94% of B2B buying groups rank their shortlist before talking to sellers. How robotics vendors can win that silent phase.
Most robotics companies still treat the first sales call as the start of the deal. The research says it is closer to the end of the decision.
In its 2025 B2B Buyer Experience Report, 6sense found that 94% of buying groups put their shortlist in order of preference before engaging with sellers. Buyers made first contact at 61% of the way through their journey, earlier than the 69% the same research recorded before, but still well past the halfway mark. And roughly four in five buyers said their first vendor conversation was with the vendor that eventually won: 77.0% in the main survey (n=3,744) and 81% in a supplemental survey (n=766).
Read that from the side of a robotics vendor or system integrator. By the time an automation engineer fills in your contact form, a ranking already exists. You did not see it being built. You were not in the room. Your content was.
By first contact, the ranking is already done. Source: 6sense, 2025 B2B Buyer Experience Report.
The silent phase is where engineers check your claims
Gartner's buyer research keeps pointing the same way. In a survey of 646 B2B buyers fielded from August through September 2025, 67% said they prefer a rep-free experience. A year earlier the figure from a comparable Gartner survey of 632 buyers was 61%. The same 2024 survey found that 73% of buyers actively avoid suppliers who send irrelevant outreach.
For most B2B categories that means buyers read websites and comparison pages on their own. In robotics it means something more specific, because our claims are physical. A cell video asserts a cycle time. A product page asserts a payload and a reach. A configuration screen in a demo asserts how hard the machine will be to run on a night shift. Every one of those can be checked by the people doing the silent research, and they are exactly the people best qualified to check them.
As I see it, the silent phase is not passive browsing. It is an unsupervised technical review of your marketing. The vendor who survives it gets the first call, and the first call, on the 6sense numbers, is usually with the winner.
Generative AI is now part of the buying committee
The newer shift is who, or what, does that research. Gartner reported in May 2026 that 45% of B2B buyers used generative AI during a recent purchase, primarily to gather information on vendors and products, and that buyers drew on an average of seven information sources. Forrester's 2025 Buyers' Journey Survey went further: generative AI tools were the single most cited meaningful interaction type for researching purchases.
That does not make buyers credulous. In the same Gartner research, 51% said they are more likely to encounter misleading information from generative AI, and 69% said they prefer to validate AI-generated insights with sales reps. Forrester found 20% of respondents felt less confident in a decision because they hit unreliable or inaccurate information. Inaccurate AI output specifically made 19% of all buyers less confident, and 28% of procurement respondents.
My reading: the assistant summarises you before the buyer ever sees you, and it summarises whatever is out there. If your datasheet says one payload, your old trade-fair video implies another and a distributor page lists a third, the model will happily average your inconsistency into something nobody at your company ever claimed. Gartner's 2025 release already flagged the human version of this problem: 69% of buyers reported inconsistencies between information on a supplier's website and what its sellers told them. Robert Blaisdell, VP Analyst in the Gartner Sales Practice, put it plainly: "When sellers provide information that doesn't match the organization's messaging elsewhere, it can create mistrust, potentially putting the transaction at risk."
I'd argue that in robotics this mistrust is expensive, because the claim that drifts is typically a safety-adjacent or performance number that an engineer will hold you to.
The buying group is bigger than your persona deck
The second thing the silent phase hides is how many people are involved. Forrester reports that on average 13 internal stakeholders and nine external participants influence a B2B purchase decision, and that purchases including generative AI features roughly double the size of the buying group.
Edelman and LinkedIn's 2025 B2B Thought Leadership Impact Report, drawing on nearly 2,000 professionals, describes the people your sales team never meets as hidden buyers, and states that they "actively discover, consume, and evaluate thought leadership just like target buyers." The same report cites research from The JOLT Effect that more than 40% of B2B deals stall due to internal misalignment within buying groups.
For a robot cell, the hidden buyers are not abstract. Think of who signs off on automating a line: the automation or process engineer, the plant or operations manager, maintenance, the safety officer, increasingly IT and OT security, procurement and finance, and the operators and shift leads who will live with the HMI. An external integrator, a consultant or a funding body may sit around that table too. Each has a different question, and most robotics content answers only one of them: will it go fast.
One question per role, one asset per question. A working map for robotics buying groups, built on the buying-group findings from Forrester and Edelman with LinkedIn.
