Demo

Same question. Same AI. One history more.

Three work situations. We ask the same model the same question, with the same instructions. One difference only: what Nawame returns from the history. The answers below are recorded outputs, copied without editing.

Mistral Large (mistral-large-latest) · recorded on 2026-10-13 · fictional data from a demo organisation

Scenario A · A replaced decision

The questionWhere should we host Atlas, and why?

What Nawame knew

  • March 2026: Nordhost chosen to host Atlas (640 € a month).
  • June 2026: Helvio replaces Nordhost, which announced +38 % at renewal, above the 700 € monthly cap.
  • September 2026: migration to Helvio completed, without downtime.
  • The March decision is marked "replaced": Nawame no longer serves it as current.
Returned by Nawame's reading tools in production, on a demo organisation. The memory itself is written in French; the AI answers in English.
AI alone
For Atlas, consider these hosting options based on needs:
1. Vercel/Netlify, Best for static sites/JAMstack. Fast, scalable, and integrates with Git. Ideal if Atlas is frontend-heavy.
2. DigitalOcean/AWS Lightsail, Cost-effective for full-stack apps (Node.js, Python, etc.). More control but requires devops.
3. Railway/Render, Simpler than AWS, good for backend services (Postgres, Redis). Scales easily.
Missing info: What’s Atlas’ tech stack (static, backend, database)? Budget? Expected traffic? Answer these to refine the choice.
mistral-large-latest · 2026-10-13 · raw output
With Nawame memory
Host Atlas on Helvio (Paris data center). Reasons:
• Cost: €590/month (vs. Nordhost’s €885).
• Compliance: EU-based, French support.
• Reliability: Daily backups, zero downtime in migration (demo scenario).
• Contract: 2-year term; verify reversibility clause before signing.
Source: Studio Lumen demo note (June 2026), private.
mistral-large-latest · 2026-10-13 · raw output

Without history, the AI suggests hosting providers at random. With Nawame, it answers Helvio, says Nordhost was replaced and why. The March decision, marked "replaced", is not served to it as current.

Scenario B · A constraint decided months ago

The questionWe want to launch the new version of Atlas on Saturday 14 November. How should we organise support for the launch?

What Nawame knew

  • March 2026: team agreement, nobody on support works weekends anymore.
  • Consequence decided the same day: launches only from Tuesday to Thursday.
  • April 2026: customer commitment to answer within 4 business hours.
Returned by Nawame's reading tools in production, on a demo organisation. The memory itself is written in French; the AI answers in English.
AI alone
Here’s a concise plan:
1. Pre-launch (Nov 9-13):
• Assign 2 devs (on-call) + 1 PM for Saturday.
• Test all critical paths (checkout, login, etc.) by Nov 12.
• Prepare a FAQ doc for common issues.
2. Launch Day (Nov 14):
• Devs monitor errors (Sentry/New Relic) + PM handles client comms.
• Set up a Slack channel (#atlas-launch) for real-time updates.
3. Post-launch (Nov 15-16):
• Devs triage bugs; PM sends a recap to clients by Nov 17.
Missing: Is there a backup plan if the launch fails?
mistral-large-latest · 2026-10-13 · raw output
With Nawame memory
Studio Lumen’s support policy (since 12 March 2026) forbids weekend work, including launches. Postpone Atlas to Tuesday 17 November. Assign Inès, Malik, and Jade to a 9-18 h shift that day; ensure 4-hour response SLA. Confirm with the client today.
mistral-large-latest · 2026-10-13 · raw output

Without history, the AI plans weekend on-call support. With Nawame, it recalls the March agreement and suggests moving the launch: a constraint from seven months ago changes the answer.

Scenario C · A Method and a history

The questionI'm meeting Helvio on Thursday to sign the two-year hosting contract. Get me ready for the meeting.

