AI Agents in Your CRM via Takli
AI agent orchestration between WhatsApp, the voice channel, and your CRM. Every dialog is written to CRM with lead-source tagging; the attribution loop closes with Google Ads. On the Takli platform.
Request demo → · takliapp.com →
What it is, and how it differs from a typical CRM integration
Most CRM integrations are data sync: contact left a form → landed in CRM → manager works it. Takli takes a step further: the AI agent independently works the contact in WhatsApp/voice up to the point a human is needed, and writes an already-qualified dialog to CRM with extracted fields (intent, budget, timeframe, urgency).
This is the difference between “the CRM knows there was a contact” and “the CRM knows who this customer is, what they want, and whether they’re ready to buy.”
Supported CRMs
Native connectors:
- Salesforce — Lead, Contact, Opportunity objects, custom fields
- HubSpot — Contacts, Deals, Conversations API
- Pipedrive — Persons, Deals, Activities
- Fireberry (Powerlink) — native Israeli CRM, very common in SMB
- Monday.com — boards-as-CRM (popular in Israeli tech startups)
Via webhook / REST API:
- Microsoft Dynamics 365
- Zoho CRM
- Bitrix24 (RU segment)
- any custom CRM with a REST/webhook endpoint
What Takli writes to CRM on each dialog
lead_source: "whatsapp_organic" | "whatsapp_googleads_gclid_XXX" | "voice_inbound" | ...
locale: "ru" | "en" | "he"
intent: "booking" | "pricing_inquiry" | "support" | "qualification"
extracted_fields:
service_type: "yacht_charter"
preferred_date: "2026-06-15"
group_size: 6
budget_range: "₪3000-5000"
urgency: "medium"
qualification_score: 0.78 # 0-1 likelihood to convert
transcript_url: "https://takliapp.com/dialogs/xyz123"
ai_summary: "Customer asking about a yacht charter on June 15 for 6 people..."
attribution:
gclid: "abc123"
utm_source: "google"
utm_campaign: "yacht_charter_en"This gives the sales manager a briefing-ready view before the first call: “the client asked about a charter on June 15 for 6 people, budget in range, next step is to discuss details.”
Attribution loop with Google Ads
This is critical for the honest CPA methodology:
- Customer clicks Google Ads →
gclidsaved in URL - Clicks WhatsApp on the landing page →
gclidcarried into the first message - AI agent qualifies → writes to CRM with
gclid - When CRM shows “closed-won” →
gclid+ revenue uploaded to Google Ads as offline conversion - Bidding strategy sees actual revenue via offline conversion upload to Google Ads
Without this loop — Google Ads sees “no conversions from WhatsApp” and lowers the bid. With the loop — it optimizes precisely against real revenue.
Pricing & onboarding
takliapp.com — pricing tiers, CRM connector demo. At Marketing Engineering Studio’s Scale retainer tier, this is included in the overall stack with no separate billing.
FAQ
What if I have a custom CRM?
We connect via webhook + REST API. Takli sends structured JSON to any endpoint you provide. We help with field schemas.
GDPR / PPL Amendment 13 compliance?
Yes. PII is processed with opt-in, opt-out is supported, dialogs are stored in EU/IL data centers per choice. PPL Amendment 13 (Israeli privacy regulation, in effect since 2025) — full compliance.
Can we train the AI on our historical CRM conversations?
Yes — fine-tuning the AI agent on your conversations. It’s a separate opt-in feature, discussed at the Scale tier.
What about two-way sync (CRM → AI)?
Yes. If a sales manager updates a deal status in CRM, the AI agent sees it before the next contact and adapts the dialog (e.g., doesn’t pitch already-rejected services).