Finding a better deal takes time. You need to compare prices, read conditions, understand whether the initial savings hide a later expense. An AI assistant could do some of this work and, if authorized, go as far as paying. A question, however, comes before convenience: will we have more money left in the end?
It is a perspective opened up by agentic AI, composed of systems capable of planning and executing actions on behalf of users. In his speech at the European Consumer Protection Day, the deputy director general of the Bank of Italy, Sergio Nicoletti Altimari, highlighted its potential: reducing search and transaction costs, helping to compare products and interact with intermediaries.
The benefit, however, must be measured on the economic result. A simpler operation is not necessarily a more convenient choice.
From digital payment to delegated decision making
So far we have learned above all to pay faster: a saved card, an app, a confirmation on the phone. The assistant capable of acting introduces a further step: he can participate in the choice of what to buy and under what conditions.
In the evolution of digital payments in Italy, this changes the relationship between technology and money. Not all applications already offer these functions: possibilities and permissions depend on the specific service.
The potential value is understandable. A family may spend less time making repetitive comparisons and identifying conditions they would otherwise have overlooked. For the advantage to be real, however, the assistant must know what we mean by “convenient”.
Lower initial price, lower annual cost and greater contractual flexibility are different objectives.
How much do we save after paying for the assistant too?
The first account to make is that of net savings.
Let’s imagine, with purely illustrative numbers, that a service identifies alternatives capable of reducing some expenses by 15 euros per month. The potential benefit would be 180 euros per year.
If using the assistant cost 8 euros per month, the annual expense would be 96 euros: the benefit would drop to 84 euros, before any other charges. A one-off cost of 30 euros to change service would reduce this to 54 euros in the first year.
These are not rates found on the market, but an example of the necessary comparison. It is also important to verify that the services remain equivalent: spending less by giving up coverage or a useful service does not automatically represent an improvement.
The time saved retains value, even when it doesn’t translate into money. It is useful, however, to distinguish it from the actual reduction in expenditure.
The spending limit must look beyond the first charge
An assistant can respect the budget and still take on a more onerous commitment than expected.
Suppose we authorize a purchase within 100 euros. The chosen offer costs an initial 79 euros, but involves another eleven mandatory payments of 12 euros: the total cost becomes 211 euros.
This example helps to understand a point highlighted by Altimari: even a transaction carried out with valid credentials and within the delegated limits may not correspond to the consumer’s intention.
The delegation should, therefore, clarify what expenditure we are limiting.
| What to compare | Because the result changes |
|---|---|
| Overall cost of the contract | Includes commitments after the first payment |
| Assistant costs and fees | They can absorb part of the savings |
| Charge deadlines | They determine how much liquidity remains available |
| Performance and conditions | They allow you to compare equivalent offers |
Three cheap purchases can leave you with little liquidity
Then there is a difference between getting a good price and being able to afford an expense at that time.
Let’s consider another example: 600 euros are available in the account, of which 450 will be used for commitments already planned. Three purchases of 60 euros each respect a limit of 100 euros, but absorb a total of 180 euros. 420 euros remain: less than necessary.
The per-operation cap does not replace an overall budget. A system that optimizes each individual purchase without considering other commitments can produce convenient decisions made separately, but difficult to support together.
It is an important distinction even for users accustomed to apps: as shown by the in-depth analysis of the different payment profiles of Italians, digital familiarity and money management are not the same thing.
More comparison can increase competition
On an economic level, making the comparison less tiring could reduce an advantage that suppliers benefit from: the inertia of the customer, who maintains a service because looking for another takes time.
It is a possible consequence, not a proven result. It also depends on the breadth of the comparison: which offers does the assistant consider? With what criteria does he order them? Are there commercial relationships that influence the selection?
To evaluate convenience, knowing how a proposal is created is as important as knowing its price. An assistant who compares only a few offers does not necessarily represent the entire market.
Control remains part of the performance of the service
Altimari calls for understandable limits, traceable actions, the possibility of contestation and human supervision. Responsibility towards the consumer must remain attributed to people and institutions: delegating a task does not mean leaving the customer without interlocutors.
The speech does not introduce an automatic refund for every error. Indicates the conditions necessary for reliable delegation.
The final criterion remains concrete: how much we save, what commitments we make and how much freedom we retain to intervene. The most helpful financial assistant will be the one who can make these results understandable, even before asking for permission to spend.
