Personyze
Website & app
Website PersonalizationContent, offers and layout shaped to each visitor — one profile across every page and every session.Recommendation EngineWhat each visitor wants next, ranked by the algorithm you pick or one you write.A/B TestingA winner per audience, promoted mid-campaign rather than after it.Dynamic Landing PagesThe whole page recomposed per visitor — headline, imagery and offer from the ad they clicked and their own interests. On pages you already run.AI Search & Chat AgentAnswers from your own pages and catalogue, and a search box that replies in a sentence.Banners & Pop-UpsYou decide who sees each banner or pop-up — by behavior, source, location or CRM field — and when it appears.AI Promotion ManagerYou list the offers; the AI picks the one each visitor sees and keeps learning from what converts.Social ProofLive scarcity, ratings and what other people just did.
Email, push & SMS
Email PersonalizationContent and recommendations picked per reader at open time — sent from Personyze or pasted as one block into your ESP.Push, SMS & WhatsAppFired by what they just did rather than by a schedule.Hosted Landing PagesThe same per-visitor page, served by us — your domain or ours, no site to deploy to.
Targeting & analytics
Behavioral TargetingSegments that update themselves from what people actually do.Audience DiscoveryPlain-language audiences found in your own data — visitors who convert several times more often than average, targetable in a click.ABM MarketingThe company behind an anonymous visit, joined to your CRM.Website AnalyticsSessions, bounce and revenue — then click a row and there is a person.
Platform & integrations
MCP ServerRun your account from Claude, ChatGPT or any MCP client.IntegrationsHubSpot, Salesforce, Twilio Segment, Tealium, GTM and the REST API.App MarketplaceThe integrations built in to every plan, and apps from other companies that Personyze reviews before listing.APIs, SDKs & AutomationOne REST endpoint over every object, server and mobile SDKs, rules and reports that run on their own.
Reviews, surveys & chat
Reviews & TestimonialsReviews from 17 platforms and your own site, matched to each visitor.Surveys & NPSOne question or a whole survey — every answer ready to target.AI Chat AgentAnswers visitors’ questions from your own pages, and hands over to a person.Meeting SchedulerThe right visitors book time with your team, on your own page.
PricingVideosProduct TourShort narrated videos of the real platform, one feature at a time.
← All articles
PersonalizationSeptember 28, 2026

Let Visitors Say What They Need: AI Product Finder and AI Content Finder

P
Personyze TeamPersonalization experts
Let Visitors Say What They Need: AI Product Finder and AI Content Finder

Site search is built for people who already know the name of the thing. Most visitors do not. They know the situation — “a lamp for my home office”, “how do I run an A/B test” — and a keyword box answers that with a page of loose matches and leaves the choosing to them.

AI Product Finder and AI Content Finder turn “search” into “tell me what you need”. The visitor describes the need in their own words and gets a short list, with a reason for each pick.

What the visitor sees

Gear for your next trip

Hiking in Norway this spring? Tell us the trip, and we will pick the kit.

a warm jacket for a week in the mountains, under €250
Boots for wet trailsA tent for two
Fjell Down Parka · €229

Warm to minus fifteen, and packs into its own hood.

Ridge Insulated Shell · €189

One layer instead of two on a wet day.

Tromsø Fleece · €95

The mid layer for evenings in the hut.

  • The bar sits at the foot of the page, at a side, or inside the page. It carries a title, an invitation, a question box and up to four example chips to tap.
  • The results page opens over the page: two to four sentences on what fits and why, a card per pick with a picture, the price, one line on why it fits and three reasons, a “Your picks” list, up to two buttons such as “Book a demo”, and up to three sections of your other campaign content.
  • Follow-up questions run as a short chat about the picks; a new question starts a new results page.

The Product Finder picks from your catalogue — price, stock, sale price — and can blend in the rankings your recommendation engine already keeps: best sellers, trending, most viewed. The Content Finder does the same for articles and site pages, with trending, most read and newest. Results stay for the whole visit, and the finder folds to a bar that brings them back.

A resource library page with the AI Content Finder bar at the foot: a title, an invitation, a question box with a typed question, and four example chips.
The bar at the foot of the page: a title, an invitation, the question box and up to four example chips to tap.

The summaries and reasons are written in the visitor’s language.

