How Research Tech Founders Overcome Objections

Overcoming AI Anxiety: How Research Tech Founders Overcome Objections

Per Insight platforms

• AI
• Artificial intelligence
• Conversational AI
• AI moderation
• Generative AI

It’s no secret that tools and solutions supported by artificial intelligence are becoming indispensable tools for research. After walking the podium at several conferences this year, an idea struck me: managing customer and user expectations for AI research technology must be challenging. There are more and more products and services supported by artificial intelligence, and navigating them is already a difficult undertaking. While everyone wants to be more efficient, few are willing to take the risk that comes with trying something new.

So, I sat down with a few founders of AI companies and thought about their experience with common customer complaints and managing user expectations:

Norbert: Tell me about your AI research technology offering and how it helps make the research process more efficient.

Greg (That’s what it says): That’s what it says is a platform that uses AI to moderate and analyze qualitative research interviews. It allows you to design, run and analyze conversational AI interviews in one place, making it faster, cheaper and easier to collect and interpret user insights.

Alex & Ran (SurveyMind): SurveyMind is a transcription analysis platform for focus groups and qualitative researchers. It offers self-service transcription and analysis, with optional consultation services to help users provide comprehensive data comparisons and thematic analysis.

Tovah (Fathom): Fathom is a topic-based coding and text analytics platform that combines AI with expert human oversight to reduce open analysis time while delivering high-quality text analytics for market researchers and customer insight teams.

Karlien (Hello Ara): Hello Ara uses AI for conversational artificial intelligence, AI analytics, and modeling of structured and unstructured data, also focusing on research in metaversal environments. The process emphasizes human-AI collaboration for creativity and efficiency.

Norbert: What do you see as the most common customer complaints about AI?

Greg (That’s what it says): The main challenges are concerns around InfoSec, data security and GDPR. That’s why we made the decision to build That’s what it says on Azure; this ensures that user data is never used to train AI models, and for our European customers, their data stays within the EEA (our servers are in Sweden and the Netherlands). We are also in the process of completing our ISO27001 certification, the gold standard in InfoSec, and of course, we are GDPR compliant. Some customers also think they expect AI to simply replicate traditional methods, only faster and cheaper. While this is true to a certain extent, our platform opens up new ways to collect insights, which may require a slight shift in mindset.

Alex & Ran (SurveyMind): Skepticism and uncertainty around AI is much more common than you think – many customers are not yet ready to fully “embed the process” due to data security and GDPR concerns. There is a need for hand-holding as clients expect guidance and reassurance along the way. What people really like to see are results. People’s expectations of artificial intelligence are much stricter than if it were a human. Everyone makes mistakes from time to time, but if an AI makes a single mistake, it is considered unreliable in many people’s minds.

Tovah (Fathom): On one side of the spectrum, many users have tried many text analytics solutions over the years and have been consistently dissatisfied. For these people, it’s about overcoming skepticism and showing them how Fathom it really provides unmatched nuance and accuracy. On the other end of the spectrum, some customers are hoping that AI is the magic bullet for themed coding and start out thinking they want a fully automated solution. But then when we get into it, they also want nuance and control and adaptability. So for those people, it’s about helping them understand how important the human-in-the-loop component is to delivering the kind of nuance, accuracy and contextual adaptability they need—with just a little human guidance. And of course, there’s data security and privacy – which is great! We’re SOC2 certified and make all our data security procedures super transparent – so we’re always happy to have those conversations.

Karlien (Hello Ara): We started selling conversational AI before the big AI boom, and we knew full well that part of our audience would have a hard time moving away from “traditional” research. In the last year, Generative AI has changed the landscape and accelerated the adoption curve. And before, we always targeted end customers who we knew were ahead of the curve in their thinking and openness to adopting new technology. Some of our clients still don’t want pure Generative AI interviewers – they’re worried about going astray (which can have reputational risks in sensitive categories) and want to know that the people in charge know how to use Generative AI.

Norbert: How do you usually handle these objections?

Greg (That’s what it says): We address data handling issues directly with transparency about our platform’s security and certificates. We also encourage each potential client to create space for experimentation and allow teams to pilot That’s what it says. Our one-time project fee allows organizations of any size to explore AI-moderated qualitative research without committing to a subscription.

Alex & Ran (SurveyMind): One of my tricks is to ask people to send me some old data that they want to structure – something that I can share. Then I plug that into our platform to help them see what our AI can do. People don’t like to see generic and fake information. We also emphasize that AI enhances, not replaces, researchers. Transparency with infrastructure and guarantees about data handling also build trust. We will not train our models on your data and we are very honest about this. We also always show our infrastructure diagrams to prove it.

Tovah (Fathom): Above all, showing tangible value that directly solves their business challenge! For a second I thought there was a cycle of AI hype where there was little research into AI solutions just for the sake of having an AI solution. That’s over. You need to deliver value that solves real problems. We offer a free trial so people can experience the power Fathom on their data! We support a lot of organizations that work with really sensitive data – so after that, it’s going to be about being good partners through the infosec process and making sure that users, managers, AI policy people, IT people are super confident about data security and privacy protocols .

Karlien (Hello Ara): Discussing clear AI use cases and ensuring skilled human oversight are important to convincing clients about the role of Generative AI. We are very clear with clients about how we will (and will not) use Generative AI and that our people are knowledgeable and well-versed in using Generative AI and all of its caveats.

Norbert: Do you have any practical tips or best practices for managing customer/user expectations for AI research technology solutions?

Greg (That’s what it says): Start with a pilot project and let people play. Be transparent about data security along with trial capabilities. Stay ethically sound and demonstrate the value of the solution.

Alex & Ran (SurveyMind): Free trials, money back guarantee. Don’t lock people into long term contracts with high prices. Show how powerful AI can be, but hold people’s hands. Take the risk off the client.

Tovah (Fathom): AI is not a value – it’s a way to get to value. In other words, AI opens doors and helps save time, energy and money, but to be truly valuable, you need to solve key business challenges for the customer. Always keep that front and center and as an added tangent tip, stop and think about how implementing AI processes can impact business/research operations and how to do it in ways that allow people to be their best and be happy, creative and collaborative.

Karlien (Hello Ara): Be knowledgeable, supportive and transparent. Use knowledgeable human teams to clearly explain the use of AI, ensuring clients understand and trust the process.

Final word on artificial intelligence research technology

If you take anything away from this article, let it be this: AI in research technology works not replace or intend to replace the human element; instead, it enhances the research process.

The key to successful adoption lies in transparency, collaboration and security. Solving customer problems—especially around data security and integration—requires clear communication. As the landscape continues to mature and evolve, remember that AI is not just a tool, but an integral partner in driving innovation and efficiency in the research domain. With support and a commitment to ethical use, organizations can navigate this exciting frontier with confidence, unlocking new potential and setting the stage for future advances in insights.

Looking for the next AI tool to add to your research arsenal? Check out our AI tools for market research and insights:

AI tools for research and insights: The market landscape

This market landscape of AI tools for research and insight represents more than 250 specialized solutions and software tools in 11 different categories.

FIND OUT MORE AI tools for research and insights: The market landscape

Author

Norbert Sari

Norbert (he/she) is an experienced storyteller, strategist and brand designer at Insights Space who leads the brand and marketing shop Rodeo13.

FIND OUT MORE Norbert Sari