AI Agents: Opportunities vs Risks – Key Learnings for Executives and Innovators

Reading Time: 5 minutes

Artificial Intelligence (AI) Agents have moved beyond theory – they’re here, operating inside marketing departments, procurement teams, customer service desks, and even negotiating on behalf of major retailers. 

Last week, MoEngage and Artefact hosted a webinar on “AI Agents: Opportunities vs Risks”, featuring a stellar panel:

• Amelia Kallman, futurist and top 25 Woman in the Metaverse, who advises global brands like Unilever, Allianz, and Red Bull on AI opportunities and pitfalls.
• Harry Morphakis, AI strategist helping S&P 500 and FTSE 100 companies turn emerging tech into measurable results.
• Ashwin Srinivasan, VP of Product, Marketing Technology and Digital Advertising at MoEngage.
• Gitte Ganderup, board advisor, AI startup mentor, and consultant at AI Forge and Rannoch Associates.
• Jan Hendrik Fleury, Senior Director of Data Consulting at Artefact, with over 15 years of expertise in transforming businesses through AI and analytics.

If you missed it, you can watch the on-demand recording here and access every insight we’re going to unpack below.

What real-world business tasks are AI Agents automating today?

The session wasted no time grounding us in reality: AI Agents are already delivering measurable ROI in sectors from retail to finance.  

Harry Morphakis described working with a procurement team to train AI Agents to extract structured data from thousands of unstructured invoices, then automatically process purchase orders – slashing back office time and cost.

Ashwin Srinivasan shared how some sectors are using “coding agents” to speed up engineering workflows, and how clients like Fastic are driving customer engagement with hyper-personalised communication at scale with the help of AI.  

Amelia Kallman highlighted compliance automation in marketing for regulated industries – moving campaign approvals from weeks to minutes, with billions in potential industry wide gains.

Gitte Ganderup gave perhaps the most disruptive example: an AI negotiation agent handling ongoing supplier discussions for a supermarket chain. The cascade of benefits? Category managers focus on strategy, small suppliers get regular communication, and millions in extra value are unlocked.

What top 5 opportunities AI Agents create for businesses?

From the discussion, five standout opportunity areas emerged:

1. Process efficiency gains – Automating decision making from invoices, compliance, and negotiation processes.
2. Cost reduction – Fewer manual tasks, faster cycles, leaner operations.
3. New revenue streams – Dynamic market adaptation and realtime negotiation open doors to incremental sales.
4. Hyper-personalisation at scale – AI Agents can deliver targeted offers, content, and experiences 24/7 without fatigue.
5. Strategic focus for teams – Instead of drowning in operational minutiae, skilled employees can focus on innovation, strategy, and customer relationships.

But when senior leaders hear “efficiency”, they often miss a crucial angle: cultural alignment and human empowerment.

What are the biggest AI risks for brands and how can they avoid them?

The panel was clear: with big opportunity comes… hidden danger.

The recent Mind the AI Gap: What Consumers Expect From Brands by 2030 – survey by Artefact and MoEngage uncovered some sobering truths:

• 41% of consumers will be more loyal to brands transparent about AI use.
• Nearly 1 in 3 would lose trust if they couldn’t tell whether interaction was with a human or a bot. 

Amelia Kallman’s warning: AI is not inherently good or bad – it reflects our values. Without clear protocols, guardrails, and “human in the loop” approaches, risks to jobs, profits, and long-term trust quickly mount.

Gitte Ganderup added: It’s vital to start adoption where it empowers, not replaces, staff – minimising fear and building buy-in. That means training employees not just to “prompt” but to understand what’s underneath these tools.

What are the best guardrails to innovate with AI without losing control?

This part of the conversation resonated deeply for every technology investment leader.

The Guardrails:

• Audit & traceability – Make AI decisions visible, inspectable, and explainable.  
• Input & output filters – As Ashwin put it, apply constraints at both ends to block risky prompts or unsafe outputs.  
• Scope control – Limit autonomy in high-risk areas and keep humans in the loop for critical actions. 

Harry Morphakis reframed AI risk management as an ongoing effort. AI Agents are living systems – the moment you consider them “done” is the moment you lose control.

Gitte’s advice is to choose starting points with high business value BUT low implementation complexity. Don’t unleash autonomous refunds or supplier payments on day one – match risk level to oversight capacity.

How can businesses avoid AI vendor dependency traps?

Overdependence on a single AI vendor, platform, or tightly integrated stack risks operational fragility.  

Harry’s recommendation: Architect systems with observability and rapid failover plans. Test them regularly. Decide upfront which components can be open source and which will inevitably be vendorised, then mitigate accordingly.

How can businesses build AI trust with customers and employees?

One of the most discussed threads was trust.

For customers, honesty about AI use is now a loyalty driver. For employees, assurance that AI is there to elevate and not eliminate their role is crucial.

Practical ways to build trust:

• Radical transparency about where AI is active.  
• Clear escalation rules for high risk scenarios.  
• Continuous training – from frontline prompt skills to board level AI literacy.  
• Value by design – embedding your brand’s ethics, inclusivity, and sustainability principles into AI models themselves.

What AI trends should businesses watch – and why Generative Engine Optimisation (GEO) will be critical?

Looking ahead 12–36 months, the panel painted a bold picture:

• Ashwin expects autonomous decisioning in marketing to reach a new level of precision – without adding customer fatigue along with personalized app experiences unique to each user’s workflows.
• Harry is excited about longer “context windows” for AI Agents – enabling them to autonomously complete multi-hour workflows and integrate more deeply into full stack business processes.
• Amelia foresees a new competitive space: Generative Engine Optimisation (GEO). As bots make more purchasing decisions, winning visibility in their data driven “search” becomes as critical as SEO is today.
• Gitte predicts org charts evolving into work charts, with AI Agents alongside humans – forcing HR to redesign reskilling, values enforcement, and change management.

What are the most important AI strategy takeaways for executives shared in our webinar AI Agents: Opportunities vs Risks?

If you’re leading corporate strategy, digital transformation, R&D, or technology investments, here’s the distilled blueprint from the panel:

1. Start with governance – Safety, compliance, and transparency should be foundational, not retrofitted.
2. Pilot with purpose – Choose high value, low complexity domains first. Build quick wins without creating hidden debt.
3. Educate continuously – AI literacy is now a board level necessity.
4. Design for resilience – Avoid over dependence, build backups, test failovers.
5. Measure human ROI – Time saved, headaches reduced, employee empowerment – as critical as financial ROI.
6. Prepare for GEO – Data formats optimised for machine “audiences” will be an entirely new competitive lever.

Partner with AI experts to build safer and high-impact AI Agents strategies

The difference between early adopters who succeed with leveraging AI Agents and those who struggle will come down to strategic design, ethical governance, and a clear focus on human value alongside technological capability.

That’s why partnering with the right experts matters. MoEngage and Artefact bring together deep AI engagement expertise and data‑driven business strategy to help organisations implement innovation that’s safe, scalable, and aligned with customer trust.

As Amelia Kallman reminded us in the webinar, trust is currency – and while technology may move fast, earning lasting trust takes time, transparency, and thoughtful execution.

Your next steps

• Download the Whitepaper –  Mind the AI Gap: What Consumers Expect From Brands by 2030 will give you the hard data and actionable strategies to align AI adoption with evolving customer expectations. Download here
• Watch the Webinar On-Demand – The full panel discussion goes deeper into each of these examples and offers practical frameworks to start using right now. Watch here