Login AI
AI sales agent: qualification and follow-up
We connect an AI agent to your CRM: it reads lead context, qualifies against your script, sends follow-up, updates fields, and hands the dialog to a manager with the full history.
The agent qualifies inbound leads in amoCRM and Битрикс24, runs follow-up, and updates cards before the handoff to a manager.
Pricing
from ₽900,000
per project
This is the lower bound of a project. We scope the exact quote after a brief: it depends on volume and the number of integrations.
Capabilities
Lead context from the CRM
The agent pulls conversation history, deal stage, amount, and past tasks before the first question, so the customer gets straight to the point.
Qualification by script
The agent asks the questions from your script, records budget, timeline, role, and interest in CRM fields, and marks the lead as a fit or a cold contact.
Follow-up sequence
The agent sends the first touch, a reminder on day two, and a final message on day five, and a customer reply closes the sequence.
Card updates and handoff
The agent writes the dialog outcome, creates a task, and passes the manager a transcript with the reason, collected fields, and next step.
From request to result
1
Pipeline audit
We review stages, fields, sources, and fill rules: what the agent reads, what it writes, and where the manager's responsibility starts.
2
Qualification script and data model
We assemble the script questions, map answers onto amoCRM fields, and fix the dedup rule: one lead per customer, one log per dialog.
3
Integration and pilot
We connect the agent to the API, set the technical account's rights, and run a pilot on one request queue with a daily review.
4
Launch
We extend the scope to every channel, enable follow-up sequences and the dashboard, train the team on agent rules, and hand over support.
Check if it fits you
Good fit
Inbound leads and a CRM with an API
A thousand requests per month and amoCRM with API access give the agent enough flow for qualification and follow-up from day one.
A sales process owner
The person who owns the script and the fields sets the qualification rules and decides on disputed leads.
Not a fit
Sales tracked in spreadsheets
When leads live in Excel with no API and no single data model, the project starts with a CRM migration, and the agent connects after the move.
Automatic discounts and money decisions
Discounts, refunds, and write-offs stay with a manager: the agent gathers context, prepares the estimate, and creates a task.
What you get
1 000
Leads in the base scope
The base delivery covers 1,000 inbound leads per month: the agent qualifies the flow and sends follow-ups during business hours.
2
Channels in the base scope
The base scope includes two channels, for example the website chat and Telegram. More channels come as blocks with their own mapping and tests.
3
Agent scenarios
The base delivery includes three scenarios: lead qualification, a follow-up sequence, and a manager handoff. Each scenario runs on your test pipeline.
What the delivery includes
Pipeline map
After the audit you get the pipeline schema, the list of fields to read and write, the dedup rule, and the handoff points to a manager.
Pilot agent
The agent runs on your test pipeline: it qualifies leads, sends follow-ups, and updates cards. The pilot runs for 4 to 6 weeks with a daily review.
Metrics dashboard
The dashboard shows the share of qualified leads, the number of tasks from the agent, card edits, and the time to first reply. Metrics come from amoCRM.
How the system is built
Lead queue
A new lead arrives through an amoCRM webhook, the agent matches it to an existing deal by phone and email, and queues the task in Redis. A repeated webhook updates the same card, so the CRM keeps one deal.
Script and score
The script is split into steps with required fields: after each answer the agent writes the value to the deal and scores it by weight. A lead above the threshold goes to a manager, while the rest stay in nurture.
Follow-up queue and handoff
Follow-up lives in a queue with two- and five-day delays, and a customer reply closes the task and updates the card. On handoff the agent packs the transcript, fields, and next step into one manager task.
Where the solution pushes back
- 01
CRM data quality
Data quality sets how precise qualification gets: uniform fields and complete sources give the agent a basis for exact rules, and we review the cards during the audit.
- 02
Channel and API limits
Every channel has its own message limit and webhook rate: we schedule follow-ups against the API quotas, so a sequence goes out on time and within the provider contract.
- 03
Money decisions stay with a person
A person confirms discounts and refunds: the agent gathers history, calculates discount options, and passes the estimate with the customer's arguments.
Technologies
The technologies we use in our work and their role in a project.
CRM, events, and data
The agent reads and writes in amoCRM through the API, stage changes travel through Kafka, the deal snapshot lives in PostgreSQL, the dialog queue and sessions stay in Redis, and the scenarios and model run on Python.
Cost calculator
This is an indicative range, not a quote. We scope the exact estimate after a brief: it depends on volume and the number of integrations.
Leads per month
Channels
Scenarios
Add-ons
Support
Year one: updates and monitoring, a reply within 2 business days.
Service
AI sales agent: qualification and follow-up
The estimate covers our team's work. Hardware, third-party service licences and travel are added on request.
Estimated cost
₽1.25M
Timeline
6 wk
Frequently asked questions
Which channels does the agent cover?
The agent replies in the website chat, Telegram, WhatsApp, and email, and handles calls through telephony with a transcript. The base delivery includes two channels, and we add the rest as blocks.
Which CRMs do you support?
amoCRM, Битрикс24, and Salesforce. During the audit we check the API method list and show which fields the agent reads and writes in your version.
How long does the pilot take?
The pilot on one request queue takes 4 to 6 weeks: two weeks for integration, two for data collection, and two for review and scenario tuning. The full launch follows.
What does a manager get on handoff?
The dialog transcript, the collected fields, the reason for the handoff, and the proposed next step. The manager continues from where the agent stopped.
What teams add to this service
Related Services
Part of Solutions
Implementation Examples
−70%
Request handling time
Ready to discuss your task?
Leave a request – we'll get back with a proposal and a preliminary estimate.
Contact us