Category robots

What made Hexagon the right partner was their deep expertise in precision measurement and their willingness to collaborate on adapting their metrology solutions for the demanding conditions of construction sites.

It’s never been faster to build an AI agent — some teams can now do it in weeks. But that speed creates a new problem: performance measurement. Once agents start handling production workloads, how do you prove they’re delivering real business value?

Maybe your agents are fielding customer requests, processing invoices, and routing support tickets wherever they need to go. It may look like your agent workforce is driving ROI, but without the right performance metrics, you’re operating in the dark. 

Measuring AI agent productivity isn’t like measuring traditional software. Agents are nondeterministic, collaborative, and dynamic, and their impact shows up in how they drive outcomes, not how often they run. 

So, your traditional metrics like uptime and response times? They fall short. They capture system efficiency, but not enterprise impact. They won’t tell you if your agents are moving the needle as you scale — whether that’s helping human team members work faster, make better decisions, or spend more time on innovative, high-value work. 

Focusing on outcomes instead of outputs is what turns visibility into trust, which is ultimately the foundation for governance, scalability, and long-term business confidence.

Welcome to the fourth and final post in our Agent Workforce series — a blueprint for agent workforce management and success measurement.

Essential agent performance metrics

Forget the traditional software metrics playbook. Enterprise-ready AI agents need measurements that capture autonomous decision-making and integration with human workflows — defined at deployment to guide every governance and improvement cycle that follows. 

1. Goal accuracy is your primary performance metric. This measures how often agents achieve their intended outcome, not just complete a task (which could be totally inaccurate). For a customer service agent, response speed isn’t enough — resolution quality is the real measure of success. 

Formula: (Successful goal completions / Total goal attempts) × 100

Benchmark at 85%+ for production agents. Anything below 80% signals issues that need immediate attention.

Goal accuracy should be defined before deployment and tracked iteratively across the agent lifecycle to verify that retraining and environmental changes continue to improve (and not degrade) performance.

2. Task adherence measures whether agents follow prescribed workflows. Agents can drift from instructions in unexpected ways, especially when edge cases are in the picture.

Workflow compliance rate, unauthorized action frequency, and scope boundary violations should be factored in here, with a 95%+ adherence score being the target. Agents that consistently fall outside of that boundary ultimately create compliance and security risks.

Deviations aren’t just inefficiencies — they’re governance and compliance signals that should trigger investigation before small drifts become systemic risks. 

3. Hallucination rate measures how often agents generate false or made-up responses. Tracking hallucinations should be integrated into the evaluation datasets used during guardrail testing so that factual reliability is validated continuously, and not reactively.

Formula: (Verified incorrect responses / Total responses requiring factual accuracy) × 100

Keep this below 2% for customer-facing agents to maintain factual reliability and compliance confidence. 

4. Success rate captures end-to-end task completion, while response consistency measures how reliably agents handle identical requests over time, which is a key driver of trust in enterprise workflows. 

These Day 1 metrics establish the foundation for every governance and improvement cycle that follows. 

Building guardrails that make governance measurable

Governance is what makes your data credible. Without it, you measure agent effectiveness in a silo, without accounting for operational or reputational risks that can undermine your agent workforce. 

Governance controls should be built in from Day 1 as part of deployment readiness — not added later as post-production cleanup. When embedded into performance measurement, these controls do more than prevent mistakes; they reduce downtime and accelerate decision-making because every agent operates within tested, approved parameters.

Strong guardrails turn compliance into a source of consistency and trust that give executives confidence that productivity gains from using AI agents are real, repeatable, and secure at scale. 

Here’s what strong governance looks like in practice:

• Monitor PII detection and handling continuously. Track exposure incidents, rule adherence, and response times for fixes. PII detection should enable automatic flagging and containment before issues escalate. Any mishandling should trigger immediate investigation and temporary isolation of the affected agent for review.
• Compliance testing should evolve with every model update. Requirements differ by industry, but the approach is consistent: create evaluation datasets that replay real interactions with known compliance challenges, refreshed regularly as models change. 

For financial services, test fair lending practices. For healthcare, HIPAA compliance. For retail, consumer protection standards. Compliance measurement should be just as automated and continuous as your performance tracking.

