, or the system where the transaction actually happened, lives somewhere else, if it’s retained at all.",[],{"_key":341,"_type":12,"children":342,"markDefs":347,"style":20},"a2ddd9297246",[343],{"_key":344,"_type":16,"marks":345,"text":346},"31f3f99e776c",[],"In stub, that context stays attached. Every entry carries structured details and references alongside its accounting values, so both humans and AI can understand not just what a number is, but where it came from and why it’s there.",[],{"_key":349,"_type":12,"children":350,"markDefs":355,"style":291},"54ab07cf00f6",[351],{"_key":352,"_type":16,"marks":353,"text":354},"3c99a343290d",[],"Room for lots of data",[],{"_key":357,"_type":12,"children":358,"markDefs":371,"style":20},"8efc3f154ec7",[359,363,367],{"_key":360,"_type":16,"marks":361,"text":362},"d54c8ef0ca13",[],"People can hold about seven things in their head at once. Today’s AI models can work with up to a million tokens at a time, roughly the length of several novels. That doesn’t necessarily make them ",{"_key":364,"_type":16,"marks":365,"text":366},"851422df9b65",[31],"smarter",{"_key":368,"_type":16,"marks":369,"text":370},"c65b17b6fccf",[]," than us, but it does mean they can work through huge amounts of data in ways we’re only starting to figure out. Today, finding a needle in a barn full of haystacks is not only possible, it’s trivial with AI.",[],{"_key":373,"_type":12,"children":374,"markDefs":379,"style":20},"ba499b2de809",[375],{"_key":376,"_type":16,"marks":377,"text":378},"7e0d7c9035d8",[],"So we built stub to handle a lot of data from a lot of sources without breaking a sweat. Instead of squeezing your data down to fit your accounting system, you can simply bring it all in. The result is a much richer picture of how and where your money actually moves.",[],{"_key":381,"_type":12,"children":382,"markDefs":387,"style":20},"382e6ac5f189",[383],{"_key":384,"_type":16,"marks":385,"text":386},"1ce86cc97e43",[],"Taken together, these make for a ledger that AI can genuinely work with. But the more important question is what this means for the people responsible for the numbers.",[],{"_key":389,"_type":12,"children":390,"markDefs":395,"style":46},"c0b2a3bd3eda",[391],{"_key":392,"_type":16,"marks":393,"text":394},"6413e893a807",[],"More control, not less",[],{"_key":397,"_type":12,"children":398,"markDefs":410,"style":20},"0bff8e2bea9c",[399,403,406],{"_key":400,"_type":16,"marks":401,"text":402},"1e6b8e5b737f",[],"Working this way doesn’t mean handing control over to AI. In fact, it’s exactly the opposite: when AI does the ",{"_key":404,"_type":16,"marks":405,"text":32},"8143224bc4a2",[31],{"_key":407,"_type":16,"marks":408,"text":409},"f6baf6904213",[],", you get the time and space to apply your judgement, and you’re not limited by workflows that would typically take humans a lot of time and effort. We built stub, from the ground up, to work precisely this way.",[],{"_key":412,"_type":12,"children":413,"markDefs":418,"style":291},"d104027194d1",[414],{"_key":415,"_type":16,"marks":416,"text":417},"8cf81e09d579",[],"1. A fast, easy-to-use surface",[],{"_key":420,"_type":65,"asset":421},"99282acb380b",{"_ref":422,"_type":68},"image-9ed3057f19fd39fb29666725d3d309971b41b9e5-1640x960-png",{"_key":424,"_type":12,"children":425,"markDefs":430,"style":20},"b3900e9919b6",[426],{"_key":427,"_type":16,"marks":428,"text":429},"893611080ff1",[],"The stub app is a fast, easy-to-use finance tool for your business, where you can see what’s happening and dig into any figure without fighting the software. It gives you:",[],{"_key":432,"_type":12,"children":433,"level":105,"listItem":106,"markDefs":438,"style":20},"5edf53bc84e5",[434],{"_key":435,"_type":16,"marks":436,"text":437},"f1de14914ffd",[],"Tools organized around the reality of day-to-day operations",[],{"_key":440,"_type":12,"children":441,"level":105,"listItem":106,"markDefs":445,"style":20},"bcc0ec29e07e",[442],{"_key":435,"_type":16,"marks":443,"text":444},[],"Detailed reports that load almost instantly",[],{"_key":447,"_type":12,"children":448,"level":105,"listItem":106,"markDefs":452,"style":20},"2a9a6bbdbcad",[449],{"_key":435,"_type":16,"marks":450,"text":451},[],"Clean, organised data with custom views to suit your needs",[],{"_key":454,"_type":12,"children":455,"level":105,"listItem":106,"markDefs":459,"style":20},"c5a328e3e35c",[456],{"_key":435,"_type":16,"marks":457,"text":458},[],"Automatic data import and transaction classification built in",[],{"_key":461,"_type":12,"children":462,"level":105,"listItem":106,"markDefs":466,"style":20},"2b4b4eb1ad87",[463],{"_key":435,"_type":16,"marks":464,"text":465},[],"Simple, secure