Miss Gullickson: Capacity, funding and coordination hinder turning congressional AI ideas into tools

Miss Gullickson, a witness before the House Administration committee, told lawmakers that the biggest barriers to turning staff AI ideas into long-lived tools are capacity, funding and unclear ownership. "My idea is to create a repeatable innovation cycle," she said, describing a process to evaluate ideas, test them quickly and then route successful pilots to the appropriate technical teams across the legislative branch.

Gullickson said the cycle would include rapid testing, routing items to the right tech team — including teams in other legislative-branch entities such as the Library and GPO — and investing to scale what works. She noted that the CAO has provided a vendor site that can help, but asked who within the House will collect, evaluate and implement staff ideas without adding unnecessary bureaucracy: "I think the biggest barrier from where I'm standing is who's going to do it, who's collecting those ideas and how we're going to do it," she said. She added that capacity, velocity and funding all matter and that the mismatch between congressional requests and the fiscal calendar complicates getting timely resources.

The witness also said funding had been "really flexible" for some projects she has been involved with, but the provided transcript cuts off mid-answer and does not record any committee follow-up or the next procedural steps. The record here does not identify the questioner who addressed Miss Gullickson by name, and the transcript uses the acronym CAO without spelling it out; those details were not specified in the available text.