• Red-teaming is an ongoing discipline. Regularly try to manipulate agents into unwanted behaviors and measure their resistance (or lack thereof). Track successful manipulation attempts, recovery methods, and detection times/durations to establish a baseline for improvement. 
• Evaluation datasets use recorded, real interactions to replay edge cases in a controlled environment. They create a continuous safety net, allowing you to identify and address risks systematically before they appear in production, not after customers notice. 

Evaluation methods: How to evaluate agent accuracy and ROI

Traditional monitoring captures activity, not value, and that gap can hide risks. It’s not enough to just know agents appear to be working as intended; you need quantitative and qualitative data to prove they deliver tangible business outcomes — and to feed those insights back into continuous improvement. 

Evaluation datasets are the backbone of this system. They create the controlled environment needed to measure accuracy, detect drift, validate guardrails, and continuously retrain agents with real interaction patterns.

Quantitative assessments

• Productivity metrics must balance speed and accuracy. Raw throughput is misleading if agents sacrifice quality for volume or create downstream rework for human teams.

Formula: (Accurate completions × Complexity weight) / Time invested

This approach prevents agents from gaming metrics by prioritizing easy tasks over complex ones and aligns quality expectations with goal accuracy benchmarks set from Day 1.

• 30/60/90-day trend analysis reveals whether agents are learning and improving or regressing over time. 

Track goal accuracy trends, error-pattern evolution, and efficiency improvements across continuous improvement dashboards, making lifecycle progression visible and actionable. Agents that plateau or decline likely need retraining or architectural adjustments.

• Token-based cost tracking provides full visibility into the computational expense of every agent interaction, tying it directly to business value generated.

Formula: Total token costs / Successful goal completions = Cost per successful outcome

This lets enterprises quantify agent efficiency against human equivalents, connecting technical performance to ROI. Benchmark against the fully loaded cost of a human performing the same work, including salary, benefits, training, and management overhead. It’s “cost as performance” in practice, a direct measure of operational ROI.

Qualitative assessments

• Compliance audits catch what numbers miss. Human-led sampling exposes subtle issues that automated scoring overlooks. Run audits weekly, not quarterly as AI systems drift faster than traditional software, and early detection prevents small problems from undermining trust or compliance. 
• Structured coaching adds human judgment where quantitative metrics reach their limit. By reviewing failed or inconsistent interactions, teams can spot hidden gaps in training data and prompt design that automation alone can’t catch. Because agents can incorporate feedback instantly, this becomes a continuous improvement loop — accelerating learning and keeping performance aligned with business goals. 

Building a monitoring and feedback framework

A unified monitoring and feedback framework ties all agent activity to measurable value and continuous improvement. It surfaces what’s working and what needs immediate action, much like a performance review system for digital employees. 

To make sure your monitoring and feedback framework positions human teams to get the most from digital employees, incorporate:

• Anomaly detection for early warning: Essential for managing multiple agents across different use cases. What looks like normal in one context might signal major issues in another. 

Use statistical process control methods that account for the expected variability in agent performance and set alert thresholds based on business impact, not just statistical deviations. 

• Real-time dashboards for unified visibility: Dashboards should surface any anomalies instantly and present both human and AI performance data in a single, unified view. Because agent behavior can shift rapidly with model updates, data drift, or environmental changes, include metrics like accuracy, cost burn rates, compliance alerts, and user satisfaction trends. Ensure insights are intuitive enough for executives and engineers alike to interpret within seconds.
• Automated reporting that speaks to what’s important: Reports should translate technical metrics into business language, connecting agent behavior to outcomes and ROI. 

Highlight business results, cost efficiency trends, compliance posture and actionable recommendations to make the business impact unmistakable. 

• Continuous improvement as a growth loop: Feed the best agent responses back into evaluation datasets to retrain and upskill agents. This creates a self-reinforcing system where strong performance becomes the baseline for future measurement, ensuring progress compounds over time. 
• Combined monitoring between human and AI agents: Hybrid teams perform best when both human and digital workers are measured by complementary standards. A shared monitoring system reinforces accountability and trust at scale. 

How to improve agent performance and AI outcomes

Improvement isn’t episodic. The same metrics that track performance should guide every upskilling cycle, ensuring agents learn continuously and apply new capabilities immediately across all interactions. 

Quick 30–60-day cycles can deliver measurable results while maintaining momentum. Longer improvement cycles risk losing focus and compounding inefficiencies. 

Implement targeted training and upskilling

Agents improve fastest when they learn from their best performances, not just their failures. 

