OLIVER proposes keeping the advantage of artificial intelligence inside the brand. Its AI Inside model combines people, training and technology, with customers retaining their data, intellectual property and models in their own cloud. The commercial promise is tested when that capability produces approved work the brand can use.
An image generated quickly can take days to reach a campaign. Selection, channel adaptation, product checks, permissions and approval sit between those moments. These tasks are part of production costs. If they sit outside the budget used to justify the transformation, the saving may describe only a fraction of the work.
For OLIVER, the growth opportunity is to make that entire sequence something a customer can buy and assess. A brand needs to know what it keeps, what it delegates and how it will establish whether an AI operation improves the previous process. The agency has a particular position from which to answer: its offer embeds teams within the customer’s business rather than simply supplying a tool.
Retaining data also requires being able to work with it
The public model description places governance inside the system from the outset. It distinguishes an initial efficiency phase from a second phase that reinvests the value created in experimentation and growth. That distinction helps sell and evaluate the service: freeing capacity and turning it into better decisions are different outcomes.
Hosting models in the customer’s cloud clarifies one part of control. Continuity also requires access to instructions, approved versions, review criteria and change histories. If that knowledge exists only in conversations with the supplier, technical ownership can coexist with operational dependence. This is a risk to resolve in the agreement, not an established OLIVER practice.
The first commercial deliverable should make it possible to follow an asset from request to use: who authorizes product information, which variations the brand allows and who accepts the result. Recording those decisions lets an expansion into another market reuse agreed criteria while costing what changes, such as language, channel or local restrictions.
Integration work matters before a tool creates value
Challenges reported by 201 CMOs or equivalent leaders at companies with at least £150 million in revenue. International online survey, March and April 2025.
Tap or hover for details. All values are also available in the data table.
Collapse notes
Multiple responses. Sample covers North America, the UK, Europe, the Middle East and Africa, and Asia-Pacific. These are reported perceptions, not failure rates, measured costs or OLIVER results. It does not represent Spain specifically.
Explore the data table
| Category | Respondents |
|---|---|
| Lack of time to evaluate | 60% |
| Integration complexity | 54% |
| Internal resistance to change | 49% |
| Budget constraints | 44% |
| Lack of knowledge | 39% |
| Poor data management | 23% |
In LXA and Intent HQ’s 2025 international panel, reported by Statista, 60% identified a lack of time to evaluate technologies and 54% identified integration complexity. The survey does not measure an agency’s performance. It does provide a reason for the commercial proposal to include the work a buyer must do before receiving value, rather than treating implementation as a formality.
A bounded project can make that exchange explicit. If the brand starts with adaptations of an existing campaign, scope should define formats, channels, included revisions and acceptance criteria. Pricing becomes clearer when requested assets are distinguished from accepted ones and the reasons for repeated work are recorded. Generating more variations is not required to demonstrate an operational improvement.
Cost per approved asset connects Procurement with the actual work
In its analysis of Procurement’s role , OLIVER argues that the function should participate before a supplier is selected and agree how outcomes will be measured. That position makes the economic discussion more concrete: the contracted unit can include review and usable delivery as well as generation capacity.
The comparison needs a consistent baseline. A master asset and a size adaptation are not equivalent. Nor are an already validated translation and a new commercial claim about a product. Separating those work categories prevents an apparent improvement from simply reflecting a shift toward easier tasks. Time to acceptance and total cost per accepted asset can be observed within each category.
Total cost would include preparation, licenses, human work, corrections and coordination; the denominator would be assets accepted under previously agreed criteria. This is a proposed buyer evaluation framework, not an OLIVER internal figure. It makes it possible to discuss whether technology reduces effort, merely moves work to another team or gives the brand a capability it did not previously have.
Creative work faces a commercial challenge: explaining its value to other teams
Five selected obstacles to determining the value of creative investment. Winterberry Group, US and UK, research conducted from March to May 2025.
Tap or hover for details. All values are also available in the data table.
Collapse notes
Respondents could select up to two answers. The original chart identifies 207 participants; Statista summarizes the base as 200. Distribution: 68% US and 32% UK. Exact fieldwork dates are not detailed. Industry companies funded the study, which included interviews with OLIVER and others; it does not audit the agency or demonstrate profitability. It is not combined with the other panel as if they shared a sample.
Explore the data table
| Category | Respondents |
|---|---|
| Explaining the connection to sales | 34% |
| Explaining value to Finance or the CFO | 34% |
| Aligning strategy, creative and media | 31% |
| Explaining value to Procurement | 29% |
| Aligning marketing and agency objectives | 26% |
Winterberry Group’s study helps explain why that definition matters. Explaining the connection to sales was selected by 34% among their two biggest obstacles to valuing creative investment; another 34% selected explaining value to Finance. Lack of alignment between strategy, creative and media was cited by 31%. These are responses from a 2025 panel, not an assessment of OLIVER customers or a measurement of Spain.
Approval settles delivery; the market determines effectiveness
A WFA survey published on September 30, 2026 places human control inside the operation: 75% of participating brands required human oversight of AI outputs. The base includes 54 senior respondents from 46 large brand owners. It describes that group rather than Spanish businesses as a whole, and does not validate OLIVER results; the agency’s parent group also has a partnership with the WFA.
Budgeting for control changes the incentive. If compensation rewards only the number of variations, producing more can be attractive even when it increases the customer’s review queue. An agreement centered on accepted assets and delivery times gives both parties a reason to resolve ambiguous instructions and brand limits earlier. Acceptance nevertheless establishes that delivery meets a requirement; it does not establish additional sales.
OLIVER itself argues for shifting the conversation from efficiency toward effectiveness . That second claim needs a different kind of evidence: a campaign comparison in which audience, spending, offer and period remain identifiable. Commercial outcomes cannot be attributed to creative work if discounts, distribution or advertising pressure also change without separating their effects.
Service expansion would therefore follow a reasoned order. First, demonstrate that a defined workflow delivers usable work with control and observable costs. Then establish whether that work improves a market outcome under an appropriate comparison. Only then is there a solid basis for widening the scope or reinvesting released capacity in further experimentation.
For OLIVER, the distinction between a repeatable transformation and an impressive demonstration lies in that step. The advantage stays inside the brand when its teams can explain what they received, what it cost to put it to use and what they learned from the outcome. That evidence makes the next investment, and a broader commercial relationship, easier to defend.
In this article
Marian Vasile
Marian Vasile is a MarTech and growth professional. His work spans product marketing, SEO, paid advertising and deep online tracking, including behavioral analytics and attribution.
