The three classic strategy canvases for Ariski, generated from the CEMI business profile rather than drawn by hand — so they cannot quietly disagree with the revenue page, the business plan or the round. A block with nothing true in it says «not yet defined» instead of being filled to look complete.
Osterwalder’s nine blocks. Customer segments, the value proposition, revenue streams and key resources are read from the business profile; channels, relationships, activities, partners and cost structure from its canvas facet.
Key partners
CEMI.ai, as publisher and as the source of the shared platform. It is the only relationship the model currently depends on.
Infrastructure and model suppliers — Firebase for hosting, authentication and data, and a commercial model provider for the assistant. Both are substitutable and neither is exclusive.
No carrier, brokerage, professional association, reseller, publisher or channel partner is signed. None is named here because none exists — a partnership list is exactly the thing a kit at this stage is tempted to invent.
Key activities
Writing and keeping current the editorial catalogue as regulation and tooling move. This is the activity the model lives or dies on: a stale catalogue is worth nothing, which is why a quarter of the round is content.
Verifying every jurisdiction rule against the statute or the regulator's own text before its page publishes, and labelling the rest «in verification» rather than guessing.
Designing the published Learning Program formats into deliverable units — instructional design, assessment and per-sector case material — and then delivering them.
Operating the platform: the assistant and its guardrails, accounts and usage tiers, comment moderation, full-text search and the third locale.
Commercial development into carriers, brokerages, professional associations and supervisors — the buyers the platform already reaches as readers. Planned; the round funds the first dedicated effort.
Key resources
The curator — one person's insurance-and-AI judgement, and the key-person risk.
CEMI.ai's shared machinery — the Personai engine, the design and content system, the registry.
The published catalogue as a compounding indexed asset, verified jurisdictions included.
The registered reader base and its usage data — under NDA, never published.
The editorial catalogue and the interactive tools, in English and Spanish, plus the Ariski brand and logo set.
Insurance-domain framing applied to AI practice.
Value propositions
An independent guide to AI in insurance — written for working professionals rather than for a vendor's sales funnel. Ariski explains what AI can and cannot do in underwriting, claims, pricing, compliance and service, gives practitioners tools they can use the same day, and turns that competence into structured learning programs for the organizations that employ them. Because Ariski sells no AI system, its judgement is not for sale either, and that independence is the offer.
Customer relationships
Self-serve for every reader, deliberately: nothing is gated, nothing is metered except the assistant, and no email is required to read anything.
Assisted self-serve through Ask Ariski, with published usage caps rather than a paywall — five messages a day anonymously, twenty-five with a free account and all four voices.
A moderated community: comments and likes on content, reviewed before they appear.
A dedicated, designed-per-organization relationship for a programme or consulting engagement — the relationship the business intends to be paid for. Planned; none is running.
Channels
ariski.com, free and ungated. Organic search and direct reading are the whole top of the funnel today.
The car-claims series — three reader tracks plus a page per jurisdiction, every rule cited. The widest way in, and in production now.
Ask Ariski — the four-voice assistant, where an anonymous reader becomes a registered one.
Registered accounts and the newsletter — the only owned, addressable channel.
Direct request through the contact page's three enquiry types. There is no sales team; the round funds the first.
Professional associations as a route to their members — planned. None signed, none named.
Customer segments
Consumers (mass market)
Insurers
Insurance brokers and agents
Fintechs, insurtechs and regtechs
Banks and financial institutions
Large corporations
SMEs
Freelancers and independent professionals
Government and public sector
Cost structure
People: the curator's time today, and the programme-delivery and commercial capacity the round funds — together the largest cost by far.
Content: keeping the catalogue and the jurisdiction dataset current. The largest recurring non-people cost, and the one that cannot be deferred without the asset decaying.
Infrastructure: static hosting and Firebase — small and close to fixed, which is the point of the architecture.
Model inference for Ask Ariski — the one cost that scales with readership, held down today by the published usage caps rather than by luck.
Cost of delivering an engagement — unknown, because none has been delivered. The business plan makes measuring it a first-year milestone.
Legal, entity and administration, including the investor reporting the ticket rights commit to.