Using successful interactions to create positive reinforcement loops helps models internalize effective behaviors before addressing errors.

A skill-gap analysis identifies where additional training is needed, using the evaluation datasets and performance dashboards established earlier in the lifecycle. This keeps retraining decisions driven by data, rather than instinct. 

To refine training with precision, teams should:

• Review failed interactions systematically to uncover recurring patterns such as specific error types or edge cases, and target those for retraining. 
• Track how error patterns evolve across model updates or new data sources. This shows whether retraining is strengthening performance or introducing new failure modes.
• Focus on concrete underperformance scenarios, and patch any vulnerabilities identified through red-teaming or audits before they impact outcomes. 

Use knowledge bases and automation for support

Reliable information is the foundation of high-performing agents. 

Repository management ensures agents have access to accurate, up-to-date data, preventing outdated content from degrading performance. Knowledge bases also enable AI-powered coaching that provides real-time guidance aligned with KPIs, while automation reduces errors and frees both humans and agents to focus on higher-value work.

Real-time feedback and performance reviews

Live alerts and real-time monitoring stop problems before they escalate. 

Immediate feedback enables instant correction, preventing small deviations from becoming systemic issues. Performance reviews should zero in on targeted, measurable improvements. Since agents can apply updates instantly, frequent human-led and AI-powered reviews strengthen performance and trust across the agent workforce.

This continuous feedback loop reinforces governance and accountability, keeping every improvement aligned with measurable, compliant outcomes.

Governance and ethics: Build trust into measurement 

Governance isn’t just about measurement; it’s how you sustain trust and accountability over time. Without it, fast-moving agents can turn operational gains into compliance risk. The only sustainable approach is embedding governance and ethics directly into how you build, operate, and govern agents from Day 1. 

Compliance as code embeds regulation into daily operations rather than treating it as a separate checkpoint. Integration should begin at deployment so compliance is continuous by design, not retrofitted later as a reactive adjustment.

Data privacy protection should be measured alongside accuracy and efficiency to keep sensitive data from being exposed or misused. Privacy performance belongs within the same dashboards that track quality, cost, and output across every agent. 

Fairness audits extend governance to equity and trust. They verify that agents treat all customer segments consistently and appropriately, preventing bias that can create both compliance exposure and customer dissatisfaction.

Immutable audit trails provide the documentation that turns compliance into confidence. Every agent interaction should be traceable and reviewable. That transparency is what regulators, boards, and customers expect to validate accountability.

When governance is codified rather than bolted on, it’s an advantage, not a constraint. In highly regulated industries, the ability to prove compliance and performance enables faster, safer scaling than competitors who treat governance as an afterthought.

Turning AI insights into business ROI

Once governance and monitoring are in place, the next step is turning insight into impact. The enterprises leading the way in agentic AI are using real-time data to guide decisions before problems surface. Advanced analytics move measurement from reactive reporting to AI-driven recommendations and actions that directly influence business outcomes. 

When measurement becomes intelligence, leaders can forecast staffing needs, rebalance workloads across human and AI agents, and dynamically route tasks to the most capable resource in real time. 

The result: lower cost per action, faster resolution, and tighter alignment between agent performance and business priorities. 

Here are some other tangible examples of measurable ROI:

• 40% faster resolution rates through better agent-customer matching
• 25% higher satisfaction rates through consistent performance and reduced wait times
• 50% reduction in escalation rates and call volume through improved first-contact resolution
• 30% lower operational costs through optimized human-AI collaboration

Ultimately, your metrics should tie directly to financial outcomes, such as bottom line impact, cost savings, and risk reduction traceable to specific improvements. Systematic measurement is what transforms pilot projects into scalable, enterprise-wide agent deployments.

Agentic measurement is your competitive edge

Performance measurement is the operating system for scaling a digital workforce. It gives executives visibility, accountability, and proof — transforming experimental tools into enterprise assets that can be governed, improved, and trusted. Without it, you’re managing an invisible workforce with no clear performance baseline, no improvement loop, and no way to validate ROI.

Enterprises leading in agentic AI:

• Measure both autonomous decisions and collaborative performance.
• Use guardrails that turn monitoring into continuous risk management.
• Track costs and efficiency as rigorously as revenue. 
• Build improvement loops that compound gains over time. 

This discipline separates those who scale confidently from those who stall under complexity and compliance pressure.

