Oct 2026Release log4 October 2026
The René Engine has learned from 2,238 researched journeys across 72 countries, including seven new industries. Workspaces can now give René their own business context, and every journey, stage and solution is written toward it. Mature now means the engine builds an industry’s everyday journeys on its own, and every ritual is checked against 15 rules in code before it can go Live.
Business context
9 changes- New: a Business context library in Settings › Context. Until now René only knew what was typed into a single request, so every journey started from a blank brief. Now a workspace keeps the documents that describe who it is, and René reads them every time it writes.
- Six short brand statements have quick-start cards: CX vision, mission, vision, ambition & goals, brand values and customer promise. Longer documents can be added for brand voice, strategy & priorities, customers & segments, standards & guidelines, research & insights, an existing journey, or anything else René should know.
- Add context by typing, pasting or uploading a file: PDF, Word, PowerPoint, Excel, text, CSV, Markdown, or an image, which is read with vision. Each document can be up to 10 MB and 60,000 characters, and its type is guessed from the file name (a file called "Q3 NPS survey" lands under Research & insights).
- Choose what each journey reads. Library items marked "Use automatically" go into every new journey. When you create a journey you can untick them, pick others, or attach documents to that journey only, with the option to also keep them in the library. The journey remembers its choice, so everything generated for it later reads the same context.
- It shapes the work, it isn't decoration: moments that deliver the CX vision, mission, ambition or customer promise become highlights, and moments that break them become pain points. Brand values and voice shape how touchpoints are described, and every stated standard, priority and constraint is respected.
- Documents change what René does with them. An existing journey is followed: its stages, steps and wording are kept and René only adds what is missing. Research and survey findings become evidence for pain points, highlights and emotion scores.
- Business context is used when René creates a journey, generates a stage, enhances a journey, and writes templates and solutions. Solutions must also move the brand toward its stated vision, ambition and promise, fit its values and voice, and respect its standards.
- Short brand statements are read first (up to 3,000 characters each). Longer documents share what is left (up to 12,000 characters each, 24,000 in total), and anything shortened is marked as shortened. René refers to the organisation as "the brand", never quotes your documents at length and never mentions that they were provided.
- Context stays in its workspace: a journey can only read documents from the workspace it belongs to. If a document can't be read, the journey is still made, just without it.
Industry research
9 changes- Education (new): 152 researched journeys across 68 countries, plus research that applies worldwide. It covers school admissions, university applications, scholarships and student life, from GEMS schools, Dubai College and the Zayed University admission in the UAE to India's CBSE, NIOS and AIIMS counselling, Brazil's ProUni and SiSU, Egypt's Tansik, Australia's UAC, the Toronto District School Board, ETH Zurich, TU Berlin, Politecnico di Milano, CROUS in Paris and Duolingo. About 3,800 rules were learned, such as admission and visa requirements, scholarship conditions and school inspections like ADEK's in Abu Dhabi.
- Fintech (new batches): 172 researched journeys across 25 markets, loaded only after a quality review of each batch. They cover wallets, pay-later, payments, investing and business accounts at about 100 providers, including PayPal, Mercado Pago, Trade Republic, stc bank, DBS PayLah!, Careem Pay, MB WAY, Tabby, Tamara, Kakao Pay, Weixin Pay, Alipay, PhonePe, Kaspi, GrabPay, Wealthsimple, Toss and Ziina. About 480 rules, such as UAE trade-licence and business (KYB) checks, beneficial-ownership disclosure, card-scheme dispute rules, SAMA digital-bank licensing, identity checks through Nafath and Absher, and capital-gains tax documents.
- Mixed-use destinations (new industry): 150 researched journeys across 10 markets and 52 destinations, including Hudson Yards, Canary Wharf, Battersea Power Station, Marina Bay, Tokyo Midtown, Grand Front Osaka, Potsdamer Platz, Barangaroo, Expo City Dubai, City Walk, Al Seef, Liverpool One and the V&A Waterfront. They cover visitors, residents and tenants, from a cinema night or a museum visit to a gym membership, a short stay or a corporate event, and about 780 rules for leases, tenant onboarding and move-out, and deposits.
- Spa & wellness (new industry): 150 researched journeys across the UAE, Saudi Arabia, the UK, the US, Australia and Singapore, at 60 spas, hotels and clubs including Four Seasons, Waldorf Astoria DIFC, The Ritz-Carlton, Raffles, Banyan Tree, The Jeddah EDITION, Talise Spa, Champneys, Thermae Bath Spa, Bannatyne, Massage Envy, Hand & Stone, endota and Fitness Time. About 770 rules, such as clear-pricing and surcharge rules, how health information may be collected, age conditions, health checks before a treatment, and gift-card refunds.
