According to Bain & Company researchers, senior executives spend on meetings an average of two days a week→. According to other sources, calls take half of all working time: up to 23 hours a week→.
But this is an average. In the sales department, meetings and calls take even longer. Planners, meetings, standups, reports include measurements of how managers comply with call scripts and procedures, training employees, and plan strategies for behavior with important customers based on previous calls...
And here you will probably be surprised — what's the problem? This is exactly how the sales department works: it calls customers, then analyzes these calls and uses them to improve processes and increase conversions.
But let's imagine that it can take 2 hours a week to process calls instead of 15. For all these protocols, adding them to CRM, analyzing the use of scripts and analyzing individual employees. Moreover, these 2 hours will give the same (or even greater) efficiency and reliability of the data!
What your employees spend their time doing
Ordinary sales staff spend up to 35% of their time at meetings→.
Imagine your regular manager — let her name, for example, Maria. She talks to several clients every day. But her call job doesn't end the moment she hits the end. For about another hour a day, she writes protocols and call summaries to send to the client, save the data in CRM, and make a report for you. Sometimes Maria gets tired or has a headache — and she makes mistakes or forgets to enter important details into the protocol, for example, the customer's requirements for the steel grade.
You know that Maria is great at communicating with clients and selling well. It is most effective on a call. But these mistakes...
You can order Maria to devote more attention and time to the protocols, listening to calls twice. But you have a sales plan.
You can hire additional managers to ensure that the plan is implemented and the protocols are perfect. But this is not cost-effective.
You can make Maria do the protocols in person. But then the best members of your team will quickly find a new job — where the boss respects their time.
Is there a fourth way? Yes, implement an AI assistant to decrypt and log calls.
Imagine: instead of listening to the call for the sake of writing a protocol (or writing it from memory), Maria sees the call summary and can quickly, accurately and without forgetting the details.
And even better: the call resume goes to all participants automatically. And Maria gets +1 hour a day for additional calls to new customers. That's what you hired her for.
Interesting, but... not enough? However, an AI assistant for sales really doesn't have to be limited to simple protocols.
The perfect (AI) sales assistant
Let's rise from Maria's level again to the level of tasks of a sales manager. What routine tasks does he perform to raise sales figures and improve the funnel?
Surely this list includes these time-consuming but inalienable tasks:
- Analyze managers' calls for compliance with regulations and rules.
- Analyze the work of each manager to make decisions: develop, promote, fire?
- Identify and resolve recurring issues in communication with customers.
- To return and recharge customers who have not yet made a purchase or have not returned for a second order.
Let's see if an AI assistant can make these tasks easier and reduce the time you invest in them.
Do your managers ever forget to offer additional services?
The sales department has a lot of regulations and scripts. You use one when you first come into contact with a warm client, the other when you first come into contact with a cold client. The third is when you have been working for a long time and want to offer a new product or service. The fourth is if you see that the client is “stuck” in a funnel.
Lots of regulations. Multiple managers. How do you check that they are all using the right script in the right situation and offering what they should offer?
In the pre-AI era, things were simple (and long). I chose random calls, listened to them, checked them according to the checklist and gave feedback to the employee. But the department has only one manager, has a lot of tasks, and managers generate hundreds of hours of calls a week. You're not going to burst!
Now imagine that you don't have to listen to anything. Nothing at all! Just ask a question to the chatbot and get a detailed answer about the call's compliance with the recommendations.
.gif)
If you are interested in cross-sections of several dozen calls, you can look at a summary table of something like this:
... and see that:
- Manager Ivan's calls suspiciously often contain markers of dissatisfaction with the interlocutors;
- Marina sometimes forgets to offer a discount to regular customers;
- Asiya does not talk about current promotions, etc.
You can view reports by employee, by specific type of call (first contact, repeat contact), or for a specific period. For example, three weeks ago you were trained by a sales manager and now you want to see how they are using their skills. This is not a problem with AI analytics.
You'll say AI decryptors don't know how to do that. Yes, indeed, this is not a basic feature. But proper implementation allows you to add marker questions to standard transcripts based on your regulations. AI will analyze the transcript by keywords and phrases and show with high accuracy how each call complies with your internal regulations and scripts.