Revenue streams
Training and events
Consulting
Sale of products
Free — no direct income
The same canvas as a print SVG — generated from the same data, one page of the kit PDF.
Lean Canvas
Maurya’s nine blocks. The unfair advantage is stated and then qualified in its own text: it is a real structural asymmetry, not a moat.
Problem
In the site's own framing: most resources available to insurance professionals fall into two camps — uncritical enthusiasm from vendors trying to sell products, or reflexive fear from commentators who do not understand the technology. Neither serves the industry. Meanwhile AI is already touching underwriting, claims, pricing and service, and the professionals accountable for those decisions have nowhere neutral to learn what it does, where it fails, and what their regulator will ask.
Solution
Ariski platform
The Learning Program
AI strategy consulting
AI products for the insurance industry
Key metrics
Registered accounts, and the share of readers who create one — the only conversion the platform controls end to end today.
Assisted sessions with Ask Ariski, and how many reach a question the catalogue does not answer. That second number is the content backlog, stated as demand rather than as opinion.
Jurisdictions published as verified against the primary source, versus those still «in verification».
Qualified enquiries from the contact page, split by its three enquiry types — the earliest signal of which stream is real.
First paid engagements signed, and whether any renews. This is the metric that settles the plan; nothing above it does.
Cost of delivery per engagement, once there is one to measure.
Unique value proposition
An independent guide to AI in insurance — written for working professionals rather than for a vendor's sales funnel. Ariski explains what AI can and cannot do in underwriting, claims, pricing, compliance and service, gives practitioners tools they can use the same day, and turns that competence into structured learning programs for the organizations that employ them. Because Ariski sells no AI system, its judgement is not for sale either, and that independence is the offer.
Unfair advantage
Structural independence: Ariski sells no AI system, so it can publish what does not work, and a carrier cannot buy that judgement from a vendor whose revenue depends on the answer. Underneath it sits a cost base no standalone competitor has — CEMI.ai's persona engine, design system and registry — so this depth of sector platform is affordable here and expensive anywhere else. Stated honestly: this is a real structural asymmetry, not a moat. A well-funded entrant could reproduce the position; what it cannot cheaply reproduce is a catalogue and a verified jurisdiction dataset already published and indexed.
Channels
ariski.com, free and ungated. Organic search and direct reading are the whole top of the funnel today.
The car-claims series — three reader tracks plus a page per jurisdiction, every rule cited. The widest way in, and in production now.
Ask Ariski — the four-voice assistant, where an anonymous reader becomes a registered one.
Registered accounts and the newsletter — the only owned, addressable channel.
Direct request through the contact page's three enquiry types. There is no sales team; the round funds the first.
Professional associations as a route to their members — planned. None signed, none named.
Customer segments
Consumers (mass market)
Insurers
Insurance brokers and agents
Fintechs, insurtechs and regtechs
Banks and financial institutions
Large corporations
SMEs
Freelancers and independent professionals
Government and public sector
Cost structure
People: the curator's time today, and the programme-delivery and commercial capacity the round funds — together the largest cost by far.
Content: keeping the catalogue and the jurisdiction dataset current. The largest recurring non-people cost, and the one that cannot be deferred without the asset decaying.
Infrastructure: static hosting and Firebase — small and close to fixed, which is the point of the architecture.
Model inference for Ask Ariski — the one cost that scales with readership, held down today by the published usage caps rather than by luck.
Cost of delivering an engagement — unknown, because none has been delivered. The business plan makes measuring it a first-year milestone.
Legal, entity and administration, including the investor reporting the ticket rights commit to.
Revenue streams
Training and events
Consulting
Sale of products
Free — no direct income
The same canvas as a print SVG — generated from the same data, one page of the kit PDF.
Value Proposition Canvases
One canvas per customer segment the profile lists — nine in all, since insurance brokers and agents were added as a segment of their own on 2026-09-13. Each is written from copy the site actually publishes for that audience, and the organizational segments are written from the organization’s perspective: its own risks and its own policies rather than an individual’s. Where a segment has no copy of its own on the site — insurtechs and regtechs, banks — the canvas says so in its own words instead of inventing an audience, and any area with nothing true in it reads «not yet defined».
