A scene, recurring in the companies we observe. The owner receives a monthly report, generated by someone (sometimes a controller, sometimes a prepared receptionist, sometimes an external accountant). The report has about ten pages, tables, charts. The owner reads it carefully the first time, quickly the others, and at a certain point they would like to know one specific thing that is not in the report. "How many customers did we lose in March?". "What was the cost per acquired customer on the Meta channel in the last quarter?". "How many quotes went out without closure in the last six weeks?".
The answer to these questions exists in the data. But it is not in that report. To have it, the owner must send a message to whoever prepares the report, wait for days, perhaps receive an approximate version, and meanwhile the decision they had to take starting from that answer has already passed.
This is not a problem of lack of data. It is a problem of interface. And the interesting thing is that, in the last two years, it has become solvable in a way that really changes the owner's working conditions.
Three interfaces to data, three profiles of who uses them
To understand where the conversational console sits in this family, it is worth seeing the three interfaces to company data that today coexist.
The first is traditional business intelligence (Power BI, Tableau, Qlik). Powerful tool, complex, reserved for those who know how to use it. It allows building customised reports, querying data in a structured way, discovering hidden patterns. For those who know how to use it, it is irreplaceable. For those who do not (and most owners of mid-sized companies do not know how to use it), it is an inaccessible interface.
The second is the dashboard. Lighter tool, shows a predefined set of numbers relevant to the business: sales, costs, conversions, KPIs (Key Performance Indicators). It works well when the owner's questions fall within the predefined cuts. It works badly as soon as the question goes out of the foreseen cuts, because modifying the dashboard requires going back to the underlying BI tool and therefore to someone who knows how to use it.
The third is the conversational console. Natural language interface that allows writing a question in your language and receiving the answer on your real data. It is not a substitute for BI or the dashboard: it is a level above, accessible to those who do not know how to use the underlying technical tools. It works because an AI system translates the question in your language into the correct technical queries on your data, runs the queries, returns the answer in your language with the necessary context.
The three interfaces coexist. BI remains the tool for those who build dashboards and data pipelines. The dashboard remains the tool for those who want to see at a glance the recurring numbers. The conversational console serves those who want to answer new, ad-hoc questions, without learning a tool.
What changes in practice for the owner
For those who are not familiar with the concept of conversational console, it is worth making concrete what changes in everyday life.
On Monday morning, the owner of a car dealership writes on the console: "how many quotes went out last week, divided by salesperson, and how many generated an appointment?". They receive the answer in twenty seconds. Seven quotes from Marco, of which four generated an appointment. Five from Anna, of which four generated an appointment. Eight from Stefano, of which two generated an appointment. The owner immediately sees that Stefano is producing many quotes but closes few, and has concrete material for the Monday afternoon conversation.
On Thursday evening, the owner of a trustee firm writes: "what was the average reply time to client emails this week, and which clients waited more than 48 hours?". They receive the list. Three clients waited over 48 hours. For two of them it was a non-urgent request. For the third it was a request on a tax deadline that should have been handled first.
On Tuesday afternoon, the commercial director of a yacht dealer writes: "how many contacts from the Genoa boat show last month turned into open negotiations, and what is the average value?". They receive the numbers and compare them with the boat show of three years ago, also available.
In all three cases, the information existed in the company's systems. What was changing was the interface to get to it. Without a console, every question of this kind required a person to translate it, hours of waiting, and therefore was often not even asked.
When it makes sense, when it doesn't
The conversational console is not a universal tool. There are conditions in which it makes sense and conditions in which it does not bring value.
It makes sense when: the company already has a structured data system (CRM, management software, integrated sources) that produces reliable information; the owner has the habit of wanting to check the numbers but does not have time to learn technical tools; the decisions the owner takes benefit from quick access to specific data (commercial sector, financial, healthcare, legal).
It does not make sense when: the company does not yet have a clean data infrastructure (in that case first you build that, then you put the console on top); the owner does not really want to read the data, but would like someone to interpret them for them (in that case a controller is needed, not a console); the owner's decisions are not driven by data but by relationships or intuitions (in that case the console becomes a toy, not a working tool).
An important technical note. The conversational console that really works is the one built on your specific data, with the Identity of your company incorporated. It is not ChatGPT that you load with some files. An effective console knows the structure of your CRM, your pricing rules, your seasonality patterns, the internal vocabulary of your company. Without this level of structured knowledge underneath, the console replies in a plausible but not useful way.
The conversational console is one of the most visible interventions of the passage from a company with data to a company with access to data. It changes how the owner takes decisions. It also changes the quality of internal meetings, because whoever arrives prepared with specific numbers shifts the conversation onto concrete decisions instead of impressions.
If you recognise yourselves in the pattern of the owner who would like to know but does not have time to ask, this is the kind of intervention we do after understanding what you already have in your data stack. The diagnosis is free.