Conversion agents with memory for your website

IntellQ by Apavin is a set of conversion agents that sit between your website and each visitor. They read intent from behaviour, remember every visitor, lead and customer, learn from businesses in the same industry, change only the page elements you marked, and keep a change only if it beats a holdout control group. It is built by Apavin in India and starts free.

Three layers. Each person it has met: whether they are a visitor, a lead or a customer, what they responded to, and any strategy they turned down, which it never shows them again. This business: the Conversion Genome of journeys, frictions, attention and which strategies worked, rebuilt every night. And businesses like this one: behaviour patterns pooled across an industry, as counts only.

No. It gets the pattern, not the data. Every brand's memory stays sealed in its own workspace. What joins the industry pool is counts: how many visitors in each context saw each strategy and converted, journeys as page roles such as product then cart then checkout, and how each context converts. No pages, copy, offers, event names, customers or memory ever cross. A pool is used only once three other businesses are in it, a brand reads it minus its own share, and the whole industry counts for at most 40 visitors per strategy, so a brand's own evidence takes over quickly. Contributing can be switched off per workspace.

Every memory the agents write is sealed with AES 256 GCM before it reaches the database: visitor, lead and customer memory, the Conversion Genome, each workspace's contribution to its industry pool and the pool itself. Each workspace has its own key, derived with HKDF from a secret held in the deployment environment and never stored. Each sealed record is bound to its workspace and the person it belongs to, so a row copied into another workspace does not open.

It reads how every visitor behaves on your website, including which blocks of a page held their attention, and puts them in a context such as pricing friction or checkout recovery. When a context has enough visitors behind it, it writes a bounded change to the page, shows it to you in a sandbox, runs it against a control group that sees the page unchanged, and keeps it only if the treatment converted better. Anything that converts worse is rolled back.

No. It can change the text of a slot you marked with data-intellq-slot, show or hide one, or point one at another page on the same site. That is the whole list. It cannot inject markup, run script, load anything from another domain, or touch an element you did not mark. The browser code matches slots by attribute rather than by a selector it built, and it writes text with textContent, so a change is text and never markup.

The script watches each block of the page with an IntersectionObserver and counts the seconds it was at least half on screen while the tab was in front, plus clicks inside it. One event per page, at most twelve blocks. Nothing is read from inside a block: no text, no field values, no cursor tracks, no session replay. That gives a heatmap of attention per block, which the agent uses to tell a page somebody read from one they scrolled past.

A credit is one unit of agent work. 1 credit is one visitor read and decided on, charged at most once per visitor per day. 25 credits is one proposed change, written with its evidence and a sandbox preview. Tracking, the live board and reports cost nothing. When the credits run out the agent stops deciding and your website carries on unchanged.

Free is 1,000 agent credits with no card and no end date. Standard is USD 79 a month with 1,000,000 credits every 30 days, with no rollover. Enterprise is a custom plan agreed with sales. Operator seats are unlimited on every plan and are never counted for billing.

Only if you set it to. Autonomy runs from level 0, where it only observes, to level 5, where it acts inside your guardrails on its own. At level 3 and below every change waits for a person to approve it. The kill switch stops every live change on the next request, and nothing is deleted when it does.

No. Country, region and city come from the CDN's own headers, and the IP is never written to the database. Visitor identity is never inferred from an IP either: a visitor counts as identified only when your own site signs an assertion with the key for that site.

A control group that is held out of every experiment and sees the page unchanged. The console shows the raw rate after the change and the control relative lift side by side, labelled, so the first can never be read as the second. Below the minimum sample per arm it refuses to call a winner and says so.

In the request path between your website and the visitor. The tag streams events to the IntellQ runtime, and the decision for that page comes back on the same response inside a 2.5 second budget. Nothing proxies your HTML and nothing sits in front of your origin: the agent only changes slots you marked, in the browser, after your page has rendered.

Not to decide what your visitors see. Strategy choice is a contextual UCB bandit over your own traffic, warm started from your industry's pooled counts, so every decision can be traced to the numbers behind it. No language model writes page content.

Yes, under one Chat tab: an inbox for your team, a widget you size per device, proactive triggers, and an AI agent that answers from your own knowledge on your own model key, calls your tools over MCP, and hands over to a person for orders and payments. The console chimes the moment a visitor starts a chat.

Yes. Import your catalog from a Shopify or WooCommerce storefront in one click, from a CSV, or add products by hand. The AI agent recommends from it and shows up to three cards with picture, price and link, your team can send any product from the inbox, and visitors can browse it from the chat. The agent may quote only a price it read in the catalog.

Chat tools such as SalesIQ, Intercom and Tidio answer the visitor; optimisation tools such as VWO and Optimizely test the page. IntellQ does both in one tag: chat with an AI sales agent and product cards, and conversion agents that change marked page elements per visitor and keep a change only if it beats control. The comparison page sets 22 tools side by side from their own sites.

An A/B testing tool waits for you to have the idea, build the variant and read the result. IntellQ finds the pattern, writes the variant, runs the experiment, reads the result and rolls back its own losers. You approve or you do not.