Destreza Bets AI Marketing Needs More Decision Discipline, Not More Output

Destreza, a new London marketing consultancy founded by former BAT and Diageo marketer Andy Parton, is entering a crowded AI advisory market with a deliberately narrow proposition: apply artificial intelligence to the decisions that shape a brand before it is used to produce another campaign asset.

The consultancy was launched on 21 September 2026. Its focus is consumer businesses and their marketing leaders, with services covering brand strategy, AI operating models, capability and agency configuration. The underlying argument is credible: access to generative tools may be widespread, but the quality of the inputs, decisions and evaluation surrounding them remains highly variable. (workai.tv)

Parton’s proposed answer is a small, senior-led model. Destreza says it combines specialist AI agents with an “Agentic Council” intended to review work, while Parton retains responsibility for diagnosis, strategic choices and final recommendations. That is a business-model claim rather than independently tested evidence, but it identifies a more useful commercial question than the usual debate about whether AI can make marketing faster: who is accountable when automated research, recommendations or creative routes shape a material brand decision?

A consultancy built around the work before the brief

Much of the current AI marketing conversation remains centred on execution: generating copy, images, presentations, research summaries and audience variants. These applications can reduce time spent on routine production, but they can also make it easier to circulate plausible work before a team has resolved the customer problem, the brand position or the evidence behind a choice.

Destreza is positioning itself further upstream. Its initial offers are described as a fixed-fee diagnostic, a defined change programme and an ongoing senior partnership. The firm also plans a ten-stage “Playbook for AI Brand Building” for founders and lean teams.

That positioning matters because brand-building has never been only a content-production task. Marketing teams need to decide which customer needs are worth serving, what signals are reliable, how a proposition should differ, where local adaptation is justified, and when an agency or internal team should take the lead. AI can widen the set of options considered. It cannot, by itself, settle a trade-off between margin, distinctiveness, customer trust and long-term brand equity.

The practical risk is that an apparently well-informed output gains authority simply because it is comprehensive, polished and immediate. For senior marketers, the control point should therefore be the decision trail: the sources used, assumptions made, alternatives rejected, human reviewers involved and the measures that will prove whether the choice worked.

A useful precedent in consumer personalisation

Parton’s connection with Diageo’s *What’s Your Whisky?* gives the launch a relevant consumer-insight precedent. The digital experience, launched in 2019, asked consumers flavour-preference questions and used AI and machine learning to recommend a suitable single malt. Contemporary reporting identified Parton, then a Diageo senior regional manager, as part of the launch. (thespiritsbusiness.com)

The initiative was more substantial than a simple novelty quiz. Diageo later described the platform as a means of matching flavour preferences to product recommendations, and said that more than one million consumers had completed the experience by 2024. The company also extended the underlying FlavorPrint approach into cocktail discovery. (diageo.com)

That history illustrates where AI can have a clearer customer consequence. A recommendation engine can reduce uncertainty in a complex category, help a consumer navigate unfamiliar choices and generate structured first-party preference data. Yet even here, the test is not whether the technology feels personalised. It is whether the recommendations are useful, intelligible and commercially responsible — especially when the range presented is inevitably shaped by the brand or retailer deploying it.

For marketers, the lesson is that personalisation should be assessed as a customer-experience and evidence system, not merely as an interface feature. Which data improves the recommendation? What is the consumer being asked to disclose? Can the business explain the basis of the suggestion? And does the experience create confidence without steering customers in ways they would not reasonably expect?

The accountability gap is the commercial opportunity

Destreza’s proposition rests on a legitimate gap in the market: many organisations have acquired tools faster than they have built operating discipline around them. Leaders are being asked to decide which work to automate, what to keep with agencies, where proprietary data belongs, and what review is needed before AI-supported work reaches customers or boards.

A senior practitioner working with specialist tools may suit a defined strategic problem, particularly for a founder or lean marketing team that does not need a large agency structure. But the model will be judged less by the number of agents involved than by its ability to make advice auditable and outcomes measurable.

That means showing clients how research has been validated, separating evidence from interpretation, recording human sign-off, and setting success measures before implementation. It also means being candid about where an AI system is unsuitable: for example, where data quality is weak, the decision has regulatory or reputational consequences, or customer understanding has not been properly tested.

The launch is therefore less interesting as another announcement about AI-enabled marketing than as a signal of where advisory demand may be moving. The value is unlikely to lie in producing more material at lower cost. It lies in helping businesses make fewer weak decisions at speed — and being able to explain why.

---

This article is based on information distributed through Pressat. It has been edited by Expert View Media for clarity, context and length.