Reading Dr. Black’s “Data-Driven Finance” framework against our own AI BASECAMP data
Summary: In a recent piece for Filene, Dr. Lamont Black lays out a three-stage shift in how credit unions work with data: from the traditional model of core systems and static reports, to “Artificial Business Intelligence” where staff query data conversationally, to agentic AI where systems act on data continuously inside workflows. Our own AI BASECAMP survey of 123 credit union executives across eight departments shows where most institutions currently sit on that spectrum, and it’s mostly stage one. Finance and Accounting, the department most tied to traditional reporting, has the lowest AI readiness score in the entire survey (2.40 out of 5), and 73.33 percent of Finance respondents pointed to automated reporting as their top AI opportunity. Technology and IT, the department that would need to build the foundation for everything Dr. Black describes, rated data analytics and reporting as the single most important area for AI improvement (100 percent), but also reported the highest rate of anticipated integration challenges (73.33 percent). The shift Dr. Black describes is directional and inevitable. Our data suggests most credit unions are still working out the foundational layer it depends on.
The Shift: Core Systems, ABI, and Agentic AI
Dr. Black’s piece traces a progression that’s worth laying out plainly. For decades, credit union data lived inside core systems and moved outward through scheduled reports. Analysts translated business questions into queries, dashboards got built, and decision-makers waited. That model is being replaced first by what Dr. Black calls Artificial Business Intelligence, or ABI, where staff and leaders interact directly with institutional data through generative AI, asking questions in plain language and getting answers conversationally instead of filing a report request and waiting.
The stage after that is agentic AI, where systems don’t just answer questions but act inside workflows on an ongoing basis. Dr. Black’s examples include agents monitoring loan pipelines and flagging emerging credit risk, adjusting fraud controls in real time, and recognizing a shift in a member’s spending pattern and proactively surfacing a relevant product. The throughline across all three stages is that the foundation matters. None of this works without credit unions first moving away from siloed, core-dependent data toward a modern data architecture (what Dr. Black refers to as warehouses, lakes, and lakehouses) that lets data move and combine across systems.
What AI BASECAMP Found: Most Credit Unions Are Still in Stage One
We surveyed 123 credit union executives across eight departments through AI BASECAMP earlier this year, and the results map onto Dr. Black’s framework more directly than we expected.
Start with Finance and Accounting, which is arguably the department most defined by the traditional reporting model Dr. Black describes as the starting point. It’s also the lowest-scoring department in our entire survey on AI readiness, at 2.40 out of 5, with zero respondents describing their department as fully prepared. When asked which areas of Finance could benefit most from AI, 73.33 percent pointed to automated reporting, more than any other option. In other words, the department most anchored to the old model is also the one most aware that the old model is the problem, and the least equipped, by its own assessment, to move past it.
Technology and IT tells a related but slightly different story. 100 percent of IT respondents identified data analytics and reporting as an area where AI could provide the greatest improvement, the only unanimous response on that question across the entire survey. IT also rated cybersecurity and threat detection at 93.33 percent. These are exactly the kinds of foundational capabilities Dr. Black’s framework depends on. But IT’s own readiness score came in at 3.13, and 73.33 percent of IT respondents said they anticipate integration challenges with existing systems, the highest rate of any department. IT knows where the value is. Whether the underlying systems can get there is a separate question.
Department by Department: Who’s Closer to ABI, Who’s Further
A few other departments are worth placing on Dr. Black’s spectrum.
Lending and Credit sits closer to the agentic end than most. 78.57 percent of Lending respondents pointed to credit scoring and underwriting as a top AI opportunity, and 42.86 percent flagged portfolio management, both of which line up closely with Dr. Black’s example of agents monitoring loan pipelines and flagging risk in real time. Lending’s readiness score came in at 3.07, roughly in the middle of the pack, with 50 percent anticipating integration challenges.