What's next: develop, praise, fire, educate—these decisions will continue to be on your side. But now you'll have more facts to accept.
How much time do you spend looking for deals with customers?
Of course, you already have a whole system for storing agreements. There are letters, there are call protocols (which are filled in by managers), and there are chat rooms. There is, after all, a CRM that aggregates and stores all this information. But...
First, are you 100% sure that your managers remember to specify anything in the minutes of meetings? People tend to forget some details and not “bring” them to the written record. I was distracted for a second while ringing the notification, and a whole layer of information was forgotten.
And secondly, even carefully stored information is not always easy to use. How many emails, protocols, and chats do you have to dig through to remember what you negotiated with your client A's manager Olga last May? Or was it June? Was this really an agreement, rather than a vague idea that “it would be nice someday...”?
Now imagine that any information regarding agreements with customers is, firstly, saved automatically without human error, and secondly, it is easily accessible.
You chat with “Remind me of such and such information over a certain period” and in 15 seconds you get a detailed summary: what you agreed on, why the agreement was not concluded at that time, and what you can do to resume cooperation.
As a result, you will be able to make more personalized offers to customers, quickly resume conversations, and identify recurring questions and requests.
According to a McKinsey study→, the use of AI in customer service can increase customer satisfaction while reducing maintenance costs.
How much time do you spend collecting data from the department to make management decisions?
Collecting data and presenting it conveniently in a single document is a big task. There are many options for solving this problem, from complex integrations to forced reporting by senior managers.
But imagine: a week ago you changed the script and sales didn't change — what's the reason? Or you've got a new product line but it's not for sale — what's the matter?
As soon as you have new data slices, you and your analysts need to figure out how to collect them, how to report them, and what indicators should be considered the norm. This is not a quick process. But decisions must be made today.
Now imagine: you can view statistics at any time in a regular table. Or ask the chatbot the question “How many calls did managers make this week offer customers such a product line?” Or, a couple of days after training, look at the statistics: who started using the new data and how.
The main thing is to find the right criteria and questions to assess the parameters you are interested in. And this, you see, is faster than making time for the IT team for a new integration or competing for business intelligence with the procurement department.
How to implement an AI assistant to get the most out
We've mentioned an AI assistant many times in this article, rather than, for example, an AI model like GPT4. This is no coincidence.
For AI to become a real assistant in the sales department, finding the right tool on the market is not enough (there is plenty of this stuff there!). It should be properly implemented into your processes and systems.
And this is where the KT.team implementation team will be useful to you.
Integration and detailed configuration
A common AI tool requires a lot of additional steps.
To get the transcript, you either need to let the AI bot into the meeting, or upload the recording of the meeting through a separate interface and then pick up the finished files.
The ideal work model for an AI assistant in the form in which we are implementing it looks like this: you or your managers simply press the “record call” button. That's it.
Ready-made transcript and protocol:
- are automatically saved to the history of your interaction with the customer in CRM;
- are automatically sent to all call participants (or not everyone, if this is your security policy);
- the call automatically receives a category: first interaction, closing a deal, etc. — it depends on the name of your funnel stages;
- in your workspace, you can track how well each manager complies with the regulations in each call;
- in a convenient interface for you, you can ask a question in human language about any of the previous calls and get an understandable answer.
To do this, we integrate an AI assistant with your CRM or any other systems that have the necessary data. We'll create a chatbot or connect an additional branch to an existing one. It will pick up records from the systems where you are used to calling and pick up customer data from the tables you want to use to improve the quality of responses.
At the same time, we will set up the language model so that it accurately deciphers your slang and terms. The minutes of the meetings will not include interjections, meaningless phrases and accidental reservations.
The ability to find answers to any question
You can use the AI assistant bot to find and analyze absolutely any information you talked about at the meeting.
Deployment flexibility
We can deploy the solution both in the cloud and on your local server. Our AI assistant can work with any calendar and conference call platforms, as well as CRM systems. We will design an AI assistant to be convenient and easy to maintain in the future — even in the development language or stack you need, with the features you need.
.png)