Standardizing how agent performance is measured keeps innovation sustainable. The longer organizations delay, the harder it becomes to maintain trust, consistency, and provable business value at scale. Learn how the Agent Workforce Platform unifies measurement, orchestration, and governance across the enterprise.

The post How to measure agent performance: metrics, methods, and ROI appeared first on DataRobot.

A warm hand is enough to drive motion in tiny Salmonella-inspired robots that harness molecular-level dynamic bonding.

A warm hand is enough to drive motion in tiny Salmonella-inspired robots that harness molecular-level dynamic bonding.

Rick Payne and team / Ai is… Banner / Licenced by CC-BY 4.0.

Earlier this year, the HRI Design Retreat brought together experts from academia and industry in the field of design for human-robot interaction (HRI). During the two-day event, which featured hands-on interactive activities, participants explored the future of design for HRI, how this could be shaped, and worked on a roadmap for the next five-ten years.

The retreat was organised by Patrícia Alves-Oliveira and Anastasia Kouvaras Ostrowski, and you can see a short documentary about it below:

Find out more about the retreat here.

As our DriveMod vehicles are approaching an intersection, they send a request to the Cyngn Traffic Broker, a feature of Cyngn Insight that manages traffic coordination through an API.

If you've ever moved into a new home, you know the challenge of packing a moving truck—it's like solving a giant, three-dimensional puzzle. Everything needs to fit just right, and nothing can be left loose or unbalanced, or it risks shifting and breaking in transit.

Atlas, CU Denver's robotic dog, trotted in a crawlspace of the Anythink Nature Library construction site in Thornton last month, lights blinking as it maneuvered through tight, dark passageways. Back at the entrance, university engineering Associate Professor Moatassem Abdallah and seven students watched Atlas's live feed, discussing how its 360° video and data could inform the project's next steps.

Designed to deliver exceptional performance, industry-leading compactness, and advanced high-end technology, built for demanding industrial and harsh-environment applications.

After years of watching glumly from the sidelines as a nimble new start-up – ChatGPT – ate its lunch and soared to record-breaking, worldwide popularity, Google has finally decried ‘enough is enough’ and released a new chatbot that’s literally in a league its own.

Dubbed Gemini 3.0, the new AI definitively dusts its nearest overall competitor – ChatGPT-5.1 – in so many benchmark tests, its as if ChatGPT-5.1 has been relegated to boxing in a pick-up match in a friend’s backyard while Gemini 3.0 shows up for a global, pay-per-view special and finds no competitor is worthy to share the ring with it.

Essentially, Gemini 3.0 finds itself punching down at ChatGPT-5.1 many times over when it comes to overall IQ intelligence, overall world knowledge, overall savvy to run a business, overall ability to work with long documents, images and search results and overall, similar skills (see ‘Under the Hood’ below).

Plus, Gemini 3.0 completely mortifies ChatGPT-5.1 on an especially key metric: The ability to understand and work with buttons, icons and other tools served up by Web sites and apps on a computer screen – a fundamental skill needed for AI agents that are looking to interact with the digital world.

Put another way, Gemini 3.0 is not just much better at working with a computer screen.

It’s night-and-day better.

Meanwhile, Gemini 3.0 and its new powers are also being integrated in apps and tools throughout the Google universe – including Google Workspace Apps – to help ensure that Google users never need to leave the Google universe, ever, to apply AI to virtually any imaginable task.

It’s as if the 800-pound gorilla in the room finally stood-up, beat its chest with its fists and bellowed: Hey, I think you forgot something. I’m the 800-pound gorilla in the room.

And with that, it became heavyweight champion of the AI world.

Short-term, it’s tough to see how ChatGPT recovers, given that ChatGPT’s last major update was released just a few months ago.

Long-term, ChatGPT – and other AI engines – will hopefully be able to lift themselves off the mat, bring themselves back up to fighting weight and give Google another taste of the ring ropes.

In the meantime, for a great, 13-minute video review on all that Gemini 3.0 has to offer, check-out this excellent take by AI expert Alex Finn.

If you’re looking to take a deeper dive from the maker’s point-of-view, Google has released 12-article collection on Gemini 3.0, which offers a number of video demos.