Consumers (mass market)
Value map
Products & services
Ariski platform
Pain relievers
One page per jurisdiction: fault system, shared-fault rule, the deadline to claim and to sue, insurer response times, minimum liability coverage and the regulator to complain to — every rule cited to its instrument, and anything unverified published as «in verification» rather than guessed.
Free tools that need no sign-up and run in the browser: a deadline calculator, a claim-readiness checklist, a total-loss estimator, a do-I-need-a-lawyer tool and a 3D crash reconstruction that produces the diagram, numbers and narrative a claim needs.
Guides written on the driver's own track — «you had an accident, here is what to do, in order» — rather than on the insurer's.
A standing notice on every guide that it is not legal advice, creates no lawyer–client relationship and states no insurer's or regulator's position — so the reader knows exactly what they are holding.
Gain creators
A «how AI changes this» section on the guides, so the reader learns where a model is in their own claim and what to ask when it is.
Three reader tracks on the same topic — drivers, companies and fleets, claims professionals — so a reader can see how the other side reads it.
Published in English and Spanish equally, with French next, so a claimant is not disadvantaged by the language of the statute.
Customer profile
Customer jobs
Get a car insurance claim filed, handled and paid — usually for the first time, and usually in a bad week.
Find out the deadline that applies where the accident happened, before it passes.
Decide whether the offer on the table is reasonable, and whether the vehicle is really a total loss.
Decide whether this needs a lawyer.
Understand what happened when a system, rather than a person, decided their claim.
Pains
Most explanations of the process are published by the party that decides whether to pay.
The deadlines that matter sit in statutes of a jurisdiction the claimant did not choose.
Valuations and fault splits arrive as conclusions, with no working shown.
Increasingly a model made the call, and nobody will explain it.
Legal information costs money the claimant is short of.
Gains
Knowing the order of the steps, so nothing is done in the wrong week.
Deadlines as actual dates for their own jurisdiction, each traced to the rule it comes from.
An independent read on the offer, from someone with nothing to gain from the answer.
Knowing when AI is in the decision, and what recourse exists when it is.
Insurers
Value map
Products & services
Ariski platform
The Learning Program
AI strategy consulting
AI products for the insurance industry
Pain relievers
Ariski sells no AI system, so it can and does publish what does not work. That is the whole offer, and a vendor structurally cannot make it.
Content organized by the work insurance people actually do — underwriting, claims, pricing, compliance, service — rather than by AI technique.
A jurisdiction guide and a car-claims rule set cited to the statute or the regulator's own text, with anything unverified labelled as such.
The readiness assessment, the adoption roadmap and the ethics simulator make the internal competence gap visible before an incident does.
The Learning Program is designed per organization rather than sold off a shelf — which is also why no price is published.
Gain creators
Eighteen published formats across five categories and three learning modes, so the same competence can reach a board, a claims floor and an actuarial team in the shape each needs.
Case material drawn from the sectors the catalogue already covers, so the examples are recognisable rather than generic.
The free platform lets anyone in the organization read the same material as the people who bought the programme — no seat licence, no gate.
Consulting available on the same relationship when a question turns out to need a decision rather than a course.
Customer profile
Customer jobs
Get AI into underwriting, claims, pricing and service without breaking a process the business already depends on.
Answer the supervisor: show how an automated decision was reached, on what data, with what recourse for the insured.
Make a build-or-buy call on vendor claims that arrive faster than anyone can evaluate them.
Bring a whole team — underwriters, adjusters, actuaries, compliance, executives — to a shared level of competence and a shared vocabulary.
Pains
Most available material is sales material, written by people paid for a yes.
Generic AI training does not map to insurance roles, and regulation and tooling move faster than internal training can be rewritten — differently in each jurisdiction.
A wrong model decision is a conduct problem, not only a technical one.
Competence is uneven across the organization, and invisible until something goes wrong.
Gains
A team that can take apart a vendor claim and say no with a reason.
A defensible answer for the supervisor, grounded in the rule as written.
Training that maps to the actual roles and sectors.
Adoption driven by judgement, not enthusiasm or fear.
