LazBlazter

@lazblazter.bsky.social

#Technologist #DigitalTransformation #SwimCoach #DefenceAI #WestonHockeyClub #SecurebyDesign #Privacy #Cyber #TechSouthWest

True wherever you look, even trades people are concerned. It will go wrong before it is perfected and equilibrium established. The technical debt in decision making spaces is massive so the ripple effect from bad decisions will need to be surfed out. #DigiBlogChat

Carol Stephen@carolstephen.bsky.social · 4d ago

A10. In 12-18 months? Massive unemployment. Jobs will be given away to the AI and there’s going to be a lot of issues. Then they’ll have to rehire a lot of people. It’s all about the bottom line. I'm already seeing it here. People are worried for their jobs. #DigiBlogChat

A10. In 12-18 months? Massive unemployment. Jobs will be given away to the AI and there’s going to be a lot of issues. Then they’ll have to rehire a lot of people. It’s all about the bottom line. I'm already seeing it here. People are worried for their jobs. #DigiBlogChat

A10: Self-optimising marketing and operations loops. Instead of humans analyzing campaign metrics every Monday and tweaking copy, agents will monitor performance in real time, generate alternative variants, allocate micro-budgets, and deploy adjustments. #DigiBlogChat

Carol Stephen@carolstephen.bsky.social · 4d ago

Q10: Looking ahead 12–18 months, what will be the single most disruptive change brought by autonomous agents in the digital workspace? #DigiBlogChat

A8: Daily roles will shift from manual execution to editorial direction and systems design. Instead of spending hours writing drafts and formatting graphics, marketers will build agent logic, test guardrails, evaluate outputs, and focus on high-level creative strategy. #DigiBlogChat

Carol Stephen@carolstephen.bsky.social · 4d ago

Q8: How will autonomous workflows shift the daily roles and skills required of content creators, social media managers, and digital marketers? #DigiBlogChat

A7: The top risks are compounding errors, e.g. where a hallucination poisons the process, runaway execution loops that potentially rack up API bills, and security vulnerabilities like prompt injection when agents process untrusted public web inputs. #DigiBlogChat

Carol Stephen@carolstephen.bsky.social · 4d ago

Q7: What are the biggest operational risks (e.g., hallucinations, API costs, runaway loops, security) you worry about with agentic AI? #DigiBlogChat

A6: Quality control requires explicit style files, prompt chaining, and evaluation loops. Give the agent strict negative constraints, e.g. "never use these words", and pass outputs through a "critic" agent to rate brand etc and keep hallucinations in check. #DigiBlogChat

Carol Stephen@carolstephen.bsky.social · 4d ago

Q6: How do you maintain quality control, brand voice, and factual accuracy when content or data workflows run autonomously? #DigiBlogChat