The 2026 Healthtech Messaging Index
Sternum
sternumiot.com·scored August 23, 2026
25/38
Approaching
Close. The bones are good. A few targeted fixes on the weakest signals would push Sternum into magnetic territory.
The page we scored
The AI Recommendation Check
When we asked AI who to recommend for Embedded IoT security and observability platform, it named:
Buyer we put in the prompt: Product Security Director, CISO, or Embedded Systems Engineering Lead at a medical device or IoT hardware manufacturer. Both the buyer and the category were inferred from Sternum’s homepage alone. The model had no web access.
Sternum was on the list.
AI named them for their own category. That's the signal every homepage is supposed to send.
AI named you when asked who it recommends for "Embedded IoT security and observability platform", and described you accurately: "Sternum provides runtime security and observability for embedded and IoT devices using lightweight on-device instrumentation with minimal performance overhead.". Your positioning is landing with the machines.
Overall assessment
Sternum's biggest strength is its combination of hard, specific numbers (96.5% attack prevention, 60% patch savings, NXP benchmark results) with named, credentialed testimonials from Medtronic and the Linux Foundation ... this is unusually concrete social proof for the category. The biggest gap is the complete absence of competitive acknowledgment and a weak conversion architecture: no primary CTA in the hero, no named alternatives, and no cost-of-inaction narrative forces the buyer to self-motivate rather than feel urgency Sternum created.
Signal by category
Narrative Clarity
8 signals
Trust Signal
4 signals
AI Signal
6 signals
Conversion Signal
1 signal
Weakest signal · fix this first
12Alternatives Acknowledged
“No mention of competitors, alternative approaches, or 'do nothing' scenario anywhere on the page.”
The page never acknowledges any alternative ... no named competitors, no comparison to traditional endpoint agents, no 'do nothing' cost ... leaving the buyer without a competitive frame.
All 19 signals, scored 0 to 2.
Total 25/38
Narrative Clarity
8 signals
- 2
01The 7-Second Test
“Embedded System Protection Against Nation State Attacks ... With just 3% overhead…”
The hero immediately tells you what it is (embedded system protection), who the threat is (nation state attacks), and a key differentiator (3% overhead), passing the 7-second test for a technical buyer.
- 1
02Rebellion / Movement
“Nation State Attacks”
There's an implied enemy (nation-state attackers) but no named industry pattern or status quo being rejected ... no statement like 'traditional endpoint security fails embedded devices' or a named broken approach.
- 2
03Owned Language & Category
“Patented EIV™ embedded integrity verification software technologies profile processes in runtime”
The page owns a specific coined term ... EIV™ (Embedded Integrity Verification) ... that is patented and named, giving AI a citable, ownable label unique to Sternum.
- 1
04ICP Clarity
“Embedded Security and Observability Platform… serves as IDS/IPS for Medical Devices”
Medical device manufacturers and IoT device makers are implied, but company size, buyer role (CISO? Product Security? Engineering?), and verticals beyond medical are never explicitly called out.
- 1
05Problem Leadership
“Embedded System Protection Against Nation State Attacks”
The hero leads with the threat/problem (nation-state attacks on embedded systems) but frames it as a solution category immediately, not as a buyer pain statement in the buyer's own language.
- 2
06Solution Clarity
“Prevent memory & command injection attacks in real-time, Detect & diagnose device & fleet-level anomalies, Cut patch-management costs by 40%”
A visitor can clearly repeat what Sternum does: runtime protection and observability for embedded/IoT devices with specific named capabilities.
- 1
07Cost of Inaction
“Cut patch-management costs by 40%”
Patch cost savings hint at cost of inaction, but the page never explicitly names what happens if you don't act ... breaches, regulatory penalties, field recalls ... leaving the stakes understated.
- 1
08Promised Land
“Make device security and performance tangible for end users… Leverage activity insights to inform future design choices”
There are glimpses of an 'after' state in the Build/Operate/Innovate section, but no vivid, specific promised-land narrative describing what life looks like post-deployment.
