HundredOneRISK
Sign inJoin

5 apps free for 3 days, no card

AI for the work that has to be right. Every answer shows its evidence.

Policy questions answered with the clause quoted, obligations mapped to controls, regulatory change tracked with dates, third parties reviewed, evidence pulled together before the auditor asks. Pick the job, make a few choices, and it is reading your material in minutes.

  • Every answer quotes its source
  • Every change is a kept version
  • Your evidence stays in your workspace
  • On-premise available
  • No coding
Join the waitlistSee how it works

Nothing here is a certification and no output is a sign-off. It does the reading and the drafting; a person still attests.

Click. Click. Done. And you can show where every line came from.

All the clever AI stuff happens behind the scenes. Nothing to configure, nothing to install, and no prompt to get right — you answer questions about the work, in the words your own framework already uses.

  1. Step 1

    Start from the obligation

    Answer questions from the policy set, map requirements onto controls, track what a regulator changed, review a supplier's pack, pull the evidence for a finding. You pick the job; there is no blank page and no framework to import first.

  2. Step 2

    Make a few choices

    How formally it writes, how far it may go, what it is allowed to read, whether it may search the public web. Every step is a choice from a short list. Upload the policies, the standards and the evidence once and it works from those.

  3. Step 3

    Use it, and keep the trail

    It opens ready to work. Every change to it is a new version with the person and the time attached, and any version can be restored — so the app that produced last quarter's answer is still there when somebody asks how it was produced.

The roles the function never has enough of

Each of these is one app, built from the same recipes everyone else uses and named for the job it does in a second-line or audit function. A couple of minutes each, and nothing to write.

  • Policy Analyst

    Hand it the policy set, the standards and the procedures. Ask what is required and get the answer with the clause quoted, so the person who asked can check it against the document rather than trusting a paraphrase.

    Policy questionsStandard interpretationGap questions

  • Regulatory Researcher

    Look a requirement up properly and get a briefing with its sources attached: what was found, where each part came from, and what the material does not settle. Every page listed with its publisher and its date.

    New regimesGuidanceJurisdiction checks

  • Compliance Summarizer

    A 200-page consultation, a standard, a supervisory letter — down to the parts that create an obligation, with every date, threshold and defined term exactly as written.

    ConsultationsStandardsSupervisory letters

  • Control Mapper

    Give it a set of requirements and your control library, and get back the mapping: which control claims to cover which obligation, where nothing does, and where the same control is carrying more than it should.

    Obligation mappingFramework crosswalksCoverage gaps

  • Regulatory Change Analyst

    Follows a regulator, a regime or a topic and reports what actually changed — separating the events from the commentary, naming who published what, and dating every page it read.

    Horizon scanningRegulator watchMonthly change note

  • Evidence Reviewer

    Point it at the evidence pack and ask whether what is in there supports the control as described — and get back the passages that do, the ones that do not, and the questions still open.

    Control testingSampling packsFindings

  • Third-Party Risk Assistant

    Hand it the suppliers on your register and get back what a review starts from: what each company does, what changed there lately, what is publicly known — organizations only, never an individual's personal details.

    Vendor due diligenceRegister refreshConcentration

  • Audit Assistant

    Say what has to be answered and get the draft: the response to a finding, the management action, the paper for the committee — in the house format, leaving a marked blank rather than asserting a control that was not evidenced.

    Findings responsesCommittee papersAttestations

  • Incident Analyst

    Paste the messages, the tickets and the logs and get the incident back as one page: the timeline, who did what and when, what was affected, and which of it is still assumption rather than record.

    TimelinesPost-incident reviewsNotifiable events

  • Risk Assessor, and anything else the function does

    Describe the job in your own words and build the app for it. A risk assessment against your own methodology, a DPIA first draft, a questionnaire scored against your criteria, every issue from a quarter grouped into themes — if you can say what goes in and what should come out, it can be built without a template.

    Risk assessmentsFirst-draft DPIAsIssue themes

One app can put the question to another

Tick which of your apps a new one may consult. It works out for itself, question by question, whether any of them would help — you never write a rule about when, and every answer names which app was asked and what it was asked.

  • Every answer carries its evidence

    Answers from your documents quote the passage they came from, and apps that search the public web list every page with its publisher and its date. A reviewer checks a citation instead of taking a claim on trust.

  • Every change is a version

    Each edit to an app is a new version with the person and the time attached, and any version can be restored. When somebody asks how an answer was produced last quarter, the app that produced it still exists.

  • Separation you can point at

    Consulting an app is not merging it. Each keeps its own documents, anyone you share the caller with reaches only the apps they could already open themselves, and each answer names what was consulted — so a second-hand conclusion is labeled as one.

Honest pricing options that fit all budgets and needs

Pick the plan you’d want and we’ll come to you first when it opens.

  • Free trial

    Free

    See whether this actually works for your problem.

    • Build 5 apps of your own
    • Our sample apps, free and unlimited
    • 3 days to make up your mind
    • Every app type, nothing held back
    • No credit card
    Join the waitlist, Free trial

  • Individual

    Economical

    $29/month

    One person, with their own apps in daily use.

    • One account, just for you
    • Build up to 40 apps — dozens of jobs, each done properly
    • Around 200 runs a month, across all of them
    • Every app type, at full quality
    • 5 GB of your documents
    • Version history, restore, and email support
    Join the waitlist, Individual

  • Teams

    Premium

    $99/month

    Up to five people, sharing everything they build — or one person who wants the lot.