Insurance brokers and agents
Written from the site's own published pages for intermediaries: the pro track hub («for adjusters, brokers and lawyers»), the about page's own list of who Ariski exists for, the adoption roadmap builder's «Broker / Agent» role and «Independent Agent» organization size, the frontline track of the published curriculum, the agency prompts in the prompt library, the agency-management entries in the tool directory, the published FAQ on telling clients about AI, and the per-jurisdiction rule pages. The honest gap: nothing on the site is sold to a broker or an agent as a product. The Learning Program reaches them through its frontline track and consulting through an enquiry type, both request-based with no price published, and there is no agency product, no placement or commission tooling and no intermediary-specific licence. So this canvas describes what an intermediary can read, use and be trained on, and says plainly that it does not describe a shelf they can buy from.
Value map
Products & services
Ariski platform
The Learning Program
AI strategy consulting
Pain relievers
Free and open, with no account required to read and no seat count to buy — which is what a one- or two-person agency can actually adopt.
A professional track on the car-claims series written, in the site's own words, for adjusters, brokers and lawyers — photo-based estimation, straight-through processing, fraud scoring and what regulators expect.
One rule page per jurisdiction — fault system, shared-fault rule, the deadline to claim and to sue, insurer response times, minimum liability cover and the regulator to complain to — each rule cited to its instrument, and anything unverified published as «in verification» rather than guessed.
A published answer to «should I tell clients I'm using AI?», and a standing «what not to do» covering client data and decisions nobody can explain.
A tool directory that gives agency-management and AI products their weaknesses as well as their strengths, because Ariski sells none of them.
Gain creators
The adoption roadmap builder takes «Broker / Agent» as a role and «Independent Agent» as an organization size, so its output is sized to an agency rather than to a carrier.
A prompt library with agency work already in it — coverage recommendation letters, policy change confirmations, client onboarding, win-back outreach, the impact of a new regulation on agents and brokers — and a builder for whatever is not there.
The programme's frontline track names brokers among its audience, in short formats an agency can release someone for.
Ask Ariski answers a specific question at the moment it comes up, which is how an agency actually learns.
Published in English and Spanish equally, with French next, so an agency is not disadvantaged by the language of the material.
Customer profile
Customer jobs
Place cover and keep a book renewing with the staff already on the payroll — one or two people, in most agencies.
Explain to a client what the insurer's automated path just did to their claim, their repair estimate or their premium.
Decide what to tell clients about AI in the process, and when — the question the site answers as a published FAQ.
Choose among the agency-management and AI tools sold to intermediaries, with no technology function to evaluate them.
Meet conduct, disclosure and data expectations without a compliance department to interpret them.
Pains
AI-in-insurance material is written by the system's sellers, or for carriers — not for the intermediary.
The automation sits inside the insurer, invisible from outside — yet the client calls the broker.
Tools marketed to agencies are scoped and priced for larger operations.
Deadlines and fault rules differ by jurisdiction, so the advice rests on memory.
Gains
Something usable in a client conversation the same day, at no cost.
A plain-language account of what the carrier automates.
A defensible position on what to tell a client about AI, and why.
Deadlines and fault rules as written, citeable in front of a client.
Fintechs, insurtechs and regtechs
Rewritten 2026-09-13 on Carlos's framing, from the organization's perspective: an insurtech or regtech is both a vendor into insurance and an insured organization of its own, and the jobs, pains and gains now read from both. The canvas is no longer «thin by design» — the material it draws on is the tool directory, the jurisdiction guide, the business track of the car-claims series, the readiness assessment and the ethics simulator, all published. The remaining gap is unchanged and recorded honestly: no page on the site addresses insurtechs or regtechs as an audience by name, so nothing here is written from copy aimed at them. If the segment matters commercially it needs its own copy on the site.
Value map
Products & services
Ariski platform
AI products for the insurance industry
AI strategy consulting
Pain relievers
Content organized by insurance work rather than by AI technique, and a tool directory that describes the category honestly — weaknesses alongside strengths, because Ariski sells none of the products in it.
A jurisdiction guide and per-jurisdiction rule pages cited to the statute or the regulator's own text, with anything unverified labelled as such.
