Generative AI has exploded into the business world, with tools like ChatGPT demonstrating powerful capabilities to generate text, code, images, and more. While the potential is massive—McKinsey estimates up to $4.4 trillion in annual value—most companies struggle to move from experimentation to real impact. A 2024 survey shows 65% Generative AI adoption, but few have a clear strategy or know how to realize ROI. This guide helps you bridge that gap.

Why this overview will be useful for you?

Based on Superlinear’s experience and industry research, this guide offers a practical roadmap: from understanding GenAI’s potential to tackling the seven most common challenges in unlocking its value. With examples from top companies and actionable advice, it’s designed to help you turn GenAI into a business advantage—not just a tech demo.

Generative AI’s business impact

Generative AI enables organizations to create content, automate tasks, and uncover insights in powerful new ways:

This opens up a wide range of business applications, for example:

• Enhanced Customer Experience: Best Buy’s virtual AI assistant troubleshoots products and frees human agents for complex issues. Carrefour’s “Hopla” helps shoppers plan meals and select products. Both examples improve personalization, convenience, and sales.

• Content and Marketing Innovation: Coca-Cola’s “Create Real Magic” campaign let consumers co-create AI-driven artwork from iconic brand assets, generating over 120,000 pieces of content. Marketers use generative AI to produce ads, visuals, and targeted copy faster and at scale.

• Knowledge Management and Decision Support: Morgan Stanley’s GPT-4-powered assistant searches vast research libraries to give wealth managers instant insights. Adoption by 98% of Morgan Stanley’s advisor teams shows how internal “AI copilots” can streamline knowledge work across industries.

• Operational Efficiency and Automation: Generative AI can also streamline internal operations. In software development, tools like GitHub Copilot and Cursor write code snippets and applications, speeding up programming tasks. In customer support, AI can draft initial responses or summarize customer tickets.

• Innovation and New Product Development: Commerzbank’s AI-driven virtual banking avatar offers personalized natural-language advice. From prototype designs to AI-based services, organizations use generative AI to differentiate themselves and develop new revenue streams.

Thus, Generative AI can drive real business value by enhancing human capabilities, as seen in examples like reduced HR support costs and improved supply chain efficiency.

But unlocking this value at scale isn’t as simple as using ChatGPT or adding new software. Many pilots show promise but fail to scale due to integration and operational hurdles. To achieve lasting impact, businesses must address key strategic and organizational challenges. The next sections explore seven common obstacles and how to overcome them with practical solutions and examples.

The 7 challenges and their solutions

1. Lack of a clear generative AI strategy and use-case focus

Challenge: Many organizations pursue AI in an ad-hoc fashion, launching pilots without linking them to strategic objectives or tangible metrics. This leads to “random acts of AI” that seldom deliver lasting value.

Solutions:

• Define a high-level AI strategy aligning with business goals (e.g., revenue growth, cost reduction, innovation).

• Identify and prioritize 2–3 high-impact use cases (e.g., automating customer support or personalizing marketing campaigns).

• Develop an AI roadmap to move from pilots to production, secure executive buy-in, and coordinate resources.

For example, Superlinear’s Applied AI Discovery workshops help organizations pinpoint and plan relevant AI initiatives that align with strategic objectives.

2. Data readiness and integration into workflows

Challenge: Generative AI is only as good as the data it has access to—and its effectiveness depends on embedding it into existing processes. Many organizations’ data is siloed or unstructured, and AI tools are often only loosely connected to systems where people work.

Solutions:

• Conduct a data audit and ensure relevant information is accessible, clean, and compliant.

• Integrate AI tools with existing CRMs, ERPs, or communication platforms via APIs or native connectors, so users can access generative AI naturally.

For example, Superlinear worked with the Port of Antwerp Bruges to develop “APICA Chat”, which uses an LLM to retrieve and query diverse data (including SQL databases and PDF files) for streamlined answers within everyday workflows.

3. Custom vs. off-the-shelf AI solutions

Challenge: Companies must decide between leveraging pre-built SaaS solutions (fast to deploy but less customizable) or developing custom AI (greater control but resource-intensive).

