of conversations scored, instead of 3%
Automatic scoring of calls, chats and video meetings, dashboards by agent and by appeal, AI analysis of the reasons people say no.
A101 groupSpeech Analytics · 2026
Where the sales team loses buyers
The analytics sorts calls into a funnel: you see at which step buyers leave, and why.
The funnel of calls for a month
- Called1 000
Every inbound call of the month. The analytics works out the subject of the call: the property, the district, “are there instalments”.
From the call«I am interested in the new blocks in Yunusabad, how much per square metre?»
- Reached a sales manager760
Missed calls that were never returned the same day. The analytics compares the list of missed calls against the outbound ones.
From the call«I have called three times and nobody picks up.»
- Were given a price540
The manager dodged the question about price: “come in and we will discuss it”. Every such conversation is flagged.
From the call«Tell me the price and then I will come in.»
- Were offered instalments310
The customer said “too expensive” and the manager did not offer instalments or a partner bank’s mortgage.
From the call«It works out expensive, for that money another development gives you more space.»
- Booked a viewing190
The conversation ended with no next step: no viewing and no time for a follow-up call.
From the call«All right, I will think about it and call you back.»
- Deal48
Customers who were never called back within three days of a viewing.
From the call«We saw the flat, waited for a call, and nobody rang.»
Every manager against one checklist
The analytics checks conversations against your sales script. Not a sample once a week, but every call.
The script
Introduced themselves, asked for a name, checked the budget and the timescale, offered a viewing, agreed a next step.
Objections
“Too expensive”, “too far”, “we are still thinking”: did the manager work with it or agree and say goodbye.
Competitors
Which developments buyers name, what they compare you with and what attracts them there.
The bot learns the budget, the timescale and the district before the manager
In the evenings and at weekends the bot takes the calls. Three questions — and in the morning the manager calls prepared, instead of starting with “tell me what you are looking for”.
The lead record for the manager
- District
- Yunusabad
- Layout
- 2 rooms
- Budget
- up to $80,000
- Instalments
- Yes
- Call
- Tomorrow 10:00
Developer bot
Which district suits you: Yunusabad, Yakkasaray or Sergeli?
Yunusabad, a two-bedroom.
What is your budget? Are you interested in instalments?
Up to eighty thousand dollars, and yes, instalments.
When are you planning to move? A manager will call tomorrow at 10:00, does that suit you?
КлиентAre there three-bedrooms in Sergeli, and over how many months are the instalments?
The prompter’s suggestionSergeli City, block 4: 6 three-bedrooms available, 78–84 m², from $11,800 per m². Instalments up to 24 months, 30% deposit.
The prompter knows the remaining units, the prices and the instalment terms
While the customer is asking, the manager can already see the free flats in that block, the price per square metre and the instalment terms. No need for “let me check and call you back”.
Related case studies
faster to shortlist properties
The bot understands “a two-bedroom in the centre under 12 million”, searches the live database and answers on the website and in Telegram; the sales team handles the difficult customers.
A large property development companyAI Agent Factory · 2025Client under NDA
Still have questions?
How does the analytics know the manager never named a price?
The model recognises the customer’s question about price and looks for an amount or a range in the manager’s answer. If there is none, the call lands in the “no price given” layer with a timestamp, and the manager can listen to that moment.
Do we have to change our CRM?
No. The result of the call review and the lead record from the bot are passed into your CRM over API. We will show the list of ready integrations at the meeting.
What if the managers speak Uzbek?
The analytics and the bot work in Uzbek and Russian, including when the languages are mixed within one sentence. That is an ordinary conversation about a flat in Tashkent.
We will build the funnel from your own calls
Send us a week of recordings. We will return the funnel with the real losses and examples of conversations in every layer.
