EP14: Automate, Optimise, Empower: Human-Centred AI with AI Tappers

Coaching, Not Consulting

AI Tappers' mission centres on building sustainable internal AI capability, rather than creating dependency on external consultants. Marcus draws a memorable analogy: like a mechanic explaining a car problem and fixing it, but also making sure the owner understands enough to handle the next issue themselves. That philosophy runs through everything from building automations to hands-on staff training.

The Discovery Phase: Understanding Before Building

Every engagement begins with a deep discovery phase, mapping an organisation's operations, workflows, and technical constraints — including what tools or platforms they're already committed to. As Marcus notes, if a company is locked into Microsoft Copilot for compliance reasons, that becomes the starting "car" they have to work with. Crucially, this phase also involves breaking down what "AI" actually means in practical terms for that specific organisation, cutting through buzzwords like "agentic AI" to establish a shared, working understanding.

AI Adoption Isn't Just an IT Department Problem

One of the more striking insights from the conversation is how AI leadership within organisations doesn't necessarily sit with the CTO. Marcus describes working with a Head of AI who comes from a completely different department — arguing that understanding how the business actually works often matters more than deep technical expertise. Successful AI transformation, in their experience, requires buy-in across the C-suite — CEO, CFO, CMO, and CTO together — not just an IT-led initiative.

The Danger of "Shadow AI"

Mattias and Marcus highlight a common and risky pattern: employees using personal AI accounts to process sensitive company information, simply because their organisation hasn't provided proper guidance or training on safe usage — so-called "shadow AI." Rather than organisations simply banning tools outright, AI Tappers argues for training staff to understand why certain restrictions exist, and what they genuinely can and can't do safely — for example, anonymising sensitive data before using a general-purpose AI tool, rather than avoiding AI-assisted work altogether.

Why Restrictive IT Policies Backfire

A recurring theme is the gap between blanket restrictions and genuine enablement. Mattias describes organisations that tell staff "you cannot use this" without ever explaining what can be done safely within those constraints — often pushing employees toward lower-capability, "sanctioned" tools without the training needed to use them effectively. The size of an organisation matters significantly here: smaller companies can make decisions and pivot quickly, while larger organisations often face months of internal approval processes before any meaningful AI initiative can move forward.

Starting Small: The "Virus" Approach to Adoption

Rather than attempting large-scale, top-down AI transformation, AI Tappers deliberately starts with hyper-specific use cases — solving one person's workflow problem, letting the results spread organically through word of mouth as colleagues notice the time savings. Marcus describes this as becoming a kind of "virus" within an organisation — genuine enthusiasm spreading from one successful, tangible use case to the next, rather than attempting to "remodel the whole house" with AI all at once.

Training as a Continuous, Human-Centred Process

Both founders emphasise that effective AI training goes well beyond a single onboarding session. Using a gym membership analogy, Matias explains that just as no one would expect to walk into a gym and immediately understand proper weightlifting technique, employees need proper, ongoing coaching to use AI tools safely and effectively — including foundational skills like prompt engineering and understanding a model's practical limitations (likened to knowing when to "change the oil" in a car, a nod to context window limits).

AI Tappers also embeds AI directly into live team projects — for example, working alongside a marketing team on a genuine product launch, using AI collaboratively from the outset rather than bolting it on partway through a process. This approach, they find, genuinely enhances team creativity rather than replacing it.

Uncovering Hidden AI Champions

A particularly valuable outcome of this collaborative approach is the emergence of informal "AI champions" within client organisations — often employees who've quietly been experimenting with AI tools on their own initiative, who then step into a more visible, influential role once given the opportunity and confidence to share what they know across their department or team.

Measuring Success: Time Saved, Redirected to Value

Asked about KPIs, Mattias and Marcus are clear that success isn't simply about replacing staff with automation — it's about freeing people from repetitive, low-value manual work so they can focus on higher-value, revenue-generating, or genuinely creative activity. They share an example of a research analyst who reduced a monthly research task from around 40 hours to just a few minutes using an automated workflow — not eliminating his role, but repositioning his time toward more strategic, actionable work, while still applying human judgement to verify the AI's output.