When generative artificial intelligence (genAI) quickly gained traction with the release of ChatGPT in 2022, tech-savvy business users quickly began experimenting. At the time, existing tools were limited to very select use cases and unreliable. With few ready-to-use genAI apps, lack of know-how has been a major barrier to pursuing genAI solutions, especially for more specialized business cases.
The core technology of text-based genAI is large-scale language models (LLMs). The complexity and resources required to create an LLM make his LLM out of reach for most companies. Currently, companies looking to implement genAI use cases can choose from existing LLMs that they can use and customize. Barriers to entry are low and within the reach of almost any mature technology team. For companies without the skills or ambition to work directly with an LLM, software platforms across nearly every business function offer out-of-the-box genAI capabilities that can be used daily with little or no specialized skills. I am.
If 2023 was the year of experimentation with genAI, 2024 is shaping up to be the year of implementing genAI into production solutions serving customers and customer-facing roles. The ability to summarize vast amounts of unstructured data and generate creative content (including text, images, and even video!) is of interest to revenue managers and broader management teams. Forrester’s September 2023 Artificial Intelligence Pulse survey showed that 70% of B2B companies are already using genAI, and another 20% are considering using it.
The common denominator is data quality
Many mistakes can occur between the user’s request, the interpretation of the question, the way the response is generated, and the way the response is returned to the user. No matter what technological path a business pursues, the primary limiting factor it faces today is the quality of its own data. The old adage “garbage in, garbage out” applies even more to genAI. GenAI places an unprecedented burden on data governance capabilities due to the following reasons:
- GenAI consumes data at new levels of speed, scale, and complexity. Data and operations teams that manage traditional business use cases focus on curation and cleansing of defined data sets. GenAI leverages structured and unstructured data alike, giving businesses access to insight generation at unprecedented speed and scale, including data types that most companies don’t actively manage. Masu.
- GenAI uses data to generate unpredictable insights. Measurement and analytics teams are accustomed to controlling the gateways through which end users query available data. Insights are delivered through reports and dashboards, each offering a limited scope and curated experience. GenAI grants access to vast data repositories and leaps intuitively to support user queries. Because data management teams no longer have control over what questions are asked, they can no longer predict what data needs to be cleansed to provide accurate insights.
- Security, privacy, and consent require new processes that don’t currently exist. Managing data security and privacy in traditional business use cases relies on controlling the source data. Data that violates compliance standards is removed, and data security relies on controlling access to specific data sets by authorized users. GenAI models do not rely on active queries of the source data to fulfill requests. Once training data is ingested, data teams can easily lose control over which users are allowed access to which data elements. Security and compliance depend on knowing the appropriate level of access for each end user. The lack of current standards for linking genAI models to source data introduces new levels of uncertainty and risk. The aforementioned AI Pulse Survey identified data privacy and security concerns as the single biggest barrier to genAI adoption by B2B companies.
GenAI’s data challenges require a different approach to data quality
Managing data quality for genAI use cases requires a different set of skills than operations teams. You will also need to retrain your team to manage new concepts. Broadly speaking, this shift in thinking has several key themes.
- Operations teams need to work more closely with technology resources. This partnership of technical skills and business insight is critical to generating reliable responses from genAI tools.
- The role of data stewards must expand their ability to provide domain expertise and take on new roles as arbiters of accurate insight generation.
- Data management must move from cleansing and controlling discrete datasets to curating ongoing, active conversations that are both immediacy and responsive.
register Join Forrester’s B2B Summit North America to learn how to prepare your business for the massive changes genAI is already driving.
This was written by Brett Kahnke, Principal Analyst, and Michelle Goetz, Vice President, Principal Analyst, and originally appeared. here.
follow me twitter Or LinkedIn. check out my website.
