How CommodityAI went from spray and pray to stacked signals on lemlist

Testing what works, without burning through funding

CommodityAI builds AI agents for physical commodity companies. They take over the admin between a trade being captured and being settled. Selling to this market is hard to standardize. Chris Bertolini, Founding GTM at CommodityAI, and his team reach many types of companies and roles across commodities and natural resources, and they sell all over the world.

What works for one sector may fail in another. So the team has to test, and test a lot.

We sell in an industry with a variety of different types of companies and roles, within physical commodities and natural resources, and very international as well. So we really need to experiment with what works for some sectors that might not work for others. That means running a ton of experiments that you have to automate as much as possible.

-Chris Bertolini, Founding GTM @CommodityAI

That is where lemlist comes in. Chris says it lets his team find out which channels and messages work for them, quickly and without pain.

With lemlist, you can experiment very fast in a painless way, to see what channels and messaging work for you that might not work for other companies.

Feeling the pain first, then automating it

Chris did not start with a tool. He started by doing outbound by hand on LinkedIn, on purpose.

At first, we just started with manual outbound on LinkedIn. We thought, we’ve got to feel the pain manually first, and then we’ll know what we need to automate.

The next step was a tool built only for LinkedIn. Then Chris found out that lemlist covers LinkedIn and email, and now calling too.

It was the first multichannel sequencer I tried. I looked at a few, but it was the one I heard the best reviews about.

One sequencer for LinkedIn, email and calls

Today the team tests all three channels in the same place. In Chris’s words, you can try LinkedIn, email and the cold call, and see what works for each sector.

Here is what that means in practice:

• •

  Build one multichannel sequence with LinkedIn, email and call steps

• •

  Send a different message to each sector, and compare the results

• •

  Keep the channel and message that work, and drop the rest

Automation keeps each experiment cheap in time and money, which is what a young team needs.

From spray and pray to stacked signals

The experiments led the team to a clear change. Before, outreach was spray and pray: different messages across channels, sent to anyone who fit the ICP. Now Chris starts from signals.

Before, it was more spray and pray. We would try LinkedIn, email and the cold call, but ultimately we only reached out because they were ICP. Now they’re ICP, and they’ve also visited our website, and they’ve engaged with our content on LinkedIn, or they’re connected with our competitors.

Each of these is a signal that a company may be ready to talk. lemlist’s intent signals show when a company is likely ready to buy, so reps can reach out at the right moment. Chris’s team stacks several signals on the same account before reaching out.

Once you stack those signals, you see reply rates skyrocket.

How CommodityAI’s outbound works

Put together, here’s the path of an account at CommodityAI:

→ Start by hand to learn what hurts

→ Move to lemlist and run LinkedIn, email and calls in one sequencer

→ Test messages and channels for each commodity sector

→ Reach out only when an ICP account also shows signals: a website visit, LinkedIn engagement, or a connection to competitors

→ Keep what works, and drop what doesn’t

Asked if he would recommend lemlist today, Chris didn’t hesitate.

Absolutely. Highly recommend lemlist.

Curious to see what your team can achieve?

Try lemlist to run LinkedIn, email and calls from one sequencer, and test what works for your market before you spend your budget.

👉 Start your 14-day free trial