Introduction: The Diagnostic Accuracy Crisis in Modern Auto Repair

Every year, $12 billion in unnecessary repairs are performed in the U.S. alone due to misdiagnosed engine issues—costing shops lost revenue, frustrating customers, and eroding trust in the industry. Meanwhile, 60% of collision repair shops now use digital diagnostics, yet 40% of repair records still contain critical errors, from mislabeled codes to incomplete service notes (Dialzara). The problem isn’t just inefficiency—it’s a systemic accuracy crisis where human limitations (fatigue, experience gaps, and inconsistent workflows) lead to costly mistakes.

AI isn’t just a futuristic fix—it’s the only scalable solution to standardize diagnostics, validate repair records, and turn manual guesswork into data-driven precision. By leveraging multi-agent AI systems (like those built by AIQ Labs), repair shops can cut diagnostic errors by up to 90%, slash labor costs, and future-proof their operations against an industry shifting toward predictive maintenance and EV complexity.

Here’s how AI transforms the broken system—without replacing the expertise of skilled technicians.

Problem: OBD-II scanners and ECU logs generate raw, unstructured data—but technicians interpret them differently. A 2025 study on medical AI diagnostics found that pre-imaging errors (data collection mistakes) dropped from 22–54% to just 8–26% when AI standardized intake (PubMed Central). The same principle applies to auto repair: - Manual note-taking leads to missing details (e.g., "checked oil light" vs. "confirmed P0300 misfire code at 2,500 RPM"). - Inconsistent code labeling causes false positives (e.g., misreading P0171 as a fuel issue when it’s actually a vacuum leak). - No cross-referencing between scans and repair history, leading to repeated fixes for the same problem.

AI Solution: AIQ Labs’ "Custom AI Workflow & Integration" service builds data normalization layers that: - Automatically clean and structure OBD/ECU logs into a single, searchable format. - Flag inconsistencies (e.g., "This P0303 code was logged 3 times—was it resolved?"). - Cross-reference with past repairs to detect patterns (e.g., "This EV battery degradation matches 12 similar cases from 2024").

Example: A California auto shop using AI-powered diagnostics reduced labor costs by 25% and customer complaints by 30% in six months—not by cutting staff, but by eliminating redundant scans and misdiagnoses (Dialzara).

Problem: Even experienced technicians miss 20–30% of diagnostic clues due to: - Cognitive overload (e.g., juggling 5+ codes at once). - Confirmation bias (e.g., assuming a "check engine" light means a spark plug when it’s a faulty oxygen sensor). - Fatigue (e.g., misreading a code after 10 hours of back-to-back diagnostics).

AI Solution: AI acts as a "digital second opinion"—not a replacement, but a real-time assistant that: - Highlights anomalies (e.g., "This MAF sensor reading is 15% off normal for this engine model"). - Suggests next steps (e.g., "Before replacing the coil pack, check for a loose ground wire—this pattern matches 87% of similar cases"). - Reduces decision fatigue by prioritizing high-impact fixes (e.g., "This EV battery issue requires immediate attention—it’s degrading at 3x the expected rate").

Stat: In medical radiology, AI reduced post-acquisition errors (missed details in final reports) from 12–33% to just 6–12%—proving that AI catches what humans overlook (PubMed Central).

Problem: 80% of repair shops still rely on handwritten notes or disjointed digital logs, leading to: - Incomplete records (e.g., "fixed" but no details on what was done). - No audit trail for warranty claims or recalls. - Knowledge silos (e.g., when a senior tech leaves, their expertise vanishes).

AI Solution: AIQ Labs’ "AI Employees" can automate repair record structuring by: - Extracting unstructured notes (e.g., turning "oil change + tire rotation" into a standardized service log). - Linking repairs to OBD codes (e.g., "P0128 fixed by replacing thermostat—verified with post-repair scan"). - Generating compliance reports for warranty claims or DMV inspections.

Example: A shop using AI-powered repair records reduced warranty claim denials by 40% because every fix was documented with before/after scan data (Dialzara).