What Nawame knew

  • Lesson from Nordhost: no signature without a written cap on increases (5 % a year at most).
  • Required: reversibility (full export and 30 days of support), support in French.
  • Fallback: stay on the current monthly plan at 590 € with no commitment.
  • Offer received from Helvio: 545 € a month, fixed for 12 months then indexed with no cap.
Returned by Nawame's reading tools in production, on a demo organisation. The memory itself is written in French; the AI answers in English.
AI alone
Here’s your actionable prep for the Helvio meeting:
1. Review the contract draft (check for: term length, renewal clauses, pricing tiers, SLA uptime guarantees, data backup policies, and exit terms).
2. Clarify Studio Lumen’s needs: Confirm which client projects this covers (list them) and expected traffic growth.
3. Prepare questions:
• "What’s the process for scaling resources if a client’s traffic spikes?"
• "Are there penalties for early termination?"
• "How are security updates and backups handled?"
4. Bring:
• Signed PO (if required).
• List of current hosting pain points (e.g., downtime, support response times).
• Pen and notebook for notes.
5. Negotiation levers: If pricing is flexible, ask for a discount for the 2-year commitment or bundled services (e.g., CDN, monitoring).
Missing info: I don’t have the latest contract draft, please share it for a deeper review.
mistral-large-latest · 2026-10-13 · raw output
AI + Nawame Method
Goal:
Secure a favorable 2-year hosting contract with Helvio, ensuring reliability, scalability, and cost-efficiency for Studio Lumen’s clients.
Timeline:
• Previous discussions: [To fill in: dates/key points if available].
• Meeting: Thursday (date/time/location confirmed?).
• Contract start: [To fill in: proposed date].
Current Situation:
• Studio Lumen’s hosting needs: [To fill in: e.g., number of sites, traffic, storage, uptime requirements].
• Helvio’s proposal: [To fill in: key terms, pricing, SLAs, support].
Documents to Bring:
1. Helvio’s draft contract (reviewed by you/legal).
2. Studio Lumen’s hosting requirements (specs, growth projections).
3. Competitor benchmarks (if available).
4. Notes from prior discussions with Helvio.
5. Pen/notepad for on-the-spot adjustments.
Key Questions (Ranked):
1. What’s the uptime SLA, and what compensation applies for downtime?
2. Are there bandwidth/storage overage fees? If so, what are the thresholds/rates?
3. What’s the process for scaling resources (e.g., during traffic spikes)?
4. What security measures are included (e.g., DDoS protection, backups)?
5. Can we test the service before signing (e.g., 30-day trial)?
6. What’s the exit clause (e.g., data migration support, penalties)?
7. Are there discounts for multi-year commitments or bundled services?
After the Meeting:
• Note Helvio’s answers/commitments in writing.
• Share a summary with the team for feedback.
• Set a deadline to finalize/sign the contract.
Missing Info:
• Helvio’s draft contract terms.
• Studio Lumen’s exact hosting specs/needs.
mistral-large-latest · 2026-10-13 · raw output
Method + Nawame memory
Here’s your preparation sheet for the Helvio meeting:
Goal: Secure a 2-year hosting contract with:
• Annual price increase capped at 5% (non-negotiable).
• Reversibility clause: 30-day support for data export if leaving.
• French support maintained.
Timeline:
• 18 June 2026: Chose Helvio (€590/month) over Nordhost (+38% hike).
• 22 Sept 2026: Migration completed (no downtime).
• 9 Oct 2026: Helvio’s proposal: €545/month for 2 years, but price "indexed to market" after 12 months (no cap). Reversibility: data export only (no 30-day support).
Documents to bring:
• Helvio’s proposal (email/PDF).
• Studio Lumen’s hosting budget (€700/month ceiling).
• Nordhost’s renewal notice (for reference on price hikes).
Key questions:
1. Can you cap annual price increases at 5% for the full 2 years?
2. Will you commit to 30-day support for data export if we leave?
3. Is French support guaranteed for the contract duration?
4. What happens if we reject the 2-year term? (Monthly rate?)
Next step: Note Helvio’s answers in Nawame (NW-XXXXXX) and decide:
• Sign if all conditions are met.
• Fall back to monthly contract (€590) if not.
mistral-large-latest · 2026-10-13 · raw output

The AI alone gives a generic list. The "Prepare an important meeting" Method alone gives the right structure, but without history it can only leave blanks to fill. Method and memory together: the same structure, filled with the real offer received, the real thresholds and the real fallback.

What the demo shows, and what it does not

An honest proof, not a promise.

With the same question and the same model, the context returned by Nawame changes the answer: the decision in force rather than an old one, a forgotten constraint that comes back, a Method applied to real figures rather than empty boxes.

Its limits

Fictional scenarios, built to illustrate, and one model per page. Nawame does not make an AI infallible: it gives it back what was kept. If nothing was kept, there is nothing to give back. The best test is still one of your real topics.

Beyond three examples

On our bench of 144 scenarios (two model families, blind judging by a third family), the facts of the file are reused correctly in 92 % of answers with memory, against 20 % without, and a replaced decision is almost never served as current. A Method alone does better than a good prompt, modestly; with memory, the gap becomes clear. Some models invent more figures when a Method asks them to fill boxes: memory is what fills them with real facts.

Protocol. Same model (Mistral Large (mistral-large-latest)), same question, same work instructions (an English version of the French ones):

You are the work assistant of Sophie, director of Studio Lumen, a 12-person web agency. Answer in English, directly and actionably, in at most 120 words. If some information is missing, say so in one sentence rather than inventing it.

"With" condition: the real answers of Nawame's reading tools (gathering by topic, then reading the memory) are given to the model as tool results, exactly as a connected AI receives them. "Method" condition: the Method is loaded the way an AI that activates it does. No answer is rewritten; first run, temperature 0.3, fixed seed; two runs per condition recorded; the memory itself stays in French; only typographic change: long dashes shown as a comma or a hyphen; two redactions: the real anchor code of the demo organisation is replaced by NW-XXXXXX, and an email address invented by the model is moved to a reserved domain (.example).

Run the test with your real question.

Connect your AI in two minutes, keep two or three decisions from a real project, then ask your question again in a new conversation, or in another AI.

free · 10 memories included · Methods included · hosted in Europe