Finally, if you want a bit more on what all the fuss is about, check-out “Key New Features/Enhancements” and “Under the Hood” –- an in-depth look at how Gemini 3.0 lunged ahead of ChatGPT-5.1 on critical benchmark tests — below:

Gemini 3.0 Key New Features/Enhancements

*Apparent Killer Creative Writing Ability: Although results are preliminary, early tests reveal that Gemini 3.0’s creative writing ability is superb. An early test by Writing Secrets, for example, revealed that the AI engine is excellent at creative writing, does about 80% of the heavy lifting for you, build memorable characters and scenes and auto-generates scores of turns-of-phrase that leave creative writers longing, ‘I wish I’d written that.”

*Brings Your Imagination to Life: Offer Gemini 3.0 a few ideas, a sketch or some scribblings/doodlings you made on the back of a napkin and it will auto-generate intelligent narrative, imaging, Web sites, apps – and more for you – on the spot.

*Master Level Analysis of Your Videos: Ask Gemini 3.0 to analyze your advertising video for you and it will come back with a detailed report featuring its view on what works and what doesn’t. Ditto for a video of your tennis, golf or pickle ball game.

*On-Board Memory That Gets to Know You: Like ChatGPT-5.1, Gemini 3.0 saves your chats to distill your likes, dislikes, work-style and similar in an effort to serve-up ever-more-customized responses over time.

*Answers Google Search Queries Using Text, Charts, Graphs, Images, Audio, Animations and/or Video, Where Applicable: This latest version of Gemini strives mightily to leverage all forms of knowledge, no matter where Gemini 3.0 pops-up in the Google universe.

Consequently, expect increasing number of responses when using Google Search in AI mode to feature answers to feature multiple forms of content in an effort to offer-up the most lucid, in-depth response as possible.

*Full Integration of Gemini 3.0 Throughout the Google Universe: This is one of Gemini 3.0’s key advantages: The ability to seamlessly integrate with tools and apps throughout the Google Universe, including Google Workspace and its apps like Google Docs, Gmail, Calendar, Contacts, Chat Sheets as well as NotebookLM, AppSheet, Apps Script. The overarching idea: You never need to leave the Google Universe, no matter what your need.

*Stronger Vibe Coding: One of the great long-term promises of AI is to offer everyday, non-technical users the ability spin-up their own apps by simply having a conversation with AI about what they want – or vibe coding. The feature has not been perfected yet, but Gemini 3.0 promises it has been enhanced with this latest version.

*Enhanced Agent Building (for Google Ultra Subscribers Only): While the promise of error-free AI agents – designed to perform a number of multi-step tasks for you without supervision – is still more of a goal that a shrink-wrapped product, Gemini 3.0 doubles down on meeting that horizon with a new agent builder that works with Gemini 3.0 – Antigravity. Alas, for now, that new capability is available only to big spender Google Ultra subscribers.

Gemini 3.0: Under the Hood

Here’s how Gemini 3.0 stacks-up against its overall closest competition, ChatGPT-5.1 on key, benchmark tests:

*Leagues above when it comes to correctly answering questions based on the onboard knowledge stored in its neural database – rather than going to the Web or another third party for help (SimpleQA Verified test).
–Gemini 3.0: 72.1% accuracy
–ChatGPT 5.1: 34.9% accuracy

This is an extremely worrisome finding if you’re the maker of ChatGPT-5.1. Essentially, Gemini 3.0 is twice as good as offering correct answers to tough questions in a head-to-head competition. Think: You’re out of the game before the other guy even knows you’re there.

*Leagues above when asked to understand – and interact with – what’s on a computer screen (ScreenSpot-Pro Test).
–Gemini 3.0: 72.7%
–ChatGPT-5.1: 3.5%

Computer screen IQ – or the ability to understand what’s on the screen before you and the intuition to know how to work all the buttons and icons to make that Web site or app work for you – is a fundamental measure of how dependable your AI will for you as an AI agent.

ChatGPT-5.1 barely put numbers on the board on this test, while Gemini 3.0 made the right choices nearly three quarters of the time.

*Leagues above when it comes to running a simple business (Vending-Bench 2 Test).
–Gemini 3.0: $5,478.16
–ChatGPT-5.1: $1,473.43

This evaluation tests the fantasy of designing an app to run a business for you without any supervision. In this case, it tests an app given seed money to run a vending machine business for a year and handle every day tasks for that business for price setting, fee paying and adjusting stock based on consumer demand. Again, Gemini 3.0 is untouchable on this benchmark compared to ChatGPT-5.1.

