Artificial intelligence is transforming the work of expert network consulting, with both opportunities and challenges arising from the use of large language models (LLMs) alongside other tools. AI has been useful for interview analysis in recent years, such as through semantic webs that help researchers visualize information from transcripts and datasets, but it has also become a buzzword. With the emergence of products that help companies interact with large amounts of data more effectively and efficiently than ever, AI has become more useful and entered a new era. Now, its potential is becoming apparent in virtually every aspect of primary research, from recruiting to data collection and analysis. As AI technologies begin to drive change in the micro-consulting industry, their implications for subject matter experts are becoming clear.
Use AI with Care
The first hint that AI could have a dramatic impact on expert-based research came in the form of rumblings about subject matter experts (SME) using OpenAI’s ChatGPT for help with interviews within just a few weeks of its release. In some ways, ChatGPT itself may have appeared to assume the role of the world’s greatest SME for a time and, within some companies, it appeared to be incredibly useful. However, issues surrounding the security of data and trustworthiness of responses revealed the need for industry to turn to AI products designed specifically for business needs, including an enterprise version of ChatGPT. It seems like AI products are being released daily to help with a multitude of business needs, such as managing customer relationships and conducting user research.
Expert networks are adding AI features to their products to help their clients gather information from SMEs more effectively. This is done primarily through retrieval-augmented generation (RAG), which can produce summaries of interviews and allow researchers to extract information from transcripts through conversations with text content, such as PDFs. And while most AI for enterprise research products don’t have the problems that ChatGPT is most associated with, such as frequent hallucinations and fake citations, most actually still use OpenAI’s language models. This is primarily due to OpenAI’s partnership with Microsoft, the world’s largest enterprise software company. And OpenAI has made it easy for developers to leverage their technology for adding AI features to products. As a result, there is a multitude of AI products available to SMEs, each with its own advantages and problems.
Takeaway 1: Use AI thoughtfully to improve how you work with clients. Competent researchers and interviewers know how to use AI tools effectively to become better prepared for expert engagements. This means that less of your time will be spent answering basic questions in favor of deeper, more specific discussions. Take advantage of AI to improve how you prepare for engagements and assume that expert clients are as well. But be careful about how you use general purpose consumer products like ChatGPT and Bard, which can hallucinate and use your prompts for training.
Make Yourself Visible to AI and Human Recruiters
It seems fitting that the first publicly discussed applications of AI by expert networks were for recruiting SMEs, near the beginning of the research process. AI can help speed up research by identifying potential experts quickly, extracting and analyzing information about them to generate lists of prospects whose expertise is relevant to virtually any research topic. However, while AI can assist researchers in finding some experts, the tradecraft of expert recruiting remains crucial for most expert-based research. Sometimes, the best experts don’t share much about their work with the general public. It may be the case that their insights are “white-labeled” by their clients, who have permission to use and even evangelize their ideas without public attribution. While the specific work they do is usually confidential, the expertise they’ve gained through their work can be invaluable in other contexts. In these cases, it’s necessary for recruiters to network within expert communities and industries to discover hard-to-find prospects and collaborators who can refer them.
Takeaway 2: If you want to be found during the SME recruiting process, which increasingly leverages AI alongside human knowhow, you need to discuss your areas of expertise publicly. While joining multiple networks will help, most projects require cold emails and social media messages inviting participation in projects. Automation has the potential to improve the expert recruiting process by helping clients identify prospects through the web more effectively. As a result, the upside of discussing your expertise publicly and publishing relevant content online has never been higher.
Protect Your Content From Web Crawlers
While some SMEs are overlooked by systems scraping the web to find interviewees, others share their work extensively through books, research reports, blogs, essays, articles, podcasts, and talks. These SMEs are most likely to appear on prospect lists. However, the era of LLMs has also created opportunities to rely less on conventional expert collaborations in favor of engaging with content from experts. This can be useful because interviews with SMEs often reference information that AI can extract from what they’ve previously said or written. For example, if an expert prospect proves to be too elusive to interview, it’s possible to use a chatbot designed for RAG for a virtual conversation with transcripts from talks and interviews they’ve given alongside books and essays they’ve written. However, the limitations to this approach are also clear to most researchers and other stakeholders who need input tailored to their unique questions. They realize that making business decisions based purely on secondary research is too risky in most cases, even when it’s conducted at an unprecedented level of effectiveness through tools that make them feel like they’ve talked to someone.
