An apple orchard heavy with ripe fruit. A picking ladder lies in the grass beside an empty basket.

Firms handed out prompt libraries while the tools that do the work went unused.

Large consulting firms gave their people prompt libraries and internal chatbots that few trust. In February 2026 the agentic tools that could do the work went unused because they carried a developer label.

Ripe apples hang on the trees, and the picking ladder lies in the grass.

Tanush Jagdish · February 2026 · First published on LinkedIn

Consultants join top firms to work on hard problems, and many expect to be trained on the best AI tools. What they get is a prompt library, a voluntary training few people attend, an internal chatbot that invents numbers, or a PowerPoint plugin that, in one consultant’s words, “could not exist and my life wouldn’t change.”

I heard the same story at firm after firm: the firm built internal AI tools, and its consultants stopped trusting them.

“Honestly, I don’t understand how it’s so bad.”

a consulting manager, on her firm’s AI slide-maker

In February 2026, tools that do work, such as Claude Code, Codex, and Claude Cowork, carried a developer label. Many firms already pay for access, yet most consultants never open them.

“I feel behind the coders. They know which problems they can use these tools for. I don’t.”

a management consultant

You do not need to code to use them. You brief them the way you would brief an analyst. In one example, Claude Code built a cybersecurity commercial due diligence, with research, a spreadsheet, and slides, in under 15 minutes.

Tanush Jagdish first published this note on LinkedIn. Read the original

More from our research.

Working on the same question?

We look for people to interview, data to compare against, and teams willing to test what we find. Leave a note and we will write back.