The business track of the car-claims series treats an employer's own motor exposure as its own subject — which is what this company is when it is the insured rather than the vendor.
Gain creators
The Learning Program's formats can bring a product team and a commercial team to the same vocabulary as the carriers they sell to.
The ethics simulator and the readiness assessment put the governance questions a carrier buyer will ask in front of the team before the buyer does.
Customer profile
Customer jobs
Understand the insurance domain the company's product is sold into — the roles, the constraints and the regulator behind them.
Build the company's own insurance literacy, which is usually thinner than its technical literacy.
Anticipate what the carrier buyer's own governance will ask of the product, because that buyer answers to a supervisor.
Insure the company itself — staff who drive for work, vehicles on the books — and run those claims with no risk function to own them.
Pains
A platform publishing what AI does not do is useful and uncomfortable at once — the buyer reads it too.
Generic AI material misses insurance roles; insurance material is written for carriers.
Regulatory expectations differ by jurisdiction and move faster than a product roadmap.
The company's own motor exposure sits between personal and commercial cover.
Gains
Domain grounding the engineering and product teams do not otherwise have.
A reading of what a carrier's compliance function will ask, available before the sales call rather than during it.
The company's own claims handled as a process rather than as an emergency.
Banks and financial institutions
Rewritten 2026-09-13 on Carlos's framing, from the organization's perspective: a bank appears here both as a distributor of insurance under its own name and as an insured organization with its own vehicles, staff and claims. The canvas is no longer «thin by design» — it draws on the jurisdiction guide, the professional and business tracks, the readiness assessment and the ethics simulator. The gap is unchanged and stated: no page on the site addresses a bank as an audience by name, and bancassurance is what makes the segment plausible rather than anything the site says about it. If it matters commercially it needs its own copy on the site.
Value map
Products & services
Ariski platform
AI strategy consulting
Pain relievers
The jurisdiction guide and the compliance material are written to the insurance instrument, not to a general financial-services summary of it.
The professional track is written for the people who meet the automation — adjusters, brokers and lawyers — which is the function a bancassurance desk performs.
The business track treats an organization's own motor exposure as its own subject, with a claim-readiness checklist and a deadline calculator to run it on.
Ariski sells no AI system, so the reading it offers is not a vendor's.
Gain creators
The readiness assessment and the ethics simulator surface the governance gap in the institution's own terms before an incident does.
Consulting is scoped per engagement, so the institution can buy a reading of its own insurance book rather than a course — which is the stream the revenue page names for this segment.
Customer profile
Customer jobs
Apply to an insurance book the same AI governance the institution already runs over credit decisions.
Answer a supervisor about an automated decision taken on an insurance product distributed under the institution's own name.
Bring the bancassurance desk to a shared vocabulary with the carrier partner whose automation decides its customers' claims.
Insure the institution itself — its vehicles and the staff who drive for work — and run those claims as a process rather than case by case.
Pains
Insurance conduct and pricing rules are not the banking rules the governance was built for.
The automation sits at the carrier; the customer complains to the branch.
Generic AI-governance material does not reach the insurance instrument.
Exposure on the institution's own vehicles is handled by whoever is nearest.
Gains
An insurance-specific reading of AI risk that a banking governance function can actually use.
A defensible answer for the supervisor, grounded in the insurance rule as written.
Branch and desk staff who can explain to a customer what the carrier's system did.
Large corporations
Reviewed 2026-09-13 on Carlos's framing. This canvas was already written from the organization's perspective — the company's vehicles, its employees who drive, its own policies and its own claims — so the review added rather than rewrote: the organization's own insurers automating underwriting and renewal pricing, and the readiness assessment read at organization level. Nothing was removed.
Value map
Products & services
Ariski platform
The Learning Program
AI strategy consulting
Pain relievers
A dedicated business track on the car-claims series — company cars, fleets and employees who drive for work: who is liable, which policy responds, and how to run the claim.
One rule page per jurisdiction, each rule cited to the statute or the regulator's own text, so a multi-jurisdiction fleet reads the same fields everywhere.
A claim-readiness checklist and a deadline calculator, free and in the browser, so the first week is a process rather than a scramble.