collaboration with others",[],{"_key":468,"_type":12,"children":469,"markDefs":474,"style":291},"5cc21bc60ee6",[470],{"_key":471,"_type":16,"marks":472,"text":473},"f5019944de81",[],"2. A comprehensive MCP",[],{"_key":476,"_type":12,"children":477,"markDefs":482,"style":20},"78f025bc67a4",[478],{"_key":479,"_type":16,"marks":480,"text":481},"31ecf4f41a4b",[],"The built-in MCP server lets you connect the AI tools you already use to stub. When your accounting is linked to the rest of your business stack, like this, it works much harder for you.",[],{"_key":484,"_type":12,"children":485,"markDefs":490,"style":291},"dbf92eaf8bc6",[486],{"_key":487,"_type":16,"marks":488,"text":489},"b5f542e6bec6",[],"3. An agent to do the work",[],{"_key":492,"_type":65,"asset":493},"fb144ad55b02",{"_ref":494,"_type":68},"image-b07ec5323739ca704c6560cdd93d7195bbad4b77-1640x960-png",{"_key":496,"_type":12,"children":497,"markDefs":502,"style":20},"96d8f8544c08",[498],{"_key":499,"_type":16,"marks":500,"text":501},"36bd8e614ac2",[],"The stub agent works alongside you inside stub. It answers questions about your numbers in plain language, takes on any work you hand it, remembers instructions that you gave it and flags the things that need your attention.",[],{"_key":504,"_type":12,"children":505,"markDefs":510,"style":20},"8b67d3ba4a88",[506],{"_key":507,"_type":16,"marks":508,"text":509},"d3aaeb7467f3",[],"And here’s the kicker: through all of this, a person stays accountable. The system does the legwork; but you make the calls. You end up with more control, more context and more insight into what’s really going on, and your judgement is brought to bear on your business.",[],{"_key":512,"_type":12,"children":513,"markDefs":518,"style":20},"a16d047f091b",[514],{"_key":515,"_type":16,"marks":516,"text":517},"e220363997d2",[],"We’ve seen this change the way many accountants and entrepreneurs manage their businesses and get things done. If there’s one thing that they all need it’s time saved, and already we’re saving them hours every month on admin, payments, accounts, reporting, payroll, VAT, Tax and more.",[],{"_key":520,"_type":12,"children":521,"markDefs":526,"style":20},"50409d4eae36",[522],{"_key":523,"_type":16,"marks":524,"text":525},"e22092c5ffd4",[],"The double-entry ledger has lasted five centuries because it captures something true about how businesses work. AI doesn’t change that. What it does is finally free the ledger from the busywork around it, so it can do what it was always meant to do: make a business clear enough that the people running it can make better decisions, focus on what matters, and chart a path to success.",[],{"_type":65,"asset":528},{"_ref":422,"_type":68},"2026-10-28",{"_type":531,"current":532},"slug","building-the-intelligent-general-ledger","What's the real opportunity for AI and accounting?","Accounting","Building the intelligent general ledger","\u003Cp>The double-entry ledger is one of the most durable data structures ever invented. It came about roughly 500 years ago in Italy, when a growing merchant class needed a better way to manage their businesses. To cut a very long story short (apologies accounting historians), it worked out pretty well: their businesses flourished and double-entry accounting became the standard for financial management and regulatory control. Today it’s everywhere. It might not be obvious, but wherever there’s financial abundance, a double-entry ledger is at work somewhere behind the scenes.\u003C\u002Fp>\u003Cp>And yet, ask any accountant how they spend their time and you’ll hear about a lot of \u003Cem>busywork\u003C\u002Fem>: capturing data, checking data, classifying data, reconciling data, auditing data, and putting controls in place to—you guessed it—manage all that data. "Month-end" is a perennial season of anguish.\u003C\u002Fp>\u003Ch2>What’s going on here?\u003C\u002Fh2>\u003Cp>The double-entry ledger is an incredible tool for financial management, but making it useful takes an awful lot of data wrangling. If you’ve ever used accounting software, you know all too well that one wrong entry knocks over all the dominoes around it. And that’s a feature of the system, not a bug: the whole point of a ledger is to make a business \u003Cem>precisely\u003C\u002Fem> legible over time.\u003C\u002Fp>\u003Cimg loading=\"lazy\" src=\"https:\u002F\u002Fcdn.sanity.io\u002Fimages\u002F59aqi1lf\u002Fproduction\u002F509104c6f02949e7dc6c967fdbbccaafc18c6e55-1640x440.png?w=1396&fit=crop&auto=format\" \u002F>\u003Cp>A business does “stuff”, that stuff shows up on a ledger, and slicing the ledger in different ways produces useful insights.