Blackbaud Unveils Lantern, a Domain-Specific Model for Fundraising

Blackbaud unveiled Lantern, a domain-specific language model the company describes as the first purpose-built for fundraising intelligence, on September 29, 2026, at its bbcon 2026 customer conference in Columbus, Ohio. Developed through a strategic collaboration with Databricks, the model is planned to begin powering select Blackbaud product capabilities in 2027.

Blackbaud positioned the release against uneven returns from AI across the social impact sector: the company said that while 85% of social impact professionals report using AI in their daily work, only about 10% of organizations are realizing significant dividends on their AI investment. Lantern, the company said, is built on more than four decades of experience in fundraising workflows, decisions and results, so that it understands giving behavior rather than buying behavior.

“Every organization has access to powerful AI, but frontier models learn from the internet, while Lantern learns from philanthropy,” said Mike Gianoni, Blackbaud’s president, CEO and vice chairman. He said more than $100 billion is raised, granted or managed through Blackbaud’s platforms every year, and that Lantern understands fundraising goals, donor dynamics and nonprofit priorities in ways general-purpose AI does not.

Post-Trained With Databricks on a Sector Signal Graph

Working with Databricks, Blackbaud is combining leading open-weight models on the Databricks Data + AI Platform with more than 40 years of social impact intelligence. Lantern is post-trained on synthetic scenarios modeled on Blackbaud’s Social Impact Signal Graph to reflect real-world fundraising patterns, an approach Blackbaud said allows the model to evolve as underlying technology advances.

“There is an enormous opportunity for the philanthropy world to benefit from AI, but seeing returns requires the right data, context and governance,” said Andy Kofoid, Databricks’ president of global field operations. He said the two companies are giving nonprofits the data foundation they need to scale their fundraising.

Blackbaud said Lantern is developed under its Responsible AI principles, with privacy, transparency and sector-appropriate governance at its core.

Designed Around Fundraising, With Customer Input

According to Blackbaud, Lantern reflects the importance of retention, upgrade paths and lifetime relationships, dynamics the company described as central to fundraising and not native to general-purpose AI models. The company said the model can surface opportunities that might otherwise go unseen and help organizations direct their effort where it can have the greatest effect.

Carrie Cobb, Blackbaud’s chief data and AI officer, said Lantern was purpose-built to reflect fundraising, drawing on decades of sector expertise and the intelligence embedded in the Social Impact Signal Graph. She described the result as intelligence that is more relevant, transparent and actionable because it understands the relationships and behaviors that drive outcomes for customers.

Lantern is being built with input from Blackbaud customers including Boston University, Hesed House and YMCA of the North. Tim Cerato, assistant vice president for constituent relationship management at Boston University, said one of the biggest unlocks for AI is its ability to understand an industry’s nuances and amplify the effectiveness of the people doing the work.

The Survey Data Behind the Adoption Figures

The adoption statistics in Blackbaud’s announcement come from the Blackbaud Institute’s Bridging the AI Effectiveness Gap report, which draws on two parallel surveys conducted in March 2026 in the United States by the institute and Edge Research. The surveys covered 1,389 social impact professionals and 1,034 donors who support social impact organizations, a category the report defines as nonprofits, healthcare organizations, K–12 schools, higher education institutions and foundations.

The report classifies about 10% of organizations as AI-Adaptive, meaning they have moved beyond experimentation to systemic AI use supported by governance, data readiness and transparency, while 75% are classified as AI-Emerging and 12% as locked down or unaware. It found that only 50% of organizations are using paid or enterprise versions of AI tools while 24% are using exclusively free versions, and that only about one-third of professionals surveyed believe their organization is using AI very effectively. On disclosure, 76% of donors said it is important for organizations to clearly state when and how AI is used, while only 26% of professionals said their organization does so.

Availability Through Platform for Good

Lantern will be embedded within Blackbaud’s Platform for Good, the unified operating system Blackbaud unveiled the same day at bbcon. That system has three layers: a Data Core combining the Blackbaud Philanthropic Dataset, Blackbaud Institute research and industry datasets, and Blackbaud Giving Search; an Intelligence Layer powered by the Social Impact Signal Graph; and an Action Layer of AI assistants, agents and workflows operating with human oversight. Blackbaud said Lantern sits in the Intelligence Layer, delivering tailored intelligence directly where customers already work.

The same-day announcements included advancements to a Development Agent that identifies and engages unassigned donors, a new Data Health Agent designed to detect and address data issues through review-and-approve workflows, and a rebuilt, native-cloud Raiser’s Edge NXT with a Strategy Assistant that Blackbaud said draws on more than $1 trillion in giving history tracked through the product.

Lantern is planned to begin powering select Blackbaud Raiser’s Edge NXT and Agents for Good capabilities in 2027, with broader availability planned across the Platform for Good.