Commoditization of Analyst Workflows

The real threat from a horizontal agent builder is not better analysis, it is cheaper replication of common analyst workflows. Donovan lets defense teams plug in their own data, choose models, create knowledge bases, and deploy agents for planning, reporting, and intelligence tasks, which means parts of a specialist product can turn into configurable features inside a broader AI stack. That shifts differentiation toward proprietary data, mission fit, and deployment inside secure environments.1 2 3 4

• Primer shows what a specialist still looks like. Delta is built to turn messy reports, open source feeds, and internal data into structured, traceable analysis, while Primer Command is tuned for narrative monitoring across news and social media. That is a narrower workflow, but one with deeper defaults for intelligence users.2 5
• Babel Street overlaps from the data side. Its products combine multilingual OSINT collection, entity resolution, relationship mapping, and identity investigation, especially for border security, screening, and law enforcement. In those cases, the hard part is often not the agent shell, it is the underlying collection, matching, and false positive control.3 6 7
• Vannevar is much smaller in capital base than Scale, with an estimated $1.5B valuation and $91.5M total funding versus Scale at an estimated $29B valuation, $1.5B revenue in 2024, and $1.6B total funding. That gap matters because a horizontal platform can subsidize product expansion across many workflows at once.4 8

The market is moving toward a split. Horizontal builders will absorb generic summarization, search, and report drafting, while category leaders keep their edge in hard to copy data pipelines, secure deployment patterns, and analyst workflows shaped around one mission. The winners will look less like model wrappers, and more like data and workflow systems with agents built in.1 2 3 4 6