Solutions: 

• Use third-party AI for quick wins on generic tasks (e.g., drafting copy, basic chatbots).

• Build custom solutions for strategic differentiation or when data sensitivity is high.

• Adopt a hybrid approach: begin with SaaS for speed, then transition to custom models for core, high-value use cases.

For example, Bloomberg built a domain-specific LLM (BloombergGPT) for specialized finance applications, while many marketing teams use off-the-shelf AI for everyday copywriting (e.g., Jasper, Copy.ai, Writer).

4. Talent and skills gaps

Challenge: Generative AI requires specialized data and engineering skills, as well as domain expertise. End-users and managers often need training to work effectively with AI.

Solutions:

• Upskill existing employees with AI basics (prompt engineering, data literacy) and hire selectively to fill critical roles like ML engineers or data architects.

• Collaborate with AI consultancies or technology partners for targeted expertise.

• Form cross-functional teams that pair AI specialists with domain experts so solutions remain relevant to the business.

For example, Johnson & Johnson’s 2025 upskilling initiative trained over 50,000 employees in generative AI usage, rapidly closing talent gaps across its global workforce.

5. GenAI governance, risk management, and ethical compliance

Challenge: Generative AI can produce biased or incorrect outputs (“hallucinations”), risk data privacy breaches, and violate regulations if left unchecked.

Solutions:

• Implement AI usage policies (e.g., no sensitive data in public models, mandatory human review).

• Maintain human-in-the-loop review, especially for customer-facing applications.

• Adopt technical controls (model filters, sandbox environments) and test AI rigorously.

• Ensure data security, compliance, and transparent oversight through a governance committee.

For example, financial services firms use private cloud instances of GPT-4 or Azure OpenAI for tighter data control and regulatory compliance.

6. Change management and user adoption

Challenge: Without cultural and procedural support, even the best AI tools can languish unused. Resistance to new technology, fear of job displacement, or lack of training often undermine AI rollouts.

Solutions:

• Secure leadership buy-in to champion AI initiatives and articulate benefits.

• Start with pilot groups to create success stories and convert colleagues into AI evangelists.

• Integrate AI into everyday workflows, making it the “default option” where possible.

• Provide thorough training and maintain feedback loops to refine the system and foster trust.

For example, Accenture leverages “change champions” across key business units to share lessons learned, train peers, and accelerate organization-wide adoption of genAI tools.

7. Measuring impact and ensuring GenAI ROI

Challenge: Organizations often fail to define clear metrics or track the results of AI projects, making it hard to prove ROI or identify when a project isn’t delivering.

Solutions:

• Set success metrics early—efficiency gains, error reduction, increased sales, etc.

• Track both quantitative (handle times, revenue lift) and qualitative (customer satisfaction, employee feedback) indicators.

• Regularly review progress and pivot if targets aren’t met.

For example, Morgan Stanley measures adoption (98% usage) and access to documents by its GPT-4 assistant, illustrating tangible business impact.

Conclusion

Generative AI can drive meaningful, transformative results—from streamlining operations to boosting customer engagement and innovation. However, success isn’t just about deploying new technology; it hinges on strategy, data readiness, governance, culture, and measurable outcomes. By focusing on these seven challenges and their solutions, organizations can shift generative AI from a novelty to a true driver of competitive advantage.

Here are some concrete next steps:

1. Assess Your Starting Point. Review existing AI activities, data quality, and skills.

2. Develop (or Refine) Your AI Strategy. Align generative AI with business goals; secure leadership buy-in for high-impact use cases.

3. Engage Experts and Build Knowledge. Consider partnerships or consultancies to accelerate value delivery and skill-building.

4. Pilot, Iterate, Scale. Launch focused pilots tied to strategic metrics; measure, refine, and expand.

5. Foster a culture of innovation and responsibility. Encourage responsible AI adoption through clear guidelines and ethics, ensuring sustainable, long-term growth.

With a disciplined, value-oriented approach, enterprises can tap into the immense potential of generative AI and position themselves to lead in the new era of intelligent innovation.

An extended version of this overview can be found here.