The readiness assessment is written at organization level across technical, cultural, policy and skills readiness, so risk and compliance can see the gap in the company's own terms.
Gain creators
The same topic readable on all three tracks, so a fleet manager can see what the driver and the adjuster are each being told.
Programme formats that can be delivered to a mixed internal audience — risk, fleet, legal and HR in one room.
A 3D crash reconstruction that produces the diagram, the numbers and the narrative a claim file needs, pedestrians and cyclists included.
Customer profile
Customer jobs
Run a claim when the company is liable for an employee's crash — establish which policy responds and get the vehicle back on the road.
Manage a fleet's claims as a repeatable process rather than as a series of emergencies.
Know what the company's insurer is automating, because it changes what evidence a claim needs.
Bring risk, fleet, legal and HR to one understanding of employee driving exposure.
Know what the company's own insurers are automating in underwriting and renewal pricing as well as in claims, because it changes what the organization has to be able to show.
Pains
Liability for an employee's driving is split across policies; which one responds is rarely obvious.
Fleet claims cross jurisdictions, and deadlines and fault rules change with each.
A first week spent gathering documents is where a claim's value leaks.
The insurer's automated triage and fraud scoring are invisible to the company.
Gains
One written procedure a driver and a fleet manager both follow, scene to repair shop.
Deadlines and fault rules per jurisdiction, as dates.
Teams that understand the automation well enough to feed it what it asks for.
A read on its own AI readiness at organization level.
SMEs
Rewritten 2026-09-13 on Carlos's framing: an SME is addressed as an insured organization — its own vehicles, its own employees who drive, its own policies and its own claims — which is the consumer offer read from the company's side rather than the individual's. The brokerage-and-agency material this canvas previously carried (serving more clients, explaining the carrier's decision to a client, agency tooling) was not an SME job at all but an intermediary's, and has moved to the insurance-brokers-agents canvas where it belongs.
Value map
Products & services
Ariski platform
The Learning Program
AI strategy consulting
Pain relievers
A business track on the car-claims series written for exactly this reader — in the site's own words, fleet managers, HR and risk owners, and small-business owners with one or two vehicles on the books.
One rule page per jurisdiction, each rule cited to the statute or the regulator's own text, so a company driving across borders reads the same fields everywhere.
A claim-readiness checklist and a deadline calculator, free and in the browser, so the company's first week is a process rather than a scramble.
Everything on the platform is free and needs no procurement, no seat count and no signature.
The readiness assessment is written at organization level — technical, cultural, policy and skills — so a small firm can see where it stands without hiring anyone to tell it.
Gain creators
The same topic readable on the consumer and the business tracks, so an owner can see what the driver is told and what the company itself must do.
A 3D crash reconstruction that produces the diagram, the numbers and the narrative a claim file needs, pedestrians and cyclists included.
Short formats in the published programme — talks, workshops, short courses — sized for a firm that cannot release staff for a diploma.
Published in English and Spanish equally, with French next, so the company reads its own rules in its own language.
Customer profile
Customer jobs
Run the claim when an employee crashes a company car, or their own car while working — establish which of the company's policies responds and get the vehicle back in service.
Decide what cover the business actually needs, with no risk manager and no broker on retainer.
Own the first hour after an incident — who is called, what is captured before the vehicle moves — when the owner is also the fleet manager and the person reporting it.
Know what the company's own insurer is automating, because it changes what a claim file has to contain.
Decide whether AI is worth adopting inside the business at all, with no budget and no technology function.
Pains
In a small company the owner is risk manager, fleet manager and the one who reports.
Liability is split between the company's policies and the employee's own.
The first week, spent gathering documents, is where the claim leaks value.
No budget, no time — and the tools sold here are built and priced for carriers.
Gains
One written procedure the driver and the owner both follow.
Knowing which company policy responds before the claim.
Deadlines and fault rules as dates, per jurisdiction driven in.
Usable the same day, at no cost, and a credible position on AI.
Freelancers and independent professionals
Reviewed 2026-09-13 on Carlos's framing, «where it applies». The employability spine of this canvas is unchanged: it is written for the individual professional whose role is being reshaped, which is who the site addresses. What the review added is the organizational side of the same person — an independent professional is their own organization, carrying their own cover and running their own claim with nobody to delegate to. Independent agents, who used to be read here, now have their own canvas at insurance-brokers-agents.
