From scattered knowledge to a scalable ideas generator
The Challenge
An award-winning, PE-backed communications consultancy wanted their account teams to proactively bring fresh ideas to clients between formal planning cycles, not just at quarterly planning meetings.
But the knowledge they needed to do that was scattered across messaging documents, planning decks, emails, client updates and daily news monitoring.
Proactive thinking relied heavily on individual memory and manual monitoring and was often too time consuming for busy account teams to carry out regularly.
What it does
We designed and built an agentic workflow called the “Proactive Ideas Agent” that combines each client’s strategy, priorities and business context with daily monitoring, then generates timely, on-strategy ideas for account leads to develop and take to clients.
Firstly, we created a ‘knowledge layer’ for each client - built from messaging documents, marketing materials, the client’s strategy, their priorities and relevant business context. This brought scattered client knowledge into one consistent, reusable ‘memory’ source for the whole team.
Next, the agency’s daily monitoring of relevant news, client updates and their client-facing work was combined with that knowledge layer, keeping the agent current on both the client and the market around them.
These components feed into their Proactive Ideas Agent, which account teams now use daily to recommend timely, on-strategy ideas for account leads to develop and take to clients.
Every output is reviewed by an account lead before it informs client work. Feedback from the internal team and clients can then be input into the agent, helping improve the quality and relevance of future ideas over time.
Embedding proactive thinking across a growing share of the agency’s client base.
The agent helps account teams identify timely, relevant opportunities for clients. AI-generated ideas have already led to securing op-eds.
Client satisfaction is improving as the agent has helped account teams bring timely ideas to clients beyond formal planning cycles.
Reusable knowledge infrastructure created giving the agency a foundation that can support additional AI tools and agents over time.
Turning one strategist's approach into scalable, sellable IP
The Challenge
21st Century Brand is a brand strategy consultancy behind some of the world’s most influential brands.
As brand discovery shifted from search engines to AI assistants (e.g. Chat GPT, Claude, Copilot), they developed a rigorous methodology for measuring and improving brand visibility and positioning in an AI era.
But the methodology took 5 days of work, and could only be delivered by one person in the business, capping revenue growth and putting valuable IP at risk if that person left the business.
The Solution
Transforming a multiple-stage process into a digital service
- We codified the methodology, mapped the manual processes across multiple tools and turned them into a productised digital service the wider strategy team could use
- The service automates much of the manual research, querying multiple Ai models concurrently, as well as the sources those models draw on, to assess how a brand's positioning is actually being shaped.
- Senior strategists helped test and refine the service throughout development. Their feedback is now embedded into a continuous improvement loop, strengthening the methodology and improving the quality of recommendations over time.
5 days work reduced to 1 freeing up capacity for higher-value strategic client work
Removed senior bottleneck and risk by transforming individual expertise into a service the whole team uses
IP strengthened through use as strategist feedback continuously refines the methodology and improves the service
New sellable IP beyond billable hours giving the consultancy a unique, repeatable digital service it can sell, refine and scale
Opens productised and recurring revenue opportunities beyond billable hours
Turning 10 years of delivery insight into searchable institutional memory
Why this matters
Teams consistently generate valuable learning through delivery retrospectives and hindsight documents, but these insights currently live in decks within nested folders.
They are only retrievable if someone already knows they exist or remembers the project.
As a result, new joiners, PMs and strategists often start from scratch, repeating mistakes or missing proven approaches. Over time, this creates knowledge leakage and slows capability building.
What it does
An internal knowledge system, acting as a memory layer for the organisation that:
- Ingests hindsight docs, project summaries and key learnings
- Allows natural-language search such as “Have we done this before?” or “What worked on similar projects?”
- Enables teams to surface relevant insights at the start of new projects or during new business
Faster onboarding and stronger decision-making for PMs and new joiners (increases ability to scale)
Reduced repetition of past mistakes
Cumulative improvement in delivery quality over time
From people-scaled agency to system-scaled growth engine
Why this matters
Performance marketing has shifted, advancements in targeting algorithms mean optimal-spend is directly linked to effective creative volume.
From 50 ads a month in the old world, to 500 ads in the new world, but advertisers:
- Rely on manual reporting and reactive optimisation
- Struggle to produce enough high-quality creative variations
- Hit growth ceilings as internal teams become bottlenecks
What it does
Completely reimaged performance marketing engine, with:
- Automated performance insight extraction at scale
- Continuous AI driven analysis of top- and low-performing ads to identify repeatable patterns (hooks, formats, offers, visual styles, messaging angles)
- AI-supported creative production workflows. Creating net new creative (video and static) leveraging Kling / Seedance / Veo3 models
- Human-in-the-loop feedback mechanisms
Scales high-performing ads without increasing headcount
Increase iteration velocity
Revenue uplift examples (+38%)
Creates defensibility through proprietary data + workflow integration
Reduce creative production time (claimed up to 70%)
Explore other case studies