Takeaway 3: The current era of AI will create opportunities for people to use your work in lieu of talking with you directly. While it’s important to share enough about your work that recruiters can find you, it’s important to consider that AI has the potential to make secondary research through publicly available content much more effective. Whenever you make in-depth information available to the general public, consider measures that prevent AI systems from completely co-opting your work by ingesting it into databases without attribution or compensation. Also, when you make in-depth content available online, include ways for your audience to connect with you. Even in an era when clients can chat with relevant content through RAG-based products, there’s still no substitute for direct communication with SMEs.
Don’t Allow Automation to Devalue Your Expertise
One of the primary threats of the current era of AI is the potential for over-reliance on it, leading to the automation of research in ways that produce erroneous or biased data. It’s essential to find the right balance between automation and human collaboration. For example, the most recent language models have the ability to generate surveys and discussion guides that can seem appropriate for research needs without much guidance or training. However, they can also make mistakes that research professionals know to avoid, such as asking questions in ways that elicit misleading responses or confuse respondents. These are issues that AI product teams can address — at least in part — by designing and training their products to ask questions in better ways, more akin to experienced human researchers. AI will create new opportunities to automate data collection whenever clients seek human input, which will change how information is collected from experts and valued by clients.
Automated data collection probably won’t impact expert-based research in the same way as consumer research because the economics are different. For example, automated interviews are beginning to allow clients to conduct in-depth qualitative research at the scale of survey research. However, micro-consulting engagements typically require much more targeted input, from fewer participants, at much higher incentive rates. Nonetheless, AI will enable new methodologies and business models across research, from user experience to micro-consulting.
Takeaway 4: Pay attention to how automation is changing how clients collect data from you and how they value your contributions. Some expert networks already encourage SMEs to accept relatively low rates. This may become more of a problem for SMEs as automation changes how experts collaborate with clients through expert networks. Be wary of approaches that have the potential to remove you from the value chain, diminish the value of your expertise, or compromise your ability to add value.
Embrace Better Models for Collaboration
AI is already starting to improve research by assisting with expert recruiting, data collection, and analysis. As AI evolves and matures, it can be expected to impact ideation, research design, and reporting as well. However, human researchers and developers of AI for research will need to work together closely to keep the technology aligned with the needs of stakeholders as they address increasingly complex problems. AI has great potential to improve how clients and expert networks collaborate with SMEs. As AI products appear increasingly capable of replacing human researchers — as well as the subject matter experts they work with — both groups will remain vitally important. However, it’s likely that the ways in which they collaborate will also need to evolve to meet the needs of businesses in the current era of AI.
Takeaway 5: Consider emerging opportunities to leverage your expertise as AI changes how clients and expert networks utilize it. While AI products have the potential to accelerate and improve research, they can also create new problems by exploiting human know-how and diminishing its value. However, equitable models will credit and compensate people who’ve consented to their work being used for AI databases. Alongside the increasingly obvious problems, AI will create new opportunities for SMEs to leverage their intellectual property and monetize their expertise.
Here’s an example:
Ferret uses AI to elevate human expertise and enhance the value of collaborations with subject matter experts by
- Accelerating analysis of diverse sources like books, articles, research reports, essays, and transcripts;
- Helping users discover the subject matter experts and authors who’ve contributed to its knowledge base;
- And enabling experts to address users’ specific research needs through interviews and bespoke datasets.
Ferret brings subject matter experts into the AI value chain as stakeholders by ensuring they receive credit and compensation for their contributions. By combining bespoke datasets and expert interviews with information from books, research reports, and other content, Ferret introduces an efficient, integrated approach to research that becomes more invaluable with each use. Ferret offers commissions for referrals to both SMEs and clients, with higher rates (10%) before publicly launching in spring of 2024.
About the Author: Phil Surles is a cultural anthropologist and innovation consultant. He delves into the intersection of culture and capitalism through Culture Capitalist, a newsletter that uses Ferret to accelerate research and content creation.
- How AI is Revolutionizing Expert Networks - December 18, 2023
- Maximizing Expert Network Earnings: Industry Executive Reveals Best Tactics - September 22, 2023
- Content Libraries: Boosting Expert Network Profits at Experts’ Expense - July 18, 2023