Research
Go beyond the AI headline.
Agentic commerce, industrial AI and the work between a promising demonstration and a governed business decision. Company cases, deployments, benchmarks and practical implications—with sources, not borrowed promises.
142 research articles
Source-linked research on what has been built, what the evidence supports and what businesses can use. External examples are not OfferSignal client results. Each article keeps its source dates, evidence limits and practical implications.
Page 6 / 12
142 articles · Showing 61–72 · Page 6/12
Gigaton at Mokra: kiln optimisation needs a governed control loop
Gigaton’s Carbon Re case at Heidelberg Materials Mokra shows how plant data, physics-aware modelling and existing automation can improve kiln operations. The reported results are specific and promising, but short-term and not a public independent audit.
Read research
Globality: agentic services sourcing still needs governed award authority
Globality shows how AI-assisted services sourcing can progress from structured briefs and supplier matching to configurable negotiation and award workflows. Its reported customer outcomes are substantial, but the evidence does not establish autonomous downstream execution or a safe substitute for industrial engineering review.
Read research
GSMA Telco AI: live deployments are growing, but network autonomy remains bounded
The GSMA’s Q1 2026 evidence points to a telecom market moving from experimentation toward live network and service applications. The practical lesson is to treat detection, recommendation and customer assistance as distinct from autonomous operational authority.
Read research
Hackett procurement AI: the pilot-to-production gap is the finding
Hackett’s 2026 procurement research shows broad interest in artificial intelligence but much narrower large-scale implementation. The useful lesson for manufacturers and other buyers is to govern one measurable workflow rather than confuse platform access or a pilot with production value.
Read research
HappyRobot: logistics agents need governed actions, not just fluent calls
HappyRobot’s Circle, WWEX and DHL cases show how voice and omnichannel agents can coordinate high-volume freight work when they can look up live records, write scoped updates and escalate exceptions. The outcomes are useful evidence for bounded automation, not proof of autonomous commercial authority.
Read research
hexafarms: yield visibility is useful before it becomes farm autonomy
The Berlin agtech’s sensors, forecasting and crop scouting show how commercial growers can turn production data into better planning. Its customer metrics remain provider-reported, so a bounded crop-and-zone pilot is the right next step.
Read research
Higharc: governed home configuration can shorten the path from plan to permit
Higharc shows how production homebuilders can connect configurable plans, options, pricing and construction documents. Its customer-reported time and conversion gains are useful evidence, but they remain provider-authored claims that require local validation.
Read research
Innovandi Open Challenge: a pitch funnel, not yet a plant result
GCCA’s 2026 cement-industry challenge narrowed 60-plus applications to 20 startups invited to pitch. The useful lesson is how to qualify industrial AI access and evidence before calling a consortium, pilot or commercial outcome real.
Read research
Instalily: faster industrial RFQs still need governed commercial authority
Instalily’s industrial-distribution cases show how cross-system AI can prepare RFQs, sales work and service actions at scale. The useful lesson is where automation stops: current data, explicit permissions and a named person must still govern engineered and binding quotes.
Read research
Kingfisher: useful shopping assistance before autonomous purchase
Kingfisher’s Castorama and B&Q assistants show how conversational advice, product retrieval and recommendations can support home-improvement buying. The evidence also shows why an announced agentic-shopping programme must not be confused with a live cart, checkout or payment capability.
Read research
Kivnon and AI-MATTERS: a credible test route, but no public validation result
AI-MATTERS offers facilities for testing industrial mobile robots, while Kivnon publishes concrete navigation and safety specifications. The retained evidence does not confirm that Kivnon completed an AI-MATTERS experiment or achieved a measured outcome.
Read research
Kiwitron agriculture: machinery safety is promising, farm outcomes remain unproven
Kiwitron has presented configurable optical sensing and partner-led navigation for agricultural machinery. The retained evidence shows a credible safety-assistance concept, but no named farm deployment, independent agricultural validation or quantified farm outcome.
Read researchFrom evidence to work
Bring a real business question.
OfferSignal connects product facts, discovery and governed enquiry work for established manufacturers and B2B/B2B2C businesses. A focused workshop can help your team choose what to test—and what not to automate.
OfferSignalBusiness & AI Workshop