Zandra Moore and Charlie Bartle explore how organisations, particularly in financial services, are moving from AI curiosity to implementation. The discussion focuses on where firms currently stand, why many executives feel pressure to act, and why the real constraint is often not technology but education, prioritisation and change management.
They argue that the strongest early returns come from practical internal use cases, especially those that improve operational workflows, unlock institutional knowledge and support compliance.
They also discuss the evolution from chat-based tools to assistants and agents, the need for governance and human involvement, and why firms should start with simple, low-risk experiments rather than waiting for perfect certainty.
Find out more about Zygens -> Here.
Key Take Aways
- Many organisations are concerned that they are behind the AI curve and are unsure whether they are moving quickly enough.
- Adoption remains uneven: some employees are experimenting actively, while others have not yet engaged with the tools available to them.
- A material education gap persists between personal experimentation with AI and effective application in the workplace.
- The defining feature of this technology wave is speed; the capability gap between early adopters and others is widening faster than in previous digital shifts.
- The core challenge is not a shortage of AI opportunities, but uncertainty over where to start and which use cases should be prioritised first.
- The most practical early use cases sit inside existing operational workflows, where AI can remove low-value manual effort and free up capacity.
- Internal workflow use cases are attractive because they are often measurable, high value and relatively lower risk than externally facing deployments.
- Domain experts should play a central role in identifying and shaping use cases, rather than leaving deployment solely to technical teams.
- AI agents should be approached as a progression, beginning with assistants and semi-autonomous systems before moving to more complex autonomous models.
- Cultural adoption matters as much as technical deployment; employees want to be involved in the conversation about how AI will affect their work.
- Governance is essential, both to manage risk and to avoid using AI as a sticking plaster for broken underlying processes.
- In financial services, the strongest near-term opportunities identified are knowledge access and compliance-related use cases.
Innovatation
- Using AI primarily to improve internal operational workflows rather than focusing only on high-profile, headline-grabbing applications.
- Treating AI deployment as a cultural transformation, with operational teams and domain experts co-designing solutions.
- Framing assistants as the first practical step in the agent journey, allowing organisations to build confidence before scaling into more autonomous systems.
- Building low-risk, high-value assistants and semi-autonomous agents with humans in the loop.
- Applying sandbox, proof-of-concept and MVP approaches to test use cases safely before broader deployment.
- Establishing AI governance councils that combine domain experts, workflow owners and technical teams to review use cases and monitor impact.
- Deploying models securely within an organisation’s own infrastructure or “walled garden” rather than relying solely on external models.
- Creating knowledge assistants to unlock unstructured information held in documents or in employees’ heads.
- Using AI in compliance processes to check activity against internal or regulatory standards and remove bottlenecks.
Key Statistics
- A single assistant has been seen to deliver 8 to 10 hours per week of time savings.
- Nearly 400 agents designed across financial services in its first year.
- The average enterprise now has around 120 SaaS platforms.
- Large corporates are described as having gone through 20 to 30 years of digital innovation, adoption and transformation cycles.
Key Discussion Points
- Why many firms feel pressure to accelerate AI adoption despite uncertainty over what “good” looks like.
- The gap between curiosity about AI and the practical competence required to use it effectively at work.
- The unusually fast pace of this technology cycle and the risk of leaving people behind.
- Why identifying the right starting point is harder than finding possible use cases.
- The case for prioritising operational productivity use cases over more fashionable but less practical applications.
- The importance of involving domain experts and frontline teams in use-case selection and design.
- How organisations should distinguish between assistants, agents and broader automation.
- Employee concerns around hallucinations, new terminology and the perceived impact on job security.
- The opportunity to use AI-driven efficiency gains to create capacity for higher-value work and new business opportunities.
- The risk of using AI to automate poor processes rather than redesigning them properly.
- How governance, security and compliance teams can support adoption without becoming a brake on experimentation.
- Why the most immediate value in financial services appears to be in knowledge access and compliance.
#Zygens
RO-AR insider newsletter
Receive notifications of new RO-AR content notifications: Also subscribe here - unsubscribe anytime
