ParallaxAdvisory
Back to selected work
Case 01 — Storytelling

Rebuilding global content strategy with agentic AI.

Three category-leading global brands with disjointed campaigning rather than coherent storytelling limiting growth potential.

Vuse, Velo and glo were active across the marketing mix, but largely through disjointed, isolated campaigns — with little of the consistent, brand-level storytelling that builds equity and drives growth. The brief: rebuild all three brands' 360° content strategies in eight weeks, ahead of a global planning milestone — and in a critical commercial year, with BAT's New Categories targeting mid-teens revenue growth,2 the bar was to improve things materially, fast. The question that framed the work — could you rebuild three brands' content strategies that quickly, to a standard that actually holds, by making AI the engine of the method itself rather than a tool bolted to its edges?

RoleGlobal Head of Content
ClientBAT · New Categories
ScopeGlobal
Year2026
ProgrammeDEPTH · FOCUS
ProofPublic method · detail on request
01The problem

Plenty of campaigns, not enough storytelling — so too little was landing.

Three category-leading brands — Vuse, Velo and glo — each activating polished, on-brand content, but as distinct campaigns, with a lot of variation across touchpoints and little coherent storytelling at a brand level. The catch is a general one: when a brand tries to say too many different things at once, no single message reaches the scale or frequency it needs to land — so consumers don't join the dots, and the consistent storytelling that builds brand equity never takes hold. The brief: rebuild all three brands' 360° content strategies in eight weeks, ahead of a global planning milestone.

02The idea

One connected story per brand, engineered as a flywheel.

The idea: give each brand its own connected story — one narrative driving every channel and market for that brand, designed as a flywheel where each turn compounds the last instead of starting from scratch. Not three brands forced under a single message, and not a brand-out process bolted to some AI tools, but a tailored storytelling system for each brand, built AI-native from the workflow up.

03The AI role

A bespoke agentic AI workflow — with a creative AI engine running alongside.

The strategy engine ran four phases: analyse the evidence base, overlay each brand's truth, shortlist the opportunity spaces, then validate against consumer and claims-risk screening — surfacing options at a depth and pace no human-only team could match. Running in parallel, a creative AI engine turned those opportunity spaces into mood boards, concept references and storyboards, bringing the thinking to life inside the business. None of it was production work: that stayed with the brands' production hub and agency village, and every senior judgement call stayed firmly human.

04The approach

Prove it on the hardest brand, then cascade.

Vuse went first by design — the world's number-one vape brand by share, but with the widest gap between that position and the strength of its storytelling, the biggest commercial scale, the deepest existing brand-architecture commitments to work with, and the most divergent content across markets. We built and ran the full four-phase workflow on Vuse end to end, producing its complete strategy first; what cleared its hardest gates became the proven template, not the finished article. glo and Velo then followed as deployment rather than parallel discovery, reusing that workflow — eight weeks, gated at every phase.

05The outcome

Three omnichannel content strategies in eight weeks — now the global standard.

All three omnichannel content strategies were delivered ahead of the global planning milestone, and the workflow became the architectural standard the brands now build from. The output can already be seen across the three brands' marketing mix, with the standard applied globally — across a portfolio reporting c.£3.6bn in annual revenue1, guided to mid-teens revenue growth in 2026.2

Sources & substantiation
  1. BAT's three New Categories brands (Vuse, Velo, glo) deliver c.£3.6bn in combined annual revenue. BAT investor materials / Annual Report. bat.com
  2. BAT's New Categories (Vuse, Velo and glo) revenue-growth guidance upgraded to mid-teens for H1 and full-year 2026, from low double-digit. BAT 2026 First Half Pre-Close Trading Update, 2 June 2026. bat.com

Diageo gold-standard benchmarks (Don Julio, Guinness, Marketing Catalyst) and the full programme detail are set out inside the requested case study. Request the full case study →

Case study describes work led by Andy Parton in senior roles at British American Tobacco plc and Diageo plc. All brands and trademarks are the property of their respective owners; Parallax Advisory is independent and not affiliated with or endorsed by them. Figures cited are drawn from the public sources noted; where results are commercially confidential they are shown directionally or available on request. Outcomes reflect a specific engagement and are not a guarantee of future results.

FAQ

What is agentic AI in marketing strategy?

Agentic AI describes systems where multiple specialised AI agents — analyse, overlay, shortlist, validate — collaborate inside a defined workflow to deliver an end-to-end output that a single model can't. In marketing strategy, that turns a process traditionally run by teams of consultants over many months into a sequenced, gated workflow that produces a strategy in weeks. The senior judgment calls stay with humans.

How is agentic AI different from generative AI?

Generative AI produces an output from a prompt. Agentic AI plans and executes a chain of tasks against a goal — calling tools, checking its own work, routing to the next agent. For strategy, it's the difference between a model drafting a deck and a workflow producing a sourced, validated, brand-specific strategy.

How can I trust agentic AI with brand and consumer data?

Two layers. Architecturally — closed environments, no model training on client data, audit trails on every agent step, and human approval gates at each phase boundary. Methodologically — every agentic output is overlaid with brand truth, screened for claims and regulatory risk, validated against existing insight, and gated by a senior call before it moves forward. The point isn't to remove the human; it's to give the human better leverage.

Does agentic AI replace the senior strategy team?

No. It changes the cost of getting to a defensible answer. Senior strategy work becomes defining the workflow, holding the brand judgment calls and validating outputs — not running the discovery from scratch. The seniority moves up, not out.

How long does it take to build an agentic-AI content strategy?

On this engagement: eight weeks to deliver three brand strategies — roughly 3× faster than the typical Tier-1 consultancy timeline for a single brand. Speed comes from defining the workflow once, proving it on the hardest brand, then cascading to the others as deployment rather than parallel discovery.

Can I see the detail behind the headline figures?

Some additional detail can be shared under NDA — please request the full case study or start a conversation below.

← Index

What's next

Two ways forward.

The full case

Method, per-brand build, validation gates and commercial ROI — shared on request.

Request the full case

A conversation

About building an agentic AI content strategy across your portfolio. DEPTH or FOCUS.

Start a conversation