Data trends: How companies are using AI today

In our latest Bessemer Community research, one key point of tension captured where most high-growth teams find themselves right now: founders and leaders are convinced operationalizing AI matters, but many are still figuring out how to make it work or the best use cases for their team.

To get a clearer picture on the state of working with AI, we surveyed 173 leaders across 113 portfolio companies on how their teams are adopting AI across six core functions: Tech and Engineering, Finance, People, GTM & Sales, Marketing & Communications, and Customer Success. This study surfaces what's working, where teams are stuck, and what separates the functions operationalizing AI quickly, from the ones still finding their footing.

Eighty-six percent of tech leaders and operators are confident AI will meaningfully change how their teams operate in the next 12 months. Nearly 60% say AI is already core to their operations or actively being deployed across their organization. And yet 43% are still experimenting or just getting started.

One trend stands out before you dive in: Engineering teams are operating in a fundamentally different reality than Finance, HR, GTM, and CS teams. But that gap isn't a failure; it reflects the fact that AI adoption looks different depending on workflows (e.g., ubiquity of code generation), the data environment, and the compliance constraints a team operates under. Every function has to figure out the workflows and use cases for the jobs to be done, given their context and conditions, and the tool stack optimal for their team.

We’re open sourcing this data to help you benchmark where your team stands, and maybe give you peace of mind that while so much has changed, there is yet to be one clear playbook.

Slide deck

Key findings: Teams working with with AI

#1 - Anthropic’s Claude is the dominant tool across functions with 73% of respondents; among Engineering teams, 77% use Claude Code and Cowork with 50% using Cursor.

#2 - AI adoption differs depending on which function you lead. The playbook for engineering doesn't transfer to finance, HR, or CS.

• Engineering is operating in a fundamentally different reality than everyone else. 90% of Tech & Eng teams are deploying or AI-core.
• Finance: 24% actively deploying.
• GTM: 45% actively deploying but 55% still experimenting.
• People: 52% still experimenting or just getting started.
• Marketing: 100% use AI for content but only 13% say it's core to operations

#3 - While “non-technical” leaders are earlier in their AI adoption, it’s not for lack of conviction. Leaders cite different blockers by function:

• Engineering: code quality and review bottlenecks (52%), measuring productivity gains (46%)
• Finance: data quality and system fragmentation (56%), security and compliance (41%)
• GTM: too many tools, no clear winner (27%), data quality and CRM hygiene (41%)
• People: "is it safe to use" — data privacy and compliance (41%)
• Marketing: brand safety and quality control (60%)
• CS: data fragmentation (40%), proving impact on NRR (30%)

#4 - Efficiency gains are already impacting future headcount growth.

• 49% report they are delivering more without adding headcount.
• 25% have upskilled team into AI-adjacent work
• 10% have created new AI workflow roles.
• 13% have slowed or paused hiring.
• 6% have backfilled roles with AI tools.
• And 26% say no material change yet, implying teams have yet to enact changes or measure the impact on the organization.

Resources

At Bessemer, we're committed to helping portfolio teams move from AI curiosity to AI execution. Subscribe to Building AI Differently for practical guides, peer examples, and operator playbooks across every function.

If you're a Bessemer portfolio founder, our community team runs ongoing AI programming designed to help your team go deeper. Reach out to Melissa Bezruchko to learn what's available for your stage and function at mbezruchko@bvp.com.