Value map
Products & services
Ariski platform
The Learning Program
Pain relievers
Free, open and bilingual, with no account required to read and no email asked for — the barrier is removed rather than lowered.
Practice items, a prompt library and a prompt builder — the material is something to do, not only something to read.
An explicit «what not to do» alongside every «what to do», including client data and unexplainable decisions.
The self-learning mode of the programme exists precisely for the professional whose employer is not paying.
Four authored voices in Ask Ariski, so an awkward question can be asked of a machine before it is asked of a manager.
The business track of the car-claims series covers an employee's own car driven for work, and the per-jurisdiction rule pages give the deadline and the fault rule as written — the same material a company uses, available to a practice of one.
Gain creators
A structured path that exists on the site already: an AI 101 for insurance, then what to do and what not to do, then prompt engineering, then practice, then the programme.
A readiness assessment and an adoption roadmap that tell a professional where they actually stand, not where a course syllabus assumes they do.
A glossary and a tool directory to settle a term or a product question in a minute.
Assessment inside the programme, so competence can be shown to an employer rather than asserted.
Customer profile
Customer jobs
Stay employable in a role that is being reshaped while they are doing it — underwriter, adjuster, actuary, compliance officer, independent agent.
Understand what the tool that appeared on their desk actually does, and where it fails.
Learn without a budget, without employer sponsorship and without leaving the job.
Know what not to do — with client data, with a model's output, with a decision they cannot explain.
Enter the profession already literate, for those still in training.
Carry their own cover and run their own claim when the vehicle they work out of is damaged — no employer's policy behind them and nobody to delegate the file to.
Pains
The honest question — «will this take my job?» — has nowhere neutral to be asked.
Generic courses have only generic examples.
Serious material is English-only, excluding most of a Spanish-speaking market.
No time: learning loses against the caseload.
A work vehicle sits between personal and commercial cover.
Gains
Competence that transfers to tomorrow's file, learned at their own pace.
The site's own thesis: professionals who understand AI will replace those who do not.
A vocabulary that holds before vendor, regulator and engineer.
Knowing which of their own covers responds before an incident.
Government and public sector
Value map
Products & services
Ariski platform
The Learning Program
AI strategy consulting
Pain relievers
Ariski is not a vendor and not a supervised entity, which is the only position from which this material can be read as disinterested.
A jurisdiction guide and a car-claims rule set built by reading the primary instrument, with anything unread published as «in verification» rather than asserted.
An ethics simulator built on a fictional scenario of an insurer's AI underwriting model going wrong and policyholder data being exposed — a training case that needs no real firm to be named.
The free platform requires no procurement at all, which is how a supervisory team can start before a budget cycle allows anything else.
Gain creators
Content that states where AI fails, which is the half of the subject a supervisor needs and a vendor will not publish.
Programme formats deliverable to a supervisory team as a closed cohort, in the modes the site already publishes.
Customer profile
Customer jobs
Supervise insurers that are automating decisions faster than supervisory guidance is written.
Judge whether an automated underwriting or claims decision is explainable, non-discriminatory and open to recourse.
Bring supervisory staff to a level of AI competence that does not depend on the supervised firm's own explanation.
Pains
The technical explanation a supervisor is given usually comes from the firm being supervised, or from its vendor.
Training budgets and procurement cycles move more slowly than the technology being supervised.
Material written for one market's rules does not transfer to another jurisdiction's instruments.
Gains
An independent technical reference that owes nothing to any supervised firm.
Concrete harm cases to reason from — algorithmic bias in pricing, opaque decisions, recourse for the insured.
Material in Spanish as fully as in English, for jurisdictions where that is the working language.
Generated by sitecraft-system/skills/investor-kit-canvases from the CEMI business profile for Ariski, this kit’s growth.json and its single ROUND. Nobody hand-edits these; to change one, change the data.
Financial terms, revenue and detailed materials are available on request. Figures describe capacity built and operating today; forward-looking statements are identified as such.
Scan for the investor kit