How do you tell curiosity apart from a real AI project?

Curiosity consumes content: likes, webinars, shared articles. A project commits resources: a hire, a role created, a vendor chosen, a budget announced, a pilot on a specific use case. As long as you only see curiosity, offer a useful point of view, not a demo.

Who should you contact at a company preparing an AI project?

At an SME, the owner, who usually decides alone. At a mid-sized company, whoever owns the use case (operations, customer service, finance) matters as much as the CIO or the data lead. Write to the person living the problem AI is meant to solve, not only to the technical owner.

Are likes on AI-related posts a good signal?

Rarely on their own. AI is a topic everyone follows, including your competitors, students and consultants. A like becomes worth acting on when it comes from a decision-maker in your target, repeats around one specific use case, or stacks with another signal such as a hire.

Which providers should be watching these signals?

IT services firms and data consultancies, software vendors building in AI, training organisations, legal and compliance firms, hosting providers and change-management consultants. Each fits a different stage: exploration, pilot or rollout.

AI project buying signals: what to watch for

AI is the topic every company talks about, and that’s exactly what makes prospecting hard. Between the owner who likes a post out of curiosity and the one who has just unlocked a budget for a pilot, the difference doesn’t show up in a contact list. It shows up in the signals.

For an IT services firm, a data consultancy, a software vendor or a training organisation, the challenge is twofold: spot the companies moving from curiosity to project, and write them something more useful than “we support your AI transformation”.

Why does an AI project show up before the tender?

Because it starts with people and words, before it starts with purchases. An owner testing tools talks about it. A company that wants to go further hires someone, or names someone, to own the topic. A team running a pilot looks for feedback from others, often in public.

By the time a tender lands, if it ever does, the use case is already chosen, and the providers who helped frame it have a head start. Many AI projects at SMEs never go through a tender at all: the provider that wins is the one already in the conversation.

Which signals point to an AI project?

Signal	What it points to	Strength
First data or AI hire (data scientist, AI engineer, data analyst)	Intent to build the skill in-house, budget approved	Strong
An AI lead or champion role created, or the mission handed to an existing manager	Someone owns the topic and is looking for solutions	Strong
A leader posting about early trials, what works and what doesn’t	Exploration under way, needs framing	Strong
A partnership announced with a lab, a school or a vendor	Structured project, often at pilot stage	Medium to strong
An AI tool named in a job posting (“familiarity with [assistant] required”)	The tool is already deployed or on its way	Medium
A firm or company adding an AI section to its expertise or its site	New positioning to back up with tools	Medium
Speaking at or attending an AI conference	Active scanning of the market, looking for partners	Moderate
A decision-maker repeatedly commenting on posts about one specific use case (AI in customer service, AI in accounting)	Real concern about that use case	Moderate, strong if repeated
An isolated like on a general AI post	Curiosity	Weak

The last signal is the most common and the least useful. On a topic this popular, you need to qualify harder than usual: role, company, repetition, and a link to a use case you can actually address. The method is covered in how to qualify a buying signal.

Curiosity, pilot or rollout: where is the company, really?

The stage decides the message. Pitching a full rollout to a company still discovering the topic is too early. Offering an introductory session to a company already industrialising is off the mark.

Curiosity. The owner reads, likes, attends webinars, trying to work out what’s possible for their business. Offer a concrete point of view on their trade: which use cases work in their sector, and which disappoint.
Exploration. Someone owns the topic, trials are under way, a first hire has been posted. Offer help choosing the right use case and framing a pilot.
Pilot. A partnership is announced, a use case is being tested. Offer expertise on what makes pilots fail: data quality, team adoption, security.
Rollout. Tools named in job postings, teams being trained. This is the moment for trainers, integrators and add-on vendors.

Who should you write to?

At an SME, almost always the owner. At a mid-sized company, don’t limit yourself to the CIO or the data lead: the person who owns the use case (head of operations, customer service lead, finance director) often carries as much weight, because it’s their team that has to change how it works. To map the roles, see who really makes B2B buying decisions.

What opener should you write?

On AI, prospecting messages all sound the same: “transformation”, “productivity gains”, “tailored support”. Your prospect gets several a week. What stands out is a line that shows you know exactly where it actually breaks down.

Illustrative example

Signal: the owner of a distribution SME posts about early AI trials for handling customer queries, with results described as “promising but patchy”.

Opener that falls flat: “Hi, I saw your post about AI. We support SMEs through their AI transformation, from strategy to rollout. Have you got 30 minutes to talk?”

Opener that gets a reply: “AI trials in customer service often handle simple questions well, then disappoint as soon as they need to lean on the real product sheets, some up to date and some not. It’s rarely the tool that’s the problem, more often the data you feed it. What kind of queries are your tests focused on?”

The second opener doesn’t mention AI in general or your offer. It names a precise problem the owner is probably living through, and ends with an easy question. Another case, built on a hiring signal:

Illustrative example

Signal: a mid-sized industrial company is hiring its first data scientist.

Opener that falls flat: “Congratulations on the new hire! We’re an AI-focused IT services firm and can strengthen your data team.”

Opener that gets a reply: “A first data scientist at an industrial group often spends the early months pulling together data scattered across production, quality and maintenance, before they can model anything. Is the first use case already chosen, or will that be part of their remit?”

How do you turn a reply into a meeting?

A prospect who replies about AI often tells you where they stand, and sometimes what worries them. Pick up on that specific point. Then offer a short, concrete format: 30 minutes to review their use cases, or to look at what stalled in their trials. An exchange that helps them decide is easier to accept than a rundown of your references.

If the company is still at the curiosity stage, don’t push for a meeting. Give them something useful, note the context, and come back when a stronger signal appears: a hire, an announced pilot.

What mistakes should you avoid?

Treating every AI-related like as intent. It’s the most-liked topic around right now. Without a relevant role or repetition, it’s noise.
Selling the technology. The prospect is buying a problem solved, less rekeying, a faster customer service. Not a model.
Ignoring the stage. A rollout demo pitched to a curious owner closes the door.
Forgetting the end user. AI projects often fail at adoption. Talking to the teams involved strengthens your case.

On LinkedIn, MeetMagnet tracks these signals daily and filters out the “everyone likes AI” noise, keeping only decision-makers in your target. For other trigger events of the same kind, see our 12 examples of trigger events.

AI project buying signals: what shows up before the tender