\u003C\u002Fp>\u003Cp>This “caveman” definition of accounting isn’t particularly profound, but it points to something important. At first glance, it might seem that most of the work happens in those big blocks; after all, they’re the \u003Cem>tangible\u003C\u002Fem> parts of the system. But in practice, all of the busywork actually happens at the seams:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Ingestion\u003C\u002Fstrong> is the work of turning business activity into double-entry ledger data.\u003C\u002Fli>\u003Cli>\u003Cstrong>Interpretation\u003C\u002Fstrong> is the work of turning that data into useful insights and actions.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Most of the time and effort goes into those hops \u003Cem>between\u003C\u002Fem> the blocks. Ingestion is high-volume and mechanical, and it’s where errors creep in. Interpretation is lower-volume and depends on context, and it’s where those errors are invariably discovered—usually far too late. Each makes the other harder, and the \u003Cem>busywork\u003C\u002Fem> compounds.\u003C\u002Fp>\u003Ch2>AI can do most of the busywork\u003C\u002Fh2>\u003Cp>When you look at it this way, the opportunity for AI is clear. AI is not a replacement for the ledger, but it is a powerful tool for working \u003Cem>with\u003C\u002Fem> the ledger.\u003C\u002Fp>\u003Cp>If the ledger is the detailed financial record of a business, AI can handle the information flowing onto it and coming from it at a scale that was not possible before now. That means pulling in data from far more sources than was previously practical, and interpreting it in ways that simply weren’t feasible until now.\u003C\u002Fp>\u003Cp>And the opportunity for finance professionals is significant too. Instead of spending their time checking and rechecking data, they have a much greater opportunity to bring their judgement, insight and context to bear on a business—which has always been the most valuable use of their skills.\u003C\u002Fp>\u003Cp>But there’s a catch.\u003C\u002Fp>\u003Cp>This only works if the ledger is something AI can actually work with, and most accounting tools were never designed for that.\u003C\u002Fp>\u003Ch2>AI can’t be bolted onto existing tools\u003C\u002Fh2>\u003Cp>Until now, accounting software was built for humans: for human workflows and human limitations. Existing tools reflect that.\u003C\u002Fp>\u003Cul>\u003Cli>They use closed data models that are hard to inspect or reason about.\u003C\u002Fli>\u003Cli>They’re designed around manual workflows, with lots of clicking and confirming.\u003C\u002Fli>\u003Cli>They expect data to be fixed and largely unchangeable.\u003C\u002Fli>\u003Cli>They don’t track where data came from, or keep detailed records of every change.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>At stub, we believe the future is one where humans and AI work together in new and more productive ways. That can’t happen by bolting AI onto existing tools, because those tools are limited in exactly the places where AI could best be applied.\u003C\u002Fp>\u003Cp>It’s time for new accounting tools, made for the modern era.\u003C\u002Fp>\u003Cp>That’s why we built stub.\u003C\u002Fp>\u003Ch2>We rebuilt the ledger for the AI era\u003C\u002Fh2>\u003Cp>At its heart, stub is an intelligent, easy-to-use general ledger. A few things set it apart.\u003C\u002Fp>\u003Ch3>Built automatically from your records\u003C\u002Fh3>\u003Cp>stub builds the ledger directly from your source records. Add an invoice, and the matching entry appears on the ledger. You never need to “post” anything. Day to day, you just work with the records you already know (invoices, bills, bank transactions, asset registers, payslips) and the ledger keeps itself up to date. There are built-in controls for closed periods to make sure historical data isn’t changed by accident, but for the most part, the practical realities of the business are just automatically reflected on the ledger.\u003C\u002Fp>\u003Ch3>Data you can change, with a full history\u003C\u002Fh3>\u003Cp>stub records every change to every piece of data: what changed, where it came from, and who (or what) made the change. You always have a complete picture of where each figure came from.\u003C\u002Fp>\u003Cp>This is what makes it safe to update data. Mistakes can be fixed at the source instead of patched over with adjusting entries, and nothing is lost along the way. You can always trace how any figure came to be what it is.