AI's great expectations

Sean Evers at Pipedrive looks at today’s emerging technology and asks: what do we want from AI?

 

Just as businesses were getting familiar with AI tools like ChatGPT, Google Gemini and Microsoft Copilot, new contenders like DeepSeek emerged, shaking up markets and reshaping expectations. The rapid evolution of AI raises the question for businesses: what do we want from AI?

 

One of the most critical decisions is between generic AI – widely available tools for broad applications – and specific AI, which is tailored to particular industries or business functions. The divide is becoming more important as businesses move beyond experimentation and look for AI solutions that deliver measurable impact.

 

Our state of AI in business report 2024 highlights how businesses are navigating this shift. According to the report, 35% of businesses have adopted AI, but a challenge remains: identifying what AI solution will genuinely work for their specific needs to make informed investment decisions.

 

 

What do businesses want from AI? 

AI adoption in business is accelerating and what companies really want from AI is becoming clearer. Right now, the majority (86%) of AI-using businesses said they leverage ChatGPT, while tools like Siri (24%) and Google Assistant (19%) are less popular.

 

Businesses are more drawn to AI tools that deliver advanced and tangible outputs from text creation to blog posts, emails and product descriptions. It’s also being used for summarisation, and filtering information into digestible insights.

 

 

The shift from generic to specific AI

Businesses are using AI in practical, assertive ways to help with administrative heavy tasks like data entry, summarisation and even research. These capabilities help free up time for high-value work, especially in industries such as sales. 

 

Yet, a number of businesses are beginning to recognise that generic AI doesn’t always understand the nuances of their industry. While general-purpose AI models like ChatGPT offer broad functionality, companies are beginning to wonder whether this AI really gets their business.

 

At this point, some companies will tweak AI to suit their specific needs. Others will wait for more tailored, industry-specific AI to be developed. A smaller group with more technical expertise will begin developing their own model to maintain a competitive edge.

 

Generally, businesses want three things from AI: 

• AI that works alongside them: Instead of replacing humans, AI should help with tasks like data entry or summarising calls, giving smarter insights that free up time for more impactful work. This ‘feels’ right, as it supports an improved but fundamentally unchanged status quo.
• AI that gets their business: Respondents agreed that the most useful tools are the ones that understand their specific industry and workflows, not just generic solutions. This is more interesting, as it’s less likely that cutting edge AI the likes of OpenAI will become widespread from a variety of vendors at pace. Some organisations are setting the pace of technological innovation and others are adopting and learning the lessons as followers. So, end user organisations may want to adopt and adapt cutting-edge solutions, wait for tailored solutions to become available, or forge ahead with their own development, if they have the expertise to do so.
• AI that’s easy to use: Great AI should integrate smoothly into existing processes and be intuitive enough that teams can jump in without needing a technology degree. OpenAI and DeepSeek are great exemplars. There is a tension between growing skills and having tools that ‘just work’ with the minimum of business disruption and time investment.

Striking this balance is key for success with emerging new tools and also ensuring humans employ their institutional, creative, and emotional skills where they can make a greater impact.

 

 

What do we want: specific, general, any way we can

Beyond productivity, AI is valued for gathering insights and improving customer interactions and satisfaction. 42% of respondents cited insight generation as a significant benefit. This can range from understanding customer preferences and behaviours to identifying market trends and optimising strategies.

 

With such insights businesses can make more informed decisions, tailor offerings to better meet customer needs, staying competitive. Tools like chatbots and virtual assistants can provide faster and more personalised customer service, responding to inquiries in real-time and offering tailored recommendations.

 

Many use cases are like this: broadly applicable to many departments and businesses. More specific use cases are also already available but are still quite new to the general user. For example, agentic AI can make decisions on its own in areas like IT operations management - failing over to a new system if a server goes down.

 

Limited use cases often provide great guardrails for AI to pick between a circumscribed set of choices and provide value from the speed of their decision-making. Consequently, more limited AIs, particularly agentic ones, can be used to raise quality of service and meet SLAs around speed of delivery and uptime.

 

The critical heuristic in investing in AI boils down to ‘where will I see the most value”. Get granular. AI doesn’t do everything well. Decide where the greatest value comes from. This might be something applicable across functions, like generating content (for 75% of AI users), content summarisation (52%), transcribing and summarising and creating actions (29%), or research purposes (24%)? Or it might involve supercharging a critical function, like sales: report generation (17%), AI-driven marketing tools (14%), and tools for data analysis identifying sales patterns, customer insights, etc (14%).

 

Some industries, like software development itself, have transformed their workflows with AI. Each business must decide where the right value and the biggest gains can come from and invest to experiment and learn what’s possible for them. Going for ‘specific AI’ choices requires careful thought and reasonable certainty before investing in something that interacts with company data or critical systems or business functions.

 

 

When do we want it: unclear

Yet, the comparatively low rate of adoption shows that while headlines blaze about AI, organisations aren’t all jumping into these changes heedlessly. Nearly half of businesses that haven’t adopted AI cite a lack of knowledge as the primary barrier. If a company doesn’t know where to start or how to integrate AI effectively into their existing processes, they are wise not to plunge headlong into a transformation that can profoundly affect business processes and customer data.

 

Legitimate concerns around trust and data privacy compound the issue: 40% of respondents are hesitant to adopt AI due to a lack of trust in the technology. Those concerns are fair, and the technology industry must allay fears with transparency, adherence to standards and best practices, collective education and honest information sharing on risks and mitigation.

 

Early adopters of AI succeed by targeting specific decision points where enhanced information processing creates measurable advantages, rather than applying the technology broadly. The true competitive edge comes from the combination of accelerated decision velocity and improved decision quality, enabling organisations to respond to market shifts faster while considering more variables and patterns.

 

Those investing in AI capabilities today aren’t just gaining immediate benefits - they’re building institutional knowledge and data assets that compound over time, creating barriers that will become increasingly difficult for followers to overcome.

 

Slower moving organisations have their reasons. Data privacy and security risks were flagged by 27% and 26% of businesses. It may be that an awareness of low-maturity IT resilience and cyber-security means some organisations delay intensive and higher risk data applications. Positively, only 4% of respondents cited “resistance from stakeholders” as a barrier to AI adoption.

 

There’s a positive trend from the bottom up to use AI. The only question is how best to integrate it for the specific needs of the business and the skills of the talent already employed - and when planning for the future state of the business. It can pay to pull that trigger and move forward, if the business has a clear enough vision for success. 

 

---

 

Sean Evers is VP of Sales at Pipedrive

 

Main image courtesy of iStockPhoto.com and Valerii Apetroaiei