Trust & Safety

AI Disclosure

Last reviewed: September 4, 2026

Think4Ei uses artificial intelligence in several places on this website and across the platform. This page explains, in plain language, when you are interacting with AI, what it can and can't do, and how to get a human involved.

When you're interacting with AI on this site

"Talk to Eli" chat widget — a conversational assistant on the homepage. Every reply is generated by a third-party AI model (see below), routed through our own server, not typed by a human in real time.
Savings / ROI estimate calculator — produces an AI-assisted narrative estimate based on the numbers you enter, alongside deterministic calculator math.
Visitor research summaries — only when you explicitly opt in via an on-page consent checkbox, we compile a research summary about your organization from public sources, using AI to help synthesize publicly available information for a sales conversation. This never runs silently or without your logged consent.

None of these AI features run inside your browser with a hidden API key — all AI calls are proxied through our own backend, so no model-provider credential is ever exposed to your device.

What model powers this

The AI features on this site are powered by Anthropic's Claude models, accessed through Think4Ei's own server-side integration. Anthropic is listed as a subprocessor at Subprocessors. Across the wider Think4Ei platform, individual client-facing "Eli" assistants embedded in specific systems may also use locally hosted, Think4Ei-operated models for some functions — where that's the case, it does not involve sending your data to Anthropic.

What's AI-generated vs. deterministic

Not everything on a page that "feels AI-driven" is generated text. For example, the savings/ROI calculator's numeric estimate is computed with a fixed formula; only the narrative explanation around it is AI-generated. Where a response is materially AI-generated or AI-assisted, it is presented inside a clearly labeled chat or assistant interface — we do not present AI-generated text as if it were a human-authored statement from Think4Ei staff.

The role of public sources

Where an AI feature references facts about your organization (the consent-gated research summary above), those facts are drawn from publicly available sources at the time of the request — things like your organization's own public website, public filings, or public news coverage. We do not use non-public data sources for this feature. Public information can be outdated, incomplete, or simply wrong at the source, and AI summarization can compound that — treat any AI-generated summary about your own organization as a conversation starter, not a verified record, and let us know if something is inaccurate (see below).

Results may be incomplete or incorrect

AI-generated and AI-assisted output on this site — chat replies, savings narratives, and research summaries alike — can be incomplete, out of date, or simply wrong. We do not claim these systems are free of error, and we do not present them as a substitute for professional legal, financial, or compliance advice.

How assumptions and sources can be reviewed

For the savings/ROI estimate, the underlying inputs you provided and the calculator logic are visible on the page itself. For chat and research-summary output, you can ask Eli directly to explain its reasoning or list what it based an answer on; we are continuing to expand automatic, always-visible source citations across AI-assisted features platform-wide, and today that coverage is not yet complete on every surface — where a citation isn't shown automatically, asking for it is the fastest way to check an answer.

When human review is required

Under our Terms of Use, AI-generated output from this site or the platform should not be used as the sole basis for a consequential legal, financial, medical, hiring, or compliance decision without human review. Within the platform itself, many workflows that carry real consequences (for example, approving a hire, releasing a budget hold, or finalizing a compliance finding) are deliberately built with an explicit human approve/reject step rather than an automatic AI action — the AI assists, a person decides.

How to report an error

If an AI-generated response on this site is inaccurate, inappropriate, or just unhelpful, email support@think4ei.com with what you asked and what was returned (a screenshot helps). We review these reports; they also feed the platform's broader effort to reduce AI errors over time.

Are conversations stored?

Yes. Chat and research-tool interactions are logged server-side so we can provide support, investigate reported errors, and improve the service, consistent with our Privacy Policy and Cookies & Analytics Notice. We do not sell this data, and typed conversation content is handled the same way as other information you directly provide to us (see "Information you give us directly" in the Privacy Policy) — it is not merged with the separate, anonymous, passive analytics pipeline described in the Cookie notice.