The proactive AI platform
that works your customer base
A-CDP collects scattered customer data into a single 360° profile, then adds a proactive AI agent that analyses your base, spots patterns and prepares segments, hypotheses and campaigns — while the final call always stays with your marketer.
What is A-CDP
At its foundation: a Customer Data Platform that merges everything your company knows about a customer — CRM, website, mobile app, loyalty programme, POS — into one clean profile, free of duplicates. On top: marketing automation, segmentation, omnichannel campaigns and a proactive AI assistant.
IDGraph technology stitches identities across all channels and systems into one Golden Record — with configurable merging logic and unlimited identifier types.
Build segments on any profile attribute or event history. AI assistant helps draft segments in plain language — no SQL needed.
Email, SMS, mobile push, web push, Telegram — mass broadcasts, trigger flows, cascading scenarios and A/B tests of full campaign branches, all from one interface.
Deploy inside your own infrastructure (Kubernetes, your S3-compatible storage) or in a managed cloud. Data never leaves your perimeter.
Three Product Pillars
Every product decision comes back to these three foundations.
Fully integrated with your CDP data, the AI agent suggests segments, generates hypotheses and drafts campaign content. External market and competitor monitoring — arriving end-2026. The final call always stays with the marketer: human-in-the-loop.
Three-perimeter model: raw personal data stays locked in the CDP; AI works only with aggregated insights and anonymised IDs; LLMs can run fully on-prem. Passes information-security audits at top-tier banks.
Tens of millions of profiles per day and thousands of events per second — the load level of the largest banks. Yet the interface is designed so a marketer, not an engineer, can build and launch a campaign in hours.
For Whom
We focus on companies where the marketing function operates at scale — millions of active customers, frequent touchpoints. Industries: retail, FMCG, banking & finance, insurance, e-commerce, automotive.
CRM, website, mobile app, loyalty programme and POS each hold a fragment. Nobody has a single, current view of the customer.
Every campaign waits in the analyst queue. By the time a segment is ready, the moment to act has already passed.
Mass broadcasts to the entire base burn reach and hurt deliverability, while relevance and open rates keep sliding.
Generic AI tools have no structured view of your base, so their suggestions are blind. Real proactivity needs a data foundation first.
Why A-CDP
The AI agent already works on your data today: it collects segments, proposes hypotheses and generates content drafts. External competitor monitoring and auto-initiated campaigns (with marketer approval) are our key direction — launching into production by end-2026. Having AI is no longer a differentiator; proactivity is.
We grew from a strong data foundation — not a bulk-sender that bolted on a data layer. The CDP core (unified profile, structure, access, audit log) is the product, not a feature. The result: the AI has something real to work with.
Suggestions are built on the structure of your specific base — but without access to raw personal data (three-perimeter model). Universal AI "without a CDP under the hood" works blind. Ours does not.
We deploy inside your infrastructure — on-prem by default for enterprise. Validated in production at top-tier banks. Integrates with your DWH and takes only what is needed for marketing; raw data never leaves your perimeter.
Tens of millions of profiles per day and thousands of events per second — the load of the largest banks. A Kubernetes-native, fault-tolerant architecture with ClickHouse, Kafka, Spark and Flink under the hood. Years of engineering maturity that are hard to copy.
Proactive AI Agent
The AI agent works in a feed-and-chat format. It itself analyses purchases, finds stable patterns, builds segments and suggests hypotheses — in plain language, no SQL. The marketer stays in control.
Strategy, complex scenarios and brand voice stay with the marketer. No communication goes to a client without a marketer's approval and full testing. AI removes the routine; the human makes the decisions.
Ask in natural language — the assistant translates to a segment builder query. No SQL, no waiting for an analyst.
"Adding the client's name to the subject line could raise conversion." Surfaces ideas from your own data patterns — things you would not otherwise have time to find.
The AI-powered message editor generates draft subject lines, body copy and push text as a starting point — the marketer edits and approves.
"A competitor just launched a −30% promo in your category." External signal → auto-drafted response campaign → marketer approves → send. This direction enters production by end-2026.
AI never touches raw personal data. Three-perimeter model: raw data locked in CDP; AI works with aggregated insights; LLMs can run fully on-prem.
Platform Modules
No separate assembly. Customer profile, AI agent, segmentation, campaign manager and analytics — by default.
Build segments from any profile attribute or event. One-click export to ad platforms (Yandex, VK) or downstream systems. Schedule auto-refresh or run on demand.
Mass broadcasts, trigger flows, cascading cross-channel journeys and A/B tests of full scenario branches. Email, SMS, mobile push, web push, Telegram, MAX.
Instrument any website or landing page with a lightweight JS snippet — no engineering sprints. Configure tags, triggers and variables in the UI.
Resolve fragmented identities — cookie, phone, email, CRM ID, loyalty ID, IDFA — into a single profile. Multi-channel deduplication. Unlimited identifier types.
Query any customer's segment membership and full profile in real time via low-latency API. Powers website personalisation, app recommendations and call-centre context.
Built-in dashboards for audience size, segments and campaigns. Deep integration with ClickHouse for end-to-end analytics across your data stack.
In-platform AI for segment search in plain language, campaign content drafts and audience recommendations. Built on your CDP data — not working blind.
Define your own data model — profile attributes, customer journey events, dictionary vocabularies — aligned to your business domain. Full schema control without code.
Role-based access per module, custom roles, SSO integration. Quarterly security scans, backup and disaster-recovery plan included.
Case Studies
Three production deployments across retail, media buying and FMCG — each implemented in 3 months, then handed to a dedicated 10xT support team.
Built a unified customer profile from site, app, offline, CRM and other sources via IDGraph. Added a predictive ML layer scoring purchase propensity by renovation stage. Segment preparation time: from 1 week (analyst queue) to 1 day (self-service by marketer).
On-prem CDP installation inside the company's perimeter. Automated key site-event collection, built interest-based segments for top products, configured auto-refresh schedules and exported segments to Yandex and VK ad cabinets. Validated with A/B testing.
Marketers build complex segments without IT in hours (not weeks). Personalised push notifications and banners based on long-term behavioural history. Analysts use the same platform to train churn-prediction and behavioural ML models.
Who Uses It
A-CDP serves the full cross-functional stack. Not just marketing.
How It Works
Turnkey implementation takes 3 months. The system is then handed over to a dedicated 10xT support team.
Technology Stack
Get Started
A typical turnkey deployment takes 3 months. We also help offset migration costs from your current vendor. Let's scope it together.