Why this matters more in Europe right now
The market context makes the silent phase harder to win. IFR's World Robotics figures, released in September 2026, show more than 600,000 industrial robots installed worldwide in 2025, up 11%, with the global operational stock reaching a record 5 million units. The European Union went the other way: installations contracted 11% to 60,500 units, even as the EU operational stock grew to 712,000 units, which IFR headlines as a record of more than 700,000. Germany, the largest robot market in the EU, installed 24,800 units, 41% of the EU total.
As I read it, a contracting regional market with a growing installed base means more buyers who have done this before, fewer new installations to compete for and more comparison between vendors. In my view, experienced buyers are also faster at spotting a render that cannot be real and a cycle time with no context.
The pre-call evidence kit
Here is the method I would use to design content for a phase you cannot attend. It is not a content calendar. It is a set of assets, each built to answer one buying-group question without a salesperson in the room.
1. One spec source of truth. Pick a single public page per product or cell type where payload, reach, cycle-time conditions, footprint and supported robots are stated once, with the test conditions. Every video caption, carousel and datasheet links to it. When something changes, it changes there first. This is the page you want a generative AI assistant to find.
2. An uncut proof clip for the engineer. One take, timer or part counter in frame, the real part. Label the conditions: part weight, gripper, cycle definition. If the clip and the spec page disagree, fix one before publishing either.
3. A real HMI walkthrough for operators and maintenance. Screen-record the actual interface: job selection, a fault, the recovery. Real latency, real dialogs. It is the easiest asset to skip, and it answers the question the people who run the cell care about most: what happens when it stops.
4. A safety approach explainer for the safety officer. Not a compliance claim. Explain how the cell is risk-assessed, which standards your process follows, and what the customer remains responsible for. Precise language beats reassuring language here.
5. A commissioning and support story for operations and finance. What the project timeline looks like, what the customer must provide, how updates and support work after go-live. No invented ROI figures. If you have a real, approved customer number, use it with its source; otherwise describe the process.
6. A limits page. What your cell does not do: part sizes, materials, cycle times or environments you will not quote. In my view, few things build trust in a silent review faster than a vendor that publishes its own boundaries.
7. A quarterly AI answer audit. Ask two or three generative AI assistants the questions your buyers ask: which vendors do this application, what payload does your cell handle, how is it programmed. Record what they say. Where an answer is wrong, find the public source it probably came from and correct or retire it. This is unglamorous work, and it is now part of brand management.
The AI answer audit as a quarterly loop. It is cheap to run and catches drift before a buyer does.
Social is where the hidden buyers see you first
Where does this content live? My answer: mostly where the hidden buyers already are. I'd expect LinkedIn and YouTube to be where an operations manager or a safety officer meets your proof clip without ever visiting your site, often forwarded by the engineer who found it. That changes how to judge a post. A technical clip with modest reach that the right maintenance lead forwards into an internal chat has done more work than a polished brand film everyone liked and nobody sent on.
It also changes who should appear in the content. If the hidden buyers include operators and maintenance, show the HMI and the recovery, not only the robot at full speed. If they include finance, show the commissioning reality. The format question is not "what performs on LinkedIn". It is "which member of the buying group is this for, and what would they need to see to stop doubting us".
As I see it, much of a robotics purchase is now evaluated like this: alone, on a small screen, without a vendor in the room.
Conclusion
My view is that robotics marketing has been optimised for the wrong moment. We pour budget into the demo, the trade-fair stand and the sales deck, all of which happen after the buyer has ranked us. The ranking itself is built from material we often treat as an afterthought: a spec page nobody owns, a video with no test conditions, an HMI nobody outside the company has ever seen, and whatever a language model has stitched together from all of it.
The fix is not more content. It is evidence designed for an audience you will never meet, consistent enough that a machine summarising it cannot get it wrong, and specific enough that every person in the buying group finds the answer to their own question. Win that silent review and the first call is yours. On the current numbers, that is most of the deal.
Sources
Gartner Sales Survey Finds 67% of B2B Buyers Prefer a Rep-Free Experience, Gartner, 9 March 2026
Forrester: B2B buying and AI in 2026, Digital Commerce 360, 22 January 2026
2025 B2B Thought Leadership Impact Report, Edelman and LinkedIn
Five Million Robots Now Operate in Factories Globally, IFR, 24 September 2026
European Union's Industrial Robot Stock Hits Record 700,000 Units, IFR, 24 September 2026