*Leagues above when it comes to solving complex math problems (Math Arena Apex Test):
–Gemini 3.0: 23.4%
–ChatGPT-5.1: 1.0%

Granted, most businesses don’t issue math tests when interviewing job candidates. But all things being equal, you’d probably want the guy flying the prop plane in a hail storm to have the IQ of a math whiz – rather than a guy stumped by measuring cups while baking. Math Arena Apex is considered to be one of the toughest math tests on the plant and until recently, most AI engines hovered near zero on their results.

*Extreme advantage when solving complex puzzles (ARC-AGI-2 Test).
–Gemini 3.0: 31.1%
–Chat-GPT-5.1: 17.6%

For this test, evaluators measure an AI engine’s puzzle-solving acuity by how adept it is as distilling puzzle rules that are embedded in tiny, colored grid puzzles. With this metric, Gemini 3.0 is nearly twice as good as ChatGPT-5.1.

*Significantly better at analyzing and working with long documents, images and search results (Facts Benchmark Suite Test).
–Gemini 3.0: 70.5%
–ChatGPT-5.1: 50.8%

AI engine makers tout their tech’s ability to understand and work with content. What they forget to mention is that their tools do not do this perfectly. Even so, Gemini 3.0’s performance scored significantly higher than ChatGPT-5.1 when measured on this task.

*Significantly better at overall world knowledge (Humanity’s Last Exam).
–Gemini 3.0: 37.5%
–ChatGPT-5.1: 26.5%

After AI engines started repeatedly acing famously tough exams regularly taken by would-be attorneys, doctors and other professionals, AI test-makers created this darkly named, super-hard exam sporting 2,500 tricky questions as the ultimate “show-me-what-you-got” test. Questions draw on expertise across the academic spectrum, including math, science, engineering, humanities and more. Once again, Gemini 3.0 bested ChatGPT-5.1 with substantial breathing room.

*A bit better when it comes to the ability to learn from educational videos (Video-MMMU Test).
–Gemini 3.0: 87.6%
–ChatGPT-5.1: 80.4%

Soon, increasing numbers of people are going to want AI engines to be able to watch educational videos and learn from them. Both AI engines did well on this test, with Gemini 3.0 doing a bit better.

*A bit better when it comes to complex science knowledge (GPQA Diamond Test).
–Gemini 3.0: 91.9%
–ChatGPT-5.1: 88.1%

In this match – which tests knowledge of physics, chemistry and biology with super-hard questions – both AI engines turned in great results, although Gemini 3.0 was a smidge better.

Share a Link:  Please consider sharing a link to https://RobotWritersAI.com from your blog, social media post, publication or emails. More links leading to RobotWritersAI.com helps everyone interested in AI-generated writing.

–Joe Dysart is editor of RobotWritersAI.com and a tech journalist with 20+ years experience. His work has appeared in 150+ publications, including The New York Times and the Financial Times of London.

The post Gemini 3.0: The New Gold Standard in AI appeared first on Robot Writers AI.

Despite aggressive AI investments, 95% of enterprise AI initiatives fail—not because the technology doesn't work, but because the organizational infrastructure beneath it was never designed for how AI actually changes work.

Claire chatted to Kevin McAleer from kevsrobots about how to get started building robots at home.

Kevin McAleer is a hobbyist robotics fanatic who likes to build robots, share videos about them on YouTube and teach people how to do the same. Kev has been building robots since 2019, when he got his first 3d printer and wanted to make more interesting builds. Kev has a degree in Computer Science, and because his day job is relatively hands-off, this hobby allows his creativity to have an outlet. Kev is a huge fan of Python and Micropython for embedded devices, and has a website – kevsrobots.com where you can learn more about how to get started in robotics.

Encoders are sensors that measure physical movement—such as the rotation of a motor—and convert it into electrical signals. This allows us to determine how much a motor has rotated and use that information as digital input in a robot’s control system. They are essential for achieving precise motion control, especially in applications that require accurate […]

Researchers have developed a new algorithm that combines two processes for personalizing robotic prosthetic devices to both optimize the movement of the prosthetic limb and—for the first time—also help a human user's body engage in a more natural walking pattern. The new approach can be used to help restore and maintain various aspects of user movement, with the goal of addressing health challenges associated with an amputation.

In this new era of human-centric automation, robotics and intelligent systems are being designed to elevate the supply chain, creating supply chains that are not only faster and more resilient but also safer, more inclusive, and deeply human.