- Media (new industry): 128 researched journeys across 11 markets, at streaming, music, sport and news services including Netflix, Spotify, OSN+, Shahid, TOD, beIN, STC tv, YouTube TV, ABC iview, Stan, NHK ONE, U-NEXT, Sky, Canal+, Le Monde and The Washington Post, with about 600 rules on subscriptions, cancellation and privacy. A further 24 journeys about cinemas and live screenings were filed under Events & Venues.
- Self storage (new industry): 93 researched journeys across 10 markets and 43 operators, including Extra Space, Public Storage, Big Yellow, Shurgard, Storage King, National Storage, Kennards, StorHub and The Box, with about 290 rules on renewals and fees, cancelling before move-in, notice periods, move-out and the final bill.
- Laundry & dry cleaning (new industry): 57 researched journeys across 10 markets and 25 brands, including Laundryheap, Washmen, Johnson Cleaners, Jim's, Rinse, Hakuyosha, Tide Cleaners and Pressed, with about 150 rules such as Japan's Cleaning Business Act, price checks before cleaning and how pick-up and locker services may use personal details.
- Earlier this month: Events & Venues (234 journeys), Banking (154), Airlines (154), E-commerce (153), Automotive (151), Car rental, Beauty salons and Business clubs (150 each), plus 58 journeys across 18 other industries, led by SaaS, government services and retail.
- In total the engine has now learned from 2,238 researched journeys across 72 countries, up from 1,318 across 52 at the start of the month, with about 15,500 market-specific rules and regulations.
Industry maturity
6 changes- Mature now means two things, not one. An industry needs a knowledge score of at least 85 (enough researched journeys, markets, independent sources and detail), and the engine must build at least half of that industry's everyday requests on its own, without handing the journey to René AI to write from scratch.
- Each researched industry has eight fixed everyday requests, written the way a client would type them, such as "Business traveller flying Emirates from Dubai to London" or "Buying a used car" in the UAE. They are tested every night and never changed to flatter a score, so the result can be compared from one night to the next.
- Under this stricter test three industries are Mature: Events & Venues, Automotive and Education (95 each).
- Nine more know enough to be Mature but are held at Established until the engine builds more of their everyday requests itself: Banking (94), Airlines, Fintech, E-commerce and Beauty salons (93), Car rental (92), Business clubs (91), Mixed-use destinations (89) and Spa & wellness (85).
- Media, Self storage and Laundry & dry cleaning are Developing. Their content already scores 70 to 86, but an industry is held back until it has a full batch of 150 researched journeys.
- The console shows what is holding each industry back, for example "Engine builds 3 of 8 everyday requests itself", "Single source, nothing independent to compare against" or how many touchpoints carry pain points and highlights.
What the engine learns from research
9 changes- Every stage is now placed in its industry's standard phases, for example 19 for Events & Venues, 15 for Car rental and 13 for E-commerce, so journeys from different sources line up.
- Some research batches give every stage its own code (1,389 codes for 1,389 stages in Mixed-use), so codes can't be compared. For those industries the engine now groups stages into standard families by what happens in them.
- Customer types are now industry-specific, such as first-time buyer, EV first adopter or company driver in Automotive, season-ticket holder or exhibitor in Events, subscription or cross-border buyer in E-commerce, and local resident or everyday visitor in Mixed-use. People with accessibility or medical needs are only matched when a journey asks for them.
- Pain points are now tagged with one of 32 standard friction types, such as waiting, unclear pricing, wayfinding or crowding, while the researcher's own wording is kept.
- Each touchpoint now records what kind it is (person, device, place, item or digital) and who runs it (the brand, a partner, a third party or the customer).
- The engine now also learns KPIs and measurements, behavioural levers, stage durations, exit signals, the reasons behind each emotion score, and the official sources cited for each rule.
- Market rules are now learned with their country, their scope (law, regulator, brand policy or event) and the date they were checked.
- Research about leaving, such as a cancellation, notice, withdrawal or refund, now teaches the engine its rules but no longer becomes the base of an everyday journey. It is only used to build a journey when the request is about ending something.
- Wording repeated across many research files now counts once, and generic disclaimers and placeholder text are left out.
Rituals engine
10 changes- Rebuilt the Rituals engine. Before, one AI pass wrote a whole ritual plan, and the only check was that each ritual pointed at a real stage and touchpoint. Now a ritual goes through seven stages: read the journey, choose where rituals belong, classify each one, brief the writer, write, critique and revise. Choosing, classifying, critiquing and the decision to go Live are worked out in code, not by the AI.