Trust Signal
4 signals
- 2
09Proof & Evidence
“96.5% Prevention in benchmark security tests… Up to 60% Savings in patch management costs… 100M+ Events observed every day… 1-3% Overhead”
Multiple specific, concrete numbers are present including benchmark test results, cost savings percentages, and scale metrics ... well above adjective-only claims.
- 2
10Social Proof
“Kyle Erickson, Former Product Security Director at Medtronic… Marc Vauclair, Senior Security System Architect and Fellow at NXP Semiconductors”
The page has six named testimonials with full names, titles, and well-known companies (Medtronic, NXP, Linux Foundation, Telit), which is strong social proof.
- 1
11Authority & Credibility
“Patented EIV™ embedded integrity verification software technologies”
A patent is referenced and prestigious organizations (Linux Foundation, NXP, Medtronic) validate the technology, but there are no founder credentials, original research reports, or named frameworks beyond EIV™.
- 0
12Alternatives Acknowledgedweakest
“No mention of competitors, alternative approaches, or 'do nothing' scenario anywhere on the page.”
The page never acknowledges any alternative ... no named competitors, no comparison to traditional endpoint agents, no 'do nothing' cost ... leaving the buyer without a competitive frame.
AI Signal
6 signals
- 1
13Customer Focus
“A customer was experiencing issues with unexpected battery depletions, leading to substantial financial losses. The team used Sternum…”
The customer stories are told from the customer's perspective, but the bulk of the page (feature lists, platform specs, compatibility) is company/product-centric, making it a mixed split.
- 1
14AI-Parmesan Index
“AI-Powered Threat Detection… AI-powered predictive analytics… AI-powered automatic alerts”
AI is mentioned three times but with no mechanistic explanation of how it works ... it's sprinkled as a label without substance, earning a middle score since the core product (EIV™) is mechanistically explained.
- 2
15LLM Quotability
“All attacks NXP crafted were blocked by Sternum with CPU overhead of less than 3%.”
This sentence from Marc Vauclair at NXP is clean, declarative, specific, and attribution-ready ... an LLM could lift it verbatim as a performance claim.
- 0
16Copyright Freshness
“No visible dates on case studies, blog posts, or copyright year found in the scraped content.”
No recency signals are present in the scraped content ... no publication dates on case studies or blog posts and no copyright year ... which weakens AI recency weighting.
- 2
17Entity Distinctiveness
“Patented EIV™ embedded integrity verification… agentless on-device solution… mitigation of known and zero-day threats… supply chain protection”
The combination of patented EIV™ technology, agentless on-device deployment, nation-state threat framing, and medical device regulatory use cases makes Sternum distinctly identifiable ... an AI could not swap this description onto a generic IoT security competitor.
- 2
18AI Recommendation Check
“AI described you as: "Sternum provides runtime security and observability for embedded and IoT devices using lightweight on-device instrumentation with minimal performance overhead.”
AI named you when asked who it recommends for "Embedded IoT security and observability platform", and described you accurately: "Sternum provides runtime security and observability for embedded and IoT devices using lightweight on-device instrumentation with minimal performance overhead.". Your positioning is landing with the machines.
Conversion Signal
1 signal
- 1
19Path & CTA Clarity
“Learn More… Read the full case study… No numbered steps or clear primary CTA visible in the hero.”
There are multiple 'Learn More' links and a case study CTA, but no numbered process, no single dominant primary CTA in the hero, and no soft secondary CTA paired with it in a connected system.
Keep the lead
AI already names Sternum. The next edition decides if it still does.
Twenty minutes with Greg. Bring the page, bring the argument. You’ll leave knowing exactly which signal is costing Sternum the AI recommendation, and what to write instead.
Not on the list? Run your homepage through the free Brand Signal Score and see where you’d land. Same 19 signals, same AI check, two minutes.