What’s Stopping Companies From Becoming Successful With AI

30 Dec

Here are the core reasons why companies are struggling to be successful rolling out AI, from not having one launch event, to having internal champions who help to push, promote and coach colleagues, to not being prepared with the right level of data and structure for AI.

Policy: Waiting for an AI policy from IT and compliance - IT ownership won’t help you grow as its hard to manage and control
Fear of shadow AI (employees using AI tools with company data) and not wanting any work on AI tools they haven’t signed off.
No Plan: Not having a clear idea of what success looks like and just wanting more output
No Process: A lack of understanding of processes and how their teams and department processes will be added into AI and optimised to help AI to be rolled out correctly and optimise processes and ops
Data: Data stored across different storage providers - on prem, different cloud platforms - if data doesn’t connect well or can be connected AI deployments and workshops will fail
Fear of AI - what it can do, what it could do, what its done in other businesses (like delete databases) - with the right partner this can be reduced and removed
AI Wants vs AI Needs - what you (as a business leader) want isn’t what you need (often your wants and team wants are off and it can do more than more output)
Briefless: Being able to brief your requirements
Documentation: Having enough documentation and processes listed to upload, critique and optimise
Productivity → ROI: Investment in time is hard to quantify, so time is often not seen as a driver of business change
ROI Hard To Track: AI ROI, especially in Marketing departments
AI Translation: People’s abilities to articulate a problem or set of problems and create workflows and understand where there can be optimisations and automations create real disconnect
Time To Train LLMs: Not being prepared with the basics to train LLMs and AI tools. Not having to right level of brand guidelines, deep domain knowledge to share and ability to onboard AI well.
Perfect Can Wait: Waiting for perfect AI - no such thing as perfect AI or AI tech stack
Forcing Cross-functional AI - when some teams will need specific sets of AI tools and others will be a completely different set, the underlying tools and tech aren’t built for multi-player yet
Replacement First Mentality: Expecting AI will replace team members before any review of the team, team process and what people are doing in their actual jobs are documented and identified to make change
Angry Automation: Lack of understanding that AI automations break and regualrly
Bridging: Expecting to AI to remove silos - if anything it can widen the silo’s without prep, foresight and thinking beyond data analysis can be done, so we can reduce the requirement on our data team
Misled: AI Tools don’t do what they say they do. Most are upto 80%. There are a lot of tools that seem helpful but there are no ways to prove amendments has really worked
AI Is Single Player: AI isn’t multi player yet, it is still single player
Culture shift! Changing Problem to PowerPoint to prototype is a shift in work and shift in culture. Building prototypes is a brand new culture for many team members and departments. This needs to be taught and part of the journey
Too High AI Exceptions: AI will instantly become agents and then agentic and automation will just happen.
Replacement As Default: many believe AI will replace time, investment, processes, headcount, hard work - over time and with training, it will. However, expecting to remove and replace in days to weeks is unrealistic and an expensive issue for many businesses
Single Roll-out Struggles - multi-layer deployment is the only way to truly have a positive rollout

Below is how I build successful AI workshops and ensure departments are driving forward, and adopt the right approach from my AI Podcast called AI moment