Platform Updates: AI-Forward Experimentation, Heatmap Diagnostics, and Region-Aware UI

Most experimentation tools were designed for a workflow that no longer exists: open the app, click around, build a variant, wait, log back in to read the results. Healthcare marketing teams running lean don't have time for that loop, and the AI tools they're already using every day can't help because nothing is exposed.

Your team is already running half its work through Claude or ChatGPT. Ours Privacy now puts experiments and heatmap diagnostics in that same workflow. Describe the test you want and your AI assistant sets it up. Ask which pages are seeing rage-click spikes and you get a ranked list back with session replay links attached. No logging in to check the dashboard. No exporting reports. Just ask.

Here's what's new:

AI-Forward Experimentation

Why experimentation, and why now. Your CDP already captures every event your team can act on. The analytics suite already shows you which channels and audiences drive attribution. The missing piece has been the ability to actually move those numbers, not just measure them. Experimentation closes that loop. Tied to session replay, you don't just see whether a change worked, you see why.

This release brings two things together: on-page content tests, and AI assistant access through Claude, ChatGPT, or any AI tool you connect.

On-page content experiments

Experiments in Ours Privacy now run two ways. Redirects send half your visitors to one URL and the other half to another. On-page experiments let you change the page itself. Swap a headline. Change a CTA. Try a different hero image. Pick the element, define the variant, and Ours Privacy handles the swap on render. No new URL, no engineering ticket.

What it does:

• Element-level variants: Change text, images, classes, or attributes on any element without redirecting

• Sits next to redirects: Both experiment types live in the same flow, with the same metrics and the same reporting

• Works with Session Replay: Watch how visitors in each variant actually behaved, not just whether they converted

For a healthcare marketing team, this is the difference between testing "does landing page A or B convert better" once a quarter and testing "does this CTA copy beat that one" every week.

Run experiments through Claude or ChatGPT

Connect Claude, ChatGPT, or any AI assistant your team already uses, and your experiments become something you manage by chat.

What you can do with it today:

• Build experiments by prompting your AI: Describe the test in plain English and your assistant creates it for you

• Read live results: Ask "how is the new pricing page test trending?" and get current conversion numbers, lift, and confidence levels back in chat

• Daily check-ins without logging in: Set up a recurring prompt and let your AI assistant summarize experiment performance every morning

• Cross-reference with everything else: Pull experiment results next to audience reports, funnel changes, and channel attribution, all from one assistant

The bigger story: every team is working inside an AI assistant now. Ours Privacy is built so the platform shows up where the work happens. Experimentation is the first place we're shipping that pattern.

AI-Forward Heatmap Diagnostics

Heatmaps tell you where visitors are fighting your pages. Until now you had to load each one and eyeball the trends. We've exposed the whole heatmap surface to your AI assistant, so you can ask the right questions instead of clicking through reports page by page.

What you can do with it today:

• List, sort, and filter every page by friction signal: Ask "which pages have the highest rage click rate this week" or "show me dead click spikes since Monday" and get a ranked list back in chat

• Slice by anything that matters: URL pattern, audience, device, traffic source, date range. The same filters the dashboard supports, available as natural language.

• Pull session replays inline: Once you spot a spike, ask for replays of engaged visitors who actually triggered those rage or dead clicks. Watch what is happening, not what you assume is happening.

• Close the loop in one conversation: Spot the friction, watch a replay, prompt your assistant to set up an on-page experiment that tests a fix. Same chat, end to end.

This is where the AI assistant earns its place in your workflow. Heatmap analysis used to be the slow, manual step between "something feels off" and "let's run an experiment." Now it is one prompt.

Funnels Now Live in Top-Level Navigation

Funnels are now a top-level item in the sidebar, next to Channels and Audiences. If you're checking funnel health every day, the saved click adds up. It also reflects how teams are actually using the platform: funnels are no longer a "drill down" view, they're a daily dashboard.

Region-Aware UI Across Jurisdictions

What you say to a patient in California isn't what you say to a patient in the EU. Different jurisdictions, different legal requirements, different obligations to disclose. Until now, that meant a separate lookup somewhere on your page or shipping different versions of the site to different regions.

Your consent platform can now tell your website what region each visitor is in, so your web team can adjust the page directly. Swap your disclosures. Change your patient-facing language. Adjust your consent flow. Meet users and patients where they are, in the language and tone the regulators in their region expect.

Behind the scenes, a new getGeo() method on the CMP SDK exposes the visitor's country, region, and any GPC signal status. Drop it into your rendering logic and let your site speak the right thing to every visitor.

HTTP API Batch Ingest

The REST ingestion API now supports batch event submission. Send a full batch of events in a single request instead of one call per event. For server-side integrations with high event volume, that means lower latency, fewer connection costs, and simpler retry logic. The batch endpoint follows the same authentication and identity resolution rules as single-event ingest, so existing pipelines drop right in.

Your Data, Your AI

What's coming together here is an end-to-end loop. See your attribution. Send it downstream securely. Ask your AI which pages are seeing rage clicks. Watch the replays of visitors who hit them. Run an experiment to fix what you saw. Watch the actual sessions that flowed through each variant. Ask your AI assistant what to do next.

The whole loop runs on data you own and an assistant you control. The AI tying it together is your team's, your enterprise's. If you don't want experiment data accessed through Claude or ChatGPT, shut off API access at the admin level. If you want to allow reads but block writes, do that. The controls live in your account.

Most platforms ask you to trust them. We're inverting that. Your data stays yours. Your AI stays yours. The leverage moves to you.

What's Next

The same AI-forward pattern is rolling out to audiences, funnels, and channels next, so any analysis a person can do in the app, an AI assistant can do for you.

If you'd like a walkthrough or want help wiring AI assistants into your team's daily workflow, book a demo.