In This Video
What Data Should You Track In A Sales Pipeline?
As much as you can. The demo board tracks details such as buyer role, industry, country, region, company segment, employee count and annual revenue against every opportunity, because the more data you have, the more accurately you can see which customers spend the most with you.
The aim is to find the group that spends more than any other, such as a particular industry, company size or revenue range, and target them more. Monthly revenue and conversion rates are useful, but knowing who to target is where the real value of sales reporting is.
The board was built to be deliberately confusing and holds test data, not real figures, but boards like it do exist in real businesses.
How Do You Connect ChatGPT To monday.com?
Add monday.com as a plugin in ChatGPT and sign in, which gives ChatGPT access to the data on your boards.
- In ChatGPT, go to Plugins and browse the available plugins.
- Search for monday.com using the search in the top right corner, then click into it.
- Press sign in or connect and go through the sign-in process.
We dictated the prompt. It told ChatGPT which workspace and board to use and asked for a dashboard of reports showing which types of business are likely to spend the most with us, by market size, annual revenue, employee count or industry, to help narrow down our ideal customer profile (ICP) and targeting.
It also asked for an AI Vibe app that filters the board's open, won and lost opportunities to show who to target. We didn't name any specific reports, only the outcome we wanted.
Can ChatGPT Build A Dashboard In monday.com?
Yes. After 23 minutes and 47 seconds on the Extra High setting, ChatGPT came back having built a dashboard, a doc and an AI Vibe app inside monday.com. Ultra might have suited a one-off job like this, but it often takes much longer and its usage is far higher.
The dashboard included win rates by employee band and by revenue band. In the test data, the 1 to 49 and 500 plus employee bands had the highest win rates and the 100 to 249 band the lowest, while the £100 million plus revenue band closed at 81.25%, the highest of any revenue band.
Because the prompt described the outcome rather than the charts, ChatGPT built reports we probably wouldn't have thought to set up ourselves.
Can You Trust The Reports ChatGPT Builds?
Not without checking them. Two of the reports disagreed: one said 67% of opportunities in the 500 plus employee band were won, and another put it at 73.684%.
The sample size charts weren't clear either. They seemed to show the total won, lost and open opportunities in each band, but they're labelled as a sample size, and if they are a sample, that might explain the skewed percentages.
There were still useful pointers. The 1 to 49 employee band had only 45 opportunities but the highest close rate, so it may be under-targeted.
How Do You Use The Reports To Decide Who To Target?
Read them together. In the test data, YouTube organic topped the revenue by acquisition channel report, sales directors had the highest average year one spend and Germany came first by country, which narrowed things to a very specific niche: sales directors in Germany, at companies of a particular size and revenue, found through YouTube.
The AI only knows what's on the board. It can't see the effort or ad spend behind each channel, so a channel you put little into might be giving a better return than one you work hard at. It also wasn't clear whether the channel report showed revenue or bookings.
Other reports covered the overall close rate and open weighted total contract value by industry. A couple were confusing, and ChatGPT can be asked to explain what they show.
How Does The AI Vibe App Find Your Best Customers?
It looks for patterns in the board's past closed deals and turns them into recommendations, starting with a top five. ChatGPT wrote the prompt AI Vibe used to build it, and it opens from the board's drop-down menu as the ICP Growth Intelligence app.
You can filter it by deal type, industry, employee band, revenue band, country and source, and it only recommends segments with more than ten closed and more than five won deals. In the test data it picked out aerospace, and its Segment Explorer listed the open deals to prioritise there.
The loss analysis sorted 67 lost deals by reason, and a lot were no decision: 16 opportunities that never got an answer, which are worth chasing. The app shows the data and reasoning behind each recommendation, and ChatGPT's doc sets out the definitions and company bands the reports use.
Is It Worth Using AI To Report On Your Sales Pipeline?
Yes, as a starting point. There's no value in data you don't use to make decisions, so if your pipeline holds a lot of it, have Claude or ChatGPT build dashboards and an AI Vibe app from it, make an initial assessment and refine from there.
A person still reads the results and decides who to target. A more specific prompt, naming the reports we wanted, might have changed the outcome.