Marketing and Member Experience also points toward the agentic end. 82.35 percent of Marketing respondents identified behavioral segmentation and targeting as a top opportunity, and 70.59 percent pointed to predictive personalization, both close cousins of Dr. Black’s example of an agent recognizing a change in a member’s spending pattern and proactively surfacing a relevant product. Marketing had the highest departmental readiness score in the survey at 3.41.
Operations leans toward the workflow-automation piece of the framework specifically. 100 percent of Operations respondents identified process automation as the area with the greatest potential for AI-driven improvement, more consensus than any other department showed on any single option. Operations readiness came in at 3.20.
The Foundation Question Nobody’s Asking Yet
Here’s where Dr. Black’s piece and our data point to the same underlying issue from different directions. His framework is explicit that ABI and agentic AI depend on a data foundation that doesn’t currently exist at most credit unions, structured, accessible, decoupled from the core. Our survey didn’t ask directly about data architecture, but the department-level answers describe an organization where everyone has identified a use case (Finance wants better reporting, IT wants better analytics, Lending wants underwriting support, Marketing wants personalization) without anyone necessarily having had the conversation about whether those use cases can be supported by a shared, modern data layer, or whether each department is implicitly assuming someone else will build it.
This is consistent with what we saw across the survey more broadly. 61.9 percent of executives cited leadership alignment and understanding as a top barrier to AI adoption, more than technology infrastructure itself. The infrastructure Dr. Black describes is foundational and crosses every department. Building it without first aligning leadership on ownership, sequencing, and shared definitions is how credit unions end up with five departments independently requesting AI tools that all assume access to data nobody has yet organized.
Where AI BASECAMP Fits
AI BASECAMP is built for the conversation that needs to happen before a credit union starts building toward Dr. Black’s vision of ABI and agentic AI. It’s an executive alignment engagement, structured surveys, department-level conversations, and working sessions, aimed at getting a leadership team aligned on what AI means for their institution, who owns what, and how the foundational work gets sequenced across departments that don’t normally coordinate closely.
Dr. Black’s piece describes where the industry is headed. Our data describes where most credit unions are starting from. AI BASECAMP is the bridge between the two.
Frequently Asked Questions
What is “Artificial Business Intelligence” (ABI)? ABI is a term used by Dr. Lamont Black to describe a stage of data maturity where staff and leaders interact directly with institutional data through generative AI, asking questions in plain language and getting answers conversationally, rather than waiting on analysts to build reports and dashboards.
How does this relate to agentic AI? Agentic AI is the stage after ABI in Dr. Black’s framework. Instead of just answering questions, AI systems act continuously inside workflows: monitoring loan pipelines, adjusting fraud controls, or proactively surfacing relevant products to members based on behavior changes, all within defined boundaries and with human oversight retained for judgment calls.
Which credit union departments are furthest along toward this kind of data use, according to AI BASECAMP’s data? Lending and Credit, Marketing and Member Experience, and Operations show the clearest alignment with ABI and agentic use cases, citing priorities like underwriting support, behavioral segmentation, predictive personalization, and process automation. Finance and Accounting and Technology and IT show the most direct ties to the foundational work (reporting and data infrastructure) but also report the lowest readiness and highest anticipated integration challenges.
Is data infrastructure something our credit union needs to solve before doing anything else with AI? Not necessarily in a strict sequence, but Dr. Black’s framework makes clear that ABI and agentic AI both depend on data that can move and combine across systems. Our survey data suggests most credit unions haven’t yet had the cross-departmental conversation about how that foundation gets built and who owns it, which is a planning and alignment gap more than a technology gap.
What is AI BASECAMP? AI BASECAMP is Wide Open Ventures’ executive alignment program for credit union leadership teams. It uses department-level surveys, structured conversations, and working sessions to help leadership reach a shared understanding of AI priorities, ownership, and sequencing, before moving into infrastructure decisions, vendor selection, or implementation.
Where can I read more from Wide Open Ventures on this? Visit our insights page for more on AI adoption, governance, and what credit unions are seeing in practice.