Watch it done · 1:53 — Setting up an AI product finder: shoppers describe a need, the assistant picks the products. See it with the rest of the tutorials →

Personal before they type

Every word the finder shows can change with the visitor, using the data you already hold:

The finder title written with ${company} and a chip written with ${industry}, as two visitors read them: a visitor from Coldwater Mechanical sees their company name and a “Case studies in HVAC” chip; a first-time visitor sees “Reading picked for you” and no industry chip.
One bar, two visitors. The visitor's company fills ${company}; a chip that needs an industry the visitor lacks is left out, with no gap.
ToolExample
A variable in any text“Reading picked for ${company}” shows the visitor’s company, with a default for everyone else
Variations per audienceOne invitation for returning customers, another for first visits
The bar’s words, written per visitorA question you write — “Invite somebody from ${company}, based in ${current_country}” — answered by the assistant as one short line
Chips that skip a missing value“Case studies in ${industry}” appears only for visitors whose industry is known, never with a gap

The designer previews it all as a real recent visitor, including the line the assistant writes for them.

Your brand, without a developer

Pick one of twelve looks in one click — Midnight, Daylight, Your colour, Ocean, Sunset, Forest, Lavender, Rose, Graphite, Editorial, Mono, Candy — then change any part: a colour for the page behind the results, the bar, the boxes and their text, the buttons, the question field and the accent. Edges, corners and fonts are settings, including the font of the site the finder is on. Text contrast is measured for every surface, so a custom colour never produces unreadable text. A logo sits at the left of the bar, and it can be a variable — a visiting company’s own logo.

The same finder bar and pick card in six looks: Midnight, Daylight, Ocean, Sunset, Forest and Editorial.
Six of the twelve looks, one click each. Any colour, corner or font can then be changed on its own.

Out of the way when they want it

The × either hides the finder or folds it into a small tab with the logo and title, which reopens it in one click. You choose when a closed finder comes back — next visit, after a day, a week or 30 days, or never on that browser — and whether it starts as the bar or as the tab. A finder placed inside the page is part of the page and has neither.

Watch it done · 1:56 — Setting up an AI content finder: readers ask a question, the assistant picks what to read. See it with the rest of the tutorials →

Inside any page you build

In the Canvas Builder, both finders are elements, beside AI Chat, AI Search and Existing action, which places any campaign content you already run — an offer, a recommendation widget, a lead form, a booking widget, reviews, a customer-logo strip — inside the design. Each piece still runs as itself, with its own A/B test and report, and a paused one leaves no gap. The “Personalized page” design in the Canvas gallery starts you with your own actions assembled into one page.

Private by design

A line many visitors share — the finder’s per-visitor invitation, and the opening answer of an AI Chat that greets with an answer — is written only from the question and your site. Anything personal comes in through the question’s own variables, which is also what gives each segment its own line. Nothing about one visitor’s profile or browsing reaches the next.

It also keeps the cost down: a shared line costs one AI answer per distinct question per hour, so a line for “a logistics company in the Netherlands” is written once, not once per visitor.

What it needs

  • The AI Chat add-on, or a trial, with conversations left. When they run out, finders switch themselves off like the chat, and the designer says so.
  • Something to pick from: a product catalogue feed for the Product Finder, articles or site pages for the Content Finder.
  • About five minutes: Campaigns → New → Targeting & Personalization, then Add action and search “Finder”.
See It In ActionBook a demo on your own catalogue — or start a free trial and add a finder in five minutes.

More: the finders on the AI Chat Agent page.

Questions people ask

How is a finder different from site search?

Search matches the words typed. A finder reads the need described, picks a few products or articles, and says why each one fits.

Does it need the AI Chat add-on?

Yes, or a trial, with conversations left. A finder is the AI assistant in another shape, and it switches itself off when the conversations run out.

Is one visitor’s data used in what another visitor reads?

No. A line shared across visitors is written only from the question and your site. Personal details reach it only through the variables you put in the question.

Can visitors close it?

Yes. The × hides it or folds it into a small tab, and you choose when it comes back, from the next visit to never.

Which languages does it answer in?

The visitor’s. Summaries, reasons and the per-visitor line are written in the language the visitor uses.

Let's talk

Book a demo with a personalization expert

30 minutes with a personalization expert. Bring your stack, your goals, your skepticism. We'll show you what changes when every visit feels like the only one.