AI employee · Advertising
AI Yandex.Direct Manager: pay for leads, not for Yandex's metrics
Our AI Yandex.Direct Manager runs your campaigns around the clock and counts results honestly: calls and form submissions, not the 'engaged sessions' and catalog opens that Yandex quietly logs as conversions of its own. In parallel, it filters out bot traffic and cleans up placements — on our own anti-fraud engine. The big decisions stay with you, and your data is processed on our infrastructure in Russia.
- We count leads, not Yandex auto-goals
- Anti-fraud against bots and click fraud
- Under human oversight · data in Russia
Why the dashboard shows conversions while the CRM stays silent
The budget leaks not only to bots — more often to whatever Yandex credits to itself.
| Характеристика | Значение |
|---|---|
| Yandex counts engagement as a conversion | Three page views, an 'engaged session', a catalog open — Yandex counts these events as goals by default. The auto-strategy dutifully optimizes toward these empty signals, the dashboard paints a high click-to-conversion rate, and yet the calls and leads never grow. |
| Bots and click fraud | In competitive niches, a sizeable share of clicks are bots and scripts burning your budget instead of your customers. You pay full price for every one of them, and not a single lead comes back. |
| Search terms, placements, and bids with no daily supervision | Without daily cleanup of negative keywords, Yandex Ad Network placements, and bid corrections, the budget bleeds into off-target impressions and overpaid positions. The contractor checks in once a week — but the budget is spent every day. |
Tasks
What our AI Yandex.Direct Manager does
We connect the AI Yandex.Direct Manager to your ad account and Metrica, and — on request, for precise lead counting — to your telephony and CRM. It counts neither clicks nor Yandex's soft auto-goals, but the calls and leads that actually happened.
- Separates real leads — calls and form submissions — from Yandex's soft auto-goals, reconciling clicks directly against telephony and CRM where needed
- Tracks and blocks bot clicks, adding their IPs and subnets to exclusions
- Cleans up search terms and negative keywords every day, and disables junk Yandex Ad Network placements
- Adjusts bids by actual payback — by leads, not by engagement
- Tests and rewrites ads, watching CTR and relevance
- Puts two numbers side by side in the report: 'conversions' per Yandex and real leads — so the gap is obvious at a glance
Human vs AI employee
| What counts as a 'conversion' | Yandex's soft auto-goals: page views, 'engaged session' | only calls and leads, for real |
|---|---|---|
| Bots and click fraud | spotted by hand, if spotted at all | the anti-fraud engine blocks them automatically |
| Watching bids and placements | in sporadic bursts, once a week | daily, around the clock |
| Where the data lives and who owns the changes | hit or miss with an outside contractor | our infrastructure in Russia, changes under human oversight |
How we launch it
| Характеристика | Значение |
|---|---|
| 1Connect the account and goals | We take access to Direct and Metrica and work out what the auto-strategy is really optimizing for right now |
| 2Audit and cleanup | We go through the campaigns and clear out the accumulated junk in search terms, negative keywords, and placements |
| 3Set up counting by real leads | We drop optimization toward soft auto-goals, reconcile clicks against telephony and CRM where we have access, and switch on anti-fraud |
| 4Reporting and adjustments | A regular summary with two figures — Yandex 'conversions' and real leads; major changes only with your sign-off |
Guided launch
The first months — with a dedicated manager
The AI is not launched “as is” and left unattended. For the first 1–3 months a dedicated manager runs your process: tunes the system to you, checks every result and fixes mistakes — while the AI learns on your real data in parallel. Nothing goes live without a human in control.
| Характеристика | Значение |
|---|---|
| 1–3 monthsDedicated manager | A dedicated specialist runs your process by hand together with the AI: tunes it to your documents and rules, checks every result, corrects mistakes. The system learns on your data. |
| Decision pointAutopilot or ongoing support | Once quality is stable, you decide: move the process to autopilot (cheaper, the AI works on its own, the manager on standby) or keep permanent support. |
| BeyondOngoing oversight, if you want it | For critical processes — primary documents, paid ads, contracts — many keep the manager for good: they backstop it, develop scenarios and own the result. |
This removes the main fear of automation — “what if it makes a mistake”. Until the AI has proven its accuracy on your tasks, a human is accountable for the result. We agree the format and cost of support around your process.
Where your data is stored
The AI Yandex.Direct Manager works with ad statistics and de-identified visit data. To reconcile real leads it can connect to your telephony and CRM, but it takes only the fact of a call or submission and its source — no call recordings or customer personal data enter the ad analysis itself. All processing runs on our infrastructure in Russia.
More about a local neural network on your own server →FAQ
How is this different from Yandex auto-strategies?
Yandex auto-strategies often optimize by default toward soft auto-goals — a few page views, an 'engaged session', a catalog open. Technically that's a conversion, so the dashboard can show a high click-to-conversion rate while the sales team hears nothing. Our AI Yandex.Direct Manager counts real leads — calls and form submissions, right down to reconciliation with telephony and CRM — and optimizes bids for exactly those.
How do you tell a real lead from a Yandex auto-goal?
We look at the fact, not the label in Metrica: a submitted form, a completed call, a move to checkout. Where we have access to telephony or CRM, we match the click directly to the call or the deal — which removes both auto-goals and fake activity at once. Without that access, we configure goals around hard events and manually exclude soft auto-goals from optimization.
How much does it actually save?
It comes from two sources: bot traffic (a practical benchmark for competitive niches is 15–30% of the budget) and the money the auto-strategy was burning to optimize toward 'conversions' that never happened. The exact figure for your account only becomes clear after an audit — we compare Yandex 'conversions' against real inquiries over the same period. We won't promise a percentage in advance.
Can it break campaigns that are already set up?
Major changes — new campaigns, serious structural edits — are agreed with you first. Only safe operations run automatically: negative keywords, excluding placements and bot IPs, and bid corrections within set limits. And behind it all is our specialist who oversees the AI's work.
What if I already have a contractor?
Then the AI Yandex.Direct Manager works as reinforcement: it takes on the daily routine, anti-fraud, and honest lead counting, while your contractor handles strategy and creative. It's usually cheaper than expanding an agency team, and it becomes immediately clear who is really moving which numbers.
Related reading
We cover this topic in more depth on the blog
See the gap between 'conversions' and real leads
We'll send you a breakdown of your account: how many 'conversions' Yandex credits to auto-goals and how many of those are genuine calls and leads. Plus how much of your budget goes to bots. The audit is free.
Contact
Let us try it on your data
Describe the task and send a couple of typical documents or examples — we will show how the system handles them, before any contract.
- info@stitex.com
- Telegram
- @StitexBot
- Response time
- Within 1–2 business hours
B2B & Commercial:
- All prices include 22% VAT
- Electronic tax invoices & formal agreement
- 3-month warranty on completed code