Our product • Call analysis • Built for Hebrew
TalkWiz: every call in the contact centre, analysed and scored
TalkWiz is SoftWiz's own product for analysing sales and service calls. It listens to every call, not a sample, and returns an agent score, objections, buying signals, a follow-up message ready for the customer and a daily report for the manager.
A contact centre manager usually listens to a handful of calls a week and judges the whole team by them. TalkWiz listens to all of them. Every call is transcribed, analysed against the business’s own criteria and ends with a next step.
Why we built TalkWiz
We built TalkWiz from years of implementing information systems in sales and service contact centres, more than 500 businesses we have worked alongside, and we learned what every manager or business owner needs to spot in their sales process, and how to analyse the call properly, with years of knowledge and experience behind it.
That experience is the basis of how TalkWiz analyses a call. The criteria, weightings and compliance points can all be adjusted for each business.
What you get from each call
- An agent score from 0 to 100, against criteria and weightings you define, with the reasoning, a quote from the call and a practical tip
- Objections and pain points: what stops customers, how agents respond and what actually works
- A customer profile: type, size, budget and urgency, cross-referenced with the outcome
- Compliance and script: whether the agent said what they must, such as the recording notice and the right to cancel
- Sentiment and risk: frustration, escalation requests, churn risk and legal threats, before the customer leaves
- A follow-up message ready to send, and a daily email to the manager with an alert on any unusual call
Every finding comes with a quote from the call, so it can be checked in seconds.
How it connects to what you already have
If the contact centre runs on CallMarker, the connection is built in: each call reaches TalkWiz automatically when it ends, already transcribed. From any other phone system, calls are sent through an API or a webhook, or recordings and files are uploaded.
The insights can be passed on to wherever the team works, such as monday.com or another CRM. SoftWiz builds that connection as part of our automation and integration work.
Who it suits
It suits sales and telemarketing teams, service and support centres, insurance and financial agencies, property, solar energy and renovation firms, and any business whose team talks to customers on the phone, in Hebrew.
It suits less a business that handles a few calls a week and hears all of them itself. There is nothing to analyse that the manager does not already see.
Getting started
Sign up on TalkWiz, create an organisation and upload a first call. It takes a few minutes, and the first 100 calls are on us. Want us to connect your phone system and set the criteria with you? Book a meeting.
FAQ
What’s worth knowing before you start
What is the relationship between TalkWiz and SoftWiz?
TalkWiz is a SoftWiz product. We build and run it. Sign-up, the account and billing all happen on the TalkWiz site, and SoftWiz works with customers who want help connecting their systems and setting the scoring criteria.
What does it cost to start?
The first 100 calls are analysed free, with no credit card and no time limit. After that you choose a monthly plan that includes a number of tokens, where a token is one minute of call analysis. Current prices are on the TalkWiz site.
Is TalkWiz in English?
No. It was built for Hebrew from the start: Hebrew transcription that separates agent from customer, analysis that copes with slang and with transcripts containing errors, and an interface, reports and messages in Hebrew. It is aimed at Israeli teams.
Which phone systems does it work with?
There is a built-in connection to CallMarker, which sends every call automatically when it ends. From any other system, calls can be sent through an API or a webhook, including via n8n, Make or Zapier, or recordings and files can be uploaded by hand.
Are our calls used to train models?
No. The analysis runs through OpenAI's API, which under OpenAI's policy does not use data sent through the API for training. Sensitive details such as card numbers and ID numbers are masked before anything is sent.
Where is the data stored?
The database and servers are in the European Union, in Frankfurt. Transcripts are stored encrypted, and each organisation sees only its own data. Recordings are deleted as soon as they have been transcribed.
Leave your details for a free initial scoping session
So we can work out together how technology can give you an edge