    • Up to 5 people, sharing one pool of work
    • Build up to 200 apps — hundreds of jobs covered
    • Around 1,000 runs a month — 5× Individual
    • Everything in Individual, for everyone
    • 25 GB of your documents
    • Connect your own storage and model keys
    Join the waitlist, Teams

  • Corporate

    Let’s talk

    Your documents can't leave your network, and you're buying for a group.

    • Your own storage and your own model endpoint
    • Usage on your provider account, not ours
    • Volume pricing across your teams
    • Single invoice, one agreement
    • A named contact, not a queue
    Contact us, Corporate

App and run numbers are estimates. Short questions go further, long documents go less far. Change plan or cancel at any time. Your data and the results you produce are yours to keep.

Three jobs, start to finish

The same work, done by one app, two and three. Whichever one you open is the only one you deal with; anything it consults is named alongside the answer.

  • One app

    Most of it looks like this

    First example. A one hundred and eighty page regulatory consultation arrives with a committee paper due on Thursday. Compliance Summarizer, built from the Summarizer recipe, works alone — the job is reading and cutting — and hands back two pages covering what creates an obligation, every threshold and date exactly as written, and what is only proposed.

    One app, nothing wired to anything, two minutes to build. You still read the consultation — you read it knowing which twenty pages matter.

  • Two apps

    When the answer has to come from the policy

    Second example. The business asks whether a supplier can be onboarded without a full review below a fifty thousand pound threshold. Audit Assistant, built from the Writer recipe, consults Policy Analyst, built from the Chat with Documents recipe, on what the third-party policy requires below that threshold. It then drafts a reply quoting the clause and the threshold, and naming what the policy does not cover.

    The drafting app never had to hold the policy set. The policy app never had to learn the house format. The answer arrives with the clause attached, which is the only form in which it is worth anything.

  • Three apps

    And when the regulator has moved

    Third example. New reporting requirements come into force in six months and the board wants the impact. Audit Assistant, built from the Writer recipe, consults three apps in turn: Regulatory Change Analyst, on what changed, who published it and when it bites; Policy Analyst, on what current policies already require; and Control Mapper, on which controls claim to cover it and where there is nothing. It then drafts an impact paper naming the source of each line, including the gaps, for a person to own. Each app was called directly, and none of them called any further.

    Three sources, one paper, and every line names where it came from — including the gaps, which are the part a committee actually needs.

Questions people ask

Does using this make us compliant, or certified?
No. HundredOne is software you build tools with. It holds no compliance certifications and claims none, and nothing an app produces is an attestation, an opinion or a sign-off — a person still owns each of those. What it does is the reading, the mapping and the drafting that sits underneath them, with the source of every line visible.
Can an answer be audited afterwards?
That is what the product is built to make possible. Answers from your documents quote the passage they came from; apps that search the public web list every page with its publisher and its date; and every answer names which other apps were consulted and the exact question put to each. Every change to an app is a new version with the person and the time attached, and any version can be restored, so the app that produced an answer last quarter is still there.
Where does our evidence go?
It stays in your workspace, visible to you and to anyone you deliberately share an app with. Answering a question means sending the relevant passages to the AI model provider that answers it; nothing is sold or shared with anyone else. If your material cannot leave your own network at all, that is what the Corporate plan is for — an on-premise or private-cloud deployment using your own storage and your own model endpoint, so nothing crosses your boundary.
Will it hallucinate a requirement that does not exist?
It answers from the material it was given and quotes the passage, which is what makes an unsupported claim visible rather than plausible. Where it searches the public web it names and dates every page, and where sources disagree it says so. It is built for a reader who checks — and in this function, that reader is the point.
Can first line, second line and audit use the same apps?
Yes, and separation still holds. Teams is $99 a month for up to 5 people sharing everything they build; anyone you share an app with reaches only the apps they could already open themselves, so a policy app can be shared widely while an evidence pack stays where it is. Individual is $29 a month, and larger functions are priced individually. Before any of it there is a free trial: 5 apps for 3 days, no credit card.
What files can it read?
PDFs, Word documents, Excel spreadsheets, PowerPoint decks, and plain text, Markdown, CSV and JSON. Files are read in your browser, so the original never leaves your computer — only the text your app searches is stored. A scanned PDF with no selectable text is refused at upload, with that reason given, rather than accepted and quietly stored as empty.
Which AI model is behind it, and can we choose?
We choose the model for each app and change it as better ones appear, which is part of what you are paying for. On the Corporate plan you can bring your own model keys and endpoints and run everything against your existing provider agreement instead — which is usually the answer where a model provider has to be one your organization has already assessed.
Can we run it inside our own network?
Yes, on the Corporate plan. On-premise and private-cloud deployments are set up with you rather than self-served, using your own storage and your own model endpoint. Send us your security questionnaire, your data-residency requirement or your procurement checklist and you will get a direct answer on each line, in writing, from someone who knows the system — including the lines where the answer is no.

The reading is done, and you can show where every line came from.

Build 5 apps, use them for 3 days, no credit card required.If they aren’t useful, you’ve lost a few minutes.If they are, the next consultation is read and the mapping is drafted before the meeting.

Join the waitlistTalk to us about your function