- Choosing where rituals go is now worked out from the journey itself, not left to the AI. Every touchpoint is scored on how bad it feels (EXIS), how severe its pain points are, how much it matters, whether the brand controls it and how many customers it reaches. Rituals then fill six slots in order: up to three top valleys, recovery for the steps most likely to fail, a bridge for low points the brand doesn't run, plus opening, closing and the core product. Pain points with the same root cause share one ritual, and moments that already go well are left alone.
- Each ritual is written as a full working brief: a trigger and who it's for, numbered steps with who does them, how long they take and what to say, an example line under 20 words for each must-say step, must-do and must-never rules, a minimum version for busy moments, variants for each archetype, training or a system handoff, how it's measured, and a Yes/No checklist for auditors. System rituals get their states, a time limit and acceptance criteria instead.
- Every ritual is checked against 15 rules written in code before it can be saved, so the AI never grades its own work. The rules include: the choreography is complete, every step is done by someone who is actually there, each step fits its time budget, every must-say step has an example line in every language, emotions and principles come from fixed lists (25 emotions, 18 behavioural principles), and auditors get at least three Yes/No checklist items tied to steps.
- The engine no longer makes things up. Any price, fee, weight, percentage or time limit that isn't in the client's input pack fails the check. So do placeholders like "[brand name]" or "the airline", and personal details such as seat numbers, passport numbers or booking references.
- Those 15 rules add up to five tests every ritual has to pass: Repeatable, Ownable (a competitor couldn't run it unchanged tomorrow), Trainable (a frontline employee can deliver it after one briefing), Observable (an auditor can score it Yes/No) and Measurable (it has a baseline, a target and a metric).
- If a ritual fails, it goes back to the writer with the reasons, up to twice. Anything still failing is saved with its reasons and stays at Draft. A ritual can only go Live once every check passes and the client has confirmed every assumption the engine had to make. Edits are checked again straight away.
- Clients can give the engine an input pack: their brand signature, real figures, offers and policies, the roles on the floor and the languages they serve. This is the only source the engine may take facts from.
- Added the Ritual Studio: a ritual map over the journey, a list, and Design, Pocket card, Guide and Audit views for each ritual, plus a critic panel showing every check and the five tests, an edit dialog, and status from Draft through Pilot to Live and Retired.
- Rituals are written in English and Arabic, and any language the journey is already translated into can be added. Example lines and system copy are translated, and the language check covers each of them. The downloadable PDF is now a full written brief with a ritual map, contents and every section of every ritual, with searchable text in every language.
Journey engine
6 changes- Changed how engine-built journeys are made: instead of copying the single nearest researched journey, the engine combines several close matches and keeps what they agree on.
- The engine builds a journey only when its research covers enough of the request, about two thirds of what the journey should be. Otherwise René AI writes it from scratch, informed by the same market rules.
- Every market rule now records when it was last checked. Rules past their shelf life are softened or dropped, rules for past events are used only as examples, and rules are re-checked against their official sources every week.
- One EXIS everywhere: stage, journey and workspace scores are now the importance-weighted average of their touchpoints, so the score on a stage header, a journey card and the Rituals map always agree. The old peak-and-end score is shown beside the journey EXIS as "Remembered experience".
- Channels and touchpoints are now separate: the channel is how a step happens and the touchpoint is who or what the customer deals with, for example "Website (Emirates)" or "Gate agent".
- Importing a journey is more forgiving: a journey map with stages and steps but no touchpoints now imports instead of failing, and files are read the same way as business context (PDF, Word, PowerPoint, Excel or an image).
Journey Vault
5 changes- Upgraded the spine of the René Engine: researched journeys are now filed into the Journey Vault as journey types, and each type's phases are written from the journeys filed under it.
- Events & Venues now has 28 journey types, Automotive and Mixed-use destinations 23 each, Airlines 21, Self storage 11 and Laundry & dry cleaning 9. Airlines grew from 3 types to 21, with 68 phases and 246 stages, and Automotive from 3 to 23, with 84 phases and 322 stages.
- Every phase is now reviewed for each market it covers, using the rules learned for that country. 27 markets were reviewed for Airlines and Automotive.
- Added 22 new journey templates and 11 new Studio archetypes built from the research.
- When a journey is built, the Vault now picks the journey type by the words that are most specific to the request, so common words like "car" or "flight" no longer pull in the wrong type.