\u003C\u002Fp>\u003Ch3>Context that stays with the data\u003C\u002Fh3>\u003Cp>Traditional ledgers strip an entry down to an account, an amount and a date. Everything else, like the agreement behind the deal, the other party, or the system where the transaction actually happened, lives somewhere else, if it’s retained at all.\u003C\u002Fp>\u003Cp>In stub, that context stays attached. Every entry carries structured details and references alongside its accounting values, so both humans and AI can understand not just what a number is, but where it came from and why it’s there.\u003C\u002Fp>\u003Ch3>Room for lots of data\u003C\u002Fh3>\u003Cp>People can hold about seven things in their head at once. Today’s AI models can work with up to a million tokens at a time, roughly the length of several novels. That doesn’t necessarily make them \u003Cem>smarter\u003C\u002Fem> than us, but it does mean they can work through huge amounts of data in ways we’re only starting to figure out. Today, finding a needle in a barn full of haystacks is not only possible, it’s trivial with AI.\u003C\u002Fp>\u003Cp>So we built stub to handle a lot of data from a lot of sources without breaking a sweat. Instead of squeezing your data down to fit your accounting system, you can simply bring it all in. The result is a much richer picture of how and where your money actually moves.\u003C\u002Fp>\u003Cp>Taken together, these make for a ledger that AI can genuinely work with. But the more important question is what this means for the people responsible for the numbers.\u003C\u002Fp>\u003Ch2>More control, not less\u003C\u002Fh2>\u003Cp>Working this way doesn’t mean handing control over to AI. In fact, it’s exactly the opposite: when AI does the \u003Cem>busywork\u003C\u002Fem>, you get the time and space to apply your judgement, and you’re not limited by workflows that would typically take humans a lot of time and effort. We built stub, from the ground up, to work precisely this way.\u003C\u002Fp>\u003Ch3>1. A fast, easy-to-use surface\u003C\u002Fh3>\u003Cimg loading=\"lazy\" src=\"https:\u002F\u002Fcdn.sanity.io\u002Fimages\u002F59aqi1lf\u002Fproduction\u002F9ed3057f19fd39fb29666725d3d309971b41b9e5-1640x960.png?w=1396&fit=crop&auto=format\" \u002F>\u003Cp>The stub app is a fast, easy-to-use finance tool for your business, where you can see what’s happening and dig into any figure without fighting the software. It gives you:\u003C\u002Fp>\u003Cul>\u003Cli>Tools organized around the reality of day-to-day operations\u003C\u002Fli>\u003Cli>Detailed reports that load almost instantly\u003C\u002Fli>\u003Cli>Clean, organised data with custom views to suit your needs\u003C\u002Fli>\u003Cli>Automatic data import and transaction classification built in\u003C\u002Fli>\u003Cli>Simple, secure collaboration with others\u003C\u002Fli>\u003C\u002Ful>\u003Ch3>2. A comprehensive MCP\u003C\u002Fh3>\u003Cp>The built-in MCP server lets you connect the AI tools you already use to stub. When your accounting is linked to the rest of your business stack, like this, it works much harder for you.\u003C\u002Fp>\u003Ch3>3. An agent to do the work\u003C\u002Fh3>\u003Cimg loading=\"lazy\" src=\"https:\u002F\u002Fcdn.sanity.io\u002Fimages\u002F59aqi1lf\u002Fproduction\u002Fb07ec5323739ca704c6560cdd93d7195bbad4b77-1640x960.png?w=1396&fit=crop&auto=format\" \u002F>\u003Cp>The stub agent works alongside you inside stub. It answers questions about your numbers in plain language, takes on any work you hand it, remembers instructions that you gave it and flags the things that need your attention.\u003C\u002Fp>\u003Cp>And here’s the kicker: through all of this, a person stays accountable. The system does the legwork; but you make the calls. You end up with more control, more context and more insight into what’s really going on, and your judgement is brought to bear on your business.\u003C\u002Fp>\u003Cp>We’ve seen this change the way many accountants and entrepreneurs manage their businesses and get things done. If there’s one thing that they all need it’s time saved, and already we’re saving them hours every month on admin, payments, accounts, reporting, payroll, VAT, Tax and more.\u003C\u002Fp>\u003Cp>The double-entry ledger has lasted five centuries because it captures something true about how businesses work. AI doesn’t change that. What it does is finally free the ledger from the busywork around it, so it can do what it was always meant to do: make a business clear enough that the people running it can make better decisions, focus on what matters, and chart a path to success.\u003C\u002Fp>","https:\u002F\u002Fcdn.sanity.io\u002Fimages\u002F59aqi1lf\u002Fproduction\u002F9ed3057f19fd39fb29666725d3d309971b41b9e5-1640x960.png?w=1280"]