Most B2B sales teams in 2026 are now paying for 4-6 AI tools. Their team's output has barely changed. The instinct is to evaluate more tools or change the stack. The actual problem is usually one of three things — none of which is "we need more software."
Sales leaders increasingly come to me with the same story: "We have all the AI tools. The reps barely use them. The ones who do use them produce slightly more output but not the 2x we expected. Should we evaluate something new?"
The answer is almost always no. The problem isn't the stack — it's how the team is set up to use it. Three patterns explain 90% of the disappointment.
Most teams using Claude or ChatGPT do so individually. Each rep has their own chats, their own prompts, their own context. Output quality varies dramatically by rep. There's no organizational learning.
The fix: shared Claude Projects with team-level system prompts, ICP, voice, and example wins loaded. Every rep operates against the same intelligence. Quality goes up; variance goes down. This single fix produces more leverage than any new tool you'd buy.
See what is Claude Projects if you don't know what these are yet.
Adding AI to a job designed around 2020 sales motions produces marginal improvement. The real lift requires redesigning the role — what AEs vs. SDRs do, what managers vs. ICs are accountable for, what activities even still make sense.
Example: most sales orgs still measure SDR activity volume (calls/emails per week). When AI does most of the research and drafting, that metric becomes meaningless. SDRs should be measured on qualified meetings booked, account-quality decisions, and creative outreach quality. Until you change what you measure, AI tools produce nothing measurable.
See how to build an AI-native sales team for the full role redesign.
Sales orgs measure things AI doesn't move directly: activity counts, pipeline value, close rates. AI affects these indirectly — through better-quality activities, more researched accounts, faster proposals.
The leading indicators that AI is actually working: (1) rep prep time per call (should be dropping 60-70%), (2) time-to-first-proposal in the cycle (should be hours, not days), (3) account-research depth per account (should be improving even as time-per-account drops), (4) manager-coaching-brief usage (should be near 100%).
If you're not measuring these, you can't tell whether AI is working. Most sales orgs aren't. The first move isn't buying more software — it's measuring the indicators that show AI's actual impact.
Audit which tools your reps are actually using daily vs. paying for and ignoring. Consolidate.
Set up shared Claude Projects with your team's ICP, voice, and proof points. Make Project usage a daily expectation, not optional.
Redesign the metrics you track to include AI leading indicators. Make them visible. Make managers accountable for them.
Run cohort training (not single workshops) for the workflows that matter.
These four moves cost essentially nothing relative to another AI tool subscription. The lift they produce dwarfs anything you'd get from adding software 7 to the stack.
Want this done for you? Our Implementation engagement is exactly this — at $3,500 fixed.
— Bill Colbert, Treetop Growth Strategy