Frequently Asked Questions

It's when AI makes something up but says it like it's true. For example, an AI chatbot might give medical advice that sounds correct but is actually wrong or dangerous. The AI doesn't know it's lying — it just generates text that sounds good. This happens more often than you'd think, and when someone gets hurt, they blame you.

Regular logs can be changed or deleted. Anyone with access — a disgruntled employee, a hacker, even your own IT team — can alter them. In court, a lawyer will ask: "How do we know these logs weren't modified?" With ProlixoTech, records are locked forever the moment they're created. Not even we can change them. That's what makes it real evidence.

The big one is the EU AI Act (Regulation 2024/1689) — if you do business in Europe, you must prove what your AI did. Article 12 requires automatic logging for high-risk systems. Penalties can be significant and vary by violation type. In the US, the Caremark doctrine requires boards to implement reasonable oversight systems for material risks.

ProlixoTech doesn't prevent AI mistakes — no one can do that. What it does is prove you weren't careless. If your AI gives bad advice, you can show: "Here's exactly what it said, here are the safety rules we had in place, and here's proof we were monitoring it." That's the difference between "you were negligent" and "it was an unfortunate accident despite your best efforts."

Think of it like a security camera for your AI. Every time your AI talks to a customer or makes a decision, ProlixoTech automatically takes a snapshot with a timestamp. This snapshot is then "sealed" using special technology (like a digital notary stamp) so no one can tamper with it. If anyone ever questions what your AI said, you have proof.

Most companies are up and running in less than a day. Your tech team adds a few lines of code to connect your AI system to ProlixoTech. There's no complicated setup, no hardware to install, and no training required. Once connected, everything happens automatically in the background.

No, not at all. ProlixoTech works quietly in the background. Your AI responds to customers just as fast as before. Your users won't notice any difference. We designed it this way because we know speed matters — we'll never be the reason your AI feels slow.

Any AI that talks to customers or makes decisions. This includes chatbots, customer service AI, AI assistants, recommendation engines, content generators, and more. If you're using OpenAI, Anthropic, Google AI, or your own custom models — ProlixoTech works with all of them.

It's an official document we create for you every month. Think of it like a report card that says: "This company was watching their AI and had safety measures in place." You can print it, show it to your board of directors, hand it to auditors, or use it in court. It's your receipt that proves you did everything right. Available on Pro and Enterprise plans.

AI laws change constantly. If you're sued in 2028 for something your AI did in 2025, the plaintiff's lawyer will try to judge you by 2028 standards — that's not fair. The "Time Machine" Defense solves this. We capture a cryptographic "snapshot" of your exact safety rules at the millisecond of each AI decision. In court, you don't just say "we were safe" — you provide a SHA-256 hash that proves your 2025 policies were followed to the letter.

Standard database logs can be edited. A lawyer will claim you "Photoshopped" your logs after the accident to make yourself look innocent. Our "Digital Wax Seal" makes that attack fail. We use AWS Hardware Security Modules (the same technology banks use) to digitally sign every log. The private key never leaves Amazon's physical hardware — it's literally impossible for anyone to export it or fake a signature. Combined with special formatting rules (RFC 8785), your evidence is unimpeachable.

The EU AI Act mandates logging and traceability for high-risk AI systems. Penalties vary by violation type and can be significant. We call this feature the "Homework Passer." Instead of reading 200 pages of regulations, you just use our SDK. We automatically track everything Article 12 requires — which databases your AI consulted, whether a human reviewed the decision, and complete decision trails. Compliance becomes a background API call, not a legal project.

Boards increasingly need documentation showing AI oversight programs are in place. Our monthly "Certificate of Truth" helps with this. Each month, we generate a signed document containing cryptographic proof of AI governance activities. This supports your documentation of reasonable oversight — but consult legal counsel for specific compliance requirements.

No — we add less than 1 millisecond of overhead. Developers hate when security tools slow things down. If logging makes your chatbot take an extra 200ms to respond, users will notice and complain. We designed ProlixoTech like a "dashcam" — it captures evidence in microseconds without blocking your main application. Your AI "throws" the log into memory and continues immediately. A background process handles the rest invisibly.

Yes — our sharded architecture handles enterprise-scale event volumes. Traditional logging creates a bottleneck. If 10,000 users log at once, they form a single line, causing the database to crash. That's a disaster during your big product launch. We use a technique called "Probabilistic Write Sharding" — think of it like having 20 checkout lanes instead of one. Logs are distributed across parallel channels so no one waits in line.

Storing millions of tiny log files directly in S3 incurs per-object overhead costs. We solved this with smart batching. Using Kinesis Data Firehose, we compress many small logs into columnar Parquet blocks before archiving. These blocks then transition through S3 storage tiers (Standard → Infrequent Access → Glacier) for long-term compliance retention at optimized costs.

You do. 100%. We store it securely on your behalf, but it's your data. We never sell it, share it, or use it for anything except protecting you. You can export it anytime, and if you ever leave (though we hope you won't), you can take your records with you.

Your data is encrypted both at rest (AES-256 via AWS KMS) and in transit (TLS). We follow security best practices aligned with industry standards, including role-based access controls and audit logging. Multi-factor authentication is available and recommended. Only you and the people you authorize can access your records.

We have a free plan for small projects (up to 1,000 AI decisions per month). Our Pro plan is $149/month and covers most growing businesses with up to 100,000 AI decisions. Enterprise plans start at $499/month. Compare that to the enterprise deals you're losing without verifiable trust — it pays for itself.

The free plan is free forever and includes:

• Up to 1,000 AI decisions per month
• Tamper-evident evidence for all decisions
• 7-day active evidence retention (archival policy)
• Basic dashboard to view your records
• Community support

It's perfect for small projects or trying ProlixoTech before upgrading.

Yes. You can upgrade or downgrade your plan at any time. If you upgrade, you get access to new features right away. If you downgrade, the change happens at the end of your billing period. All your existing data stays safe when you change plans.

Yes! We offer 50% off for the first year for early-stage startups. If you're part of an accelerator program (Y Combinator, Techstars, 500 Global, etc.), just let us know and we'll set you up with special pricing. We want to help startups build responsibly from day one.