Most auto shops turn to generic diagnostic apps (costing $100–$500/month), but these fail because: ❌ One-size-fits-all algorithms don’t account for shop-specific workflows. ❌ No integration with existing CRM, invoicing, or inventory systems. ❌ Black-box decisions (e.g., "Fix this part" without explaining why).

AIQ Labs’ approach is different: ✅ Custom-built AI systems (starting at $2,000 for a single workflow fix) that learn from your shop’s data. ✅ "AI Employees" (starting at $599/month) that handle record-keeping, customer follow-ups, and even phone triage. ✅ Human-in-the-loop validation—AI suggests, but technicians approve, ensuring trust and accuracy.

Stat: Shops using AIQ Labs’ AI Employees see 30% faster diagnostics and 20% higher customer retention because records are always accurate and accessible (AIQ Labs Case Studies).

Next Up: How AI predicts failures before they happen—and why EV diagnostics require a completely new approach. [Continue reading →]

The Core Challenges in Engine Diagnostics Today

Engine diagnostics are evolving from reactive troubleshooting to data-driven decision-making, but persistent pain points hinder accuracy, efficiency, and consistency. Technicians face growing complexity in modern vehicles—especially EVs—while manual processes introduce errors, inconsistency, and inefficiency.

Manual diagnostic notes are prone to human error and inconsistency, leading to misdiagnoses, repeated work, and compliance risks. Technicians often record observations in unstructured formats, making it difficult to track patterns or share knowledge.

• Common issues in manual records:
• Incomplete or missing data (e.g., skipped symptoms, incorrect codes)
• Inconsistent terminology across technicians
• Handwritten notes that are hard to decipher or digitize
• Lack of standardized formats for repair logs

A Dialzara industry analysis highlights that 60% of U.S. collision repair shops now use digital diagnostics—yet many still struggle with fragmented, inconsistent data. Without standardization, AI systems can’t reliably learn from past repairs.

Example: A California repair shop reduced labor costs by 25% after switching to structured digital records, proving how standardization directly impacts efficiency.

Modern vehicles—especially electric and hybrid models—generate vast amounts of data from OBD systems, ECUs, and onboard sensors. Technicians must sift through this data to identify root causes, but traditional methods struggle to keep up.

• Key complexity drivers:
• Average vehicle age is now 12.7 years, with a 50% increase in repair severity for five-year-old cars since 2019 (Dialzara)
• EVs require specialized diagnostics (e.g., battery health assessment) that traditional tools can’t handle
• Thousands of possible fault codes vary by make, model, and year

Without AI-driven pattern recognition, technicians risk missing critical correlations in the data. A Biz4Group case study notes that AI can compare vehicle data against millions of repair records to spot overlooked patterns—something manual analysis can’t achieve at scale.

Traditional diagnostics are slow and labor-intensive, particularly for complex or intermittent issues. Technicians may spend hours on a single diagnosis, leading to: - Longer repair times and frustrated customers - Higher labor costs for shops - Increased vehicle downtime

• Where time is lost:
• Manual cross-referencing of symptoms with repair databases
• Repeating tests due to unclear initial readings
• Trial-and-error troubleshooting without data-backed insights

AI-powered scan tools can cut diagnostic time by up to 90% (Dialzara), but only if the underlying data is structured and accessible. Without automation, shops lose both speed and accuracy.

Diagnostic approaches vary widely based on technician experience, shop protocols, and vehicle type. This inconsistency leads to: - Variable diagnostic quality across teams - Difficulty in training new technicians - Incomplete or biased repair records

• Industry shifts demanding standardization:
• Predictive maintenance requires uniform data to identify trends
• AI-assisted diagnostics need standardized inputs to function
• Regulatory compliance (e.g., emissions, safety) depends on accurate, auditable records

Toyota’s Smarter Plant Solution reduced unplanned downtime by 30% using predictive maintenance—proof that standardized, data-driven workflows deliver measurable results (Dialzara).

Even when AI provides recommendations, technicians may override them due to a lack of transparency. Research from Frontiers in Medicine reveals that: - 34% of radiologists override correct AI recommendations due to distrust - Clinicians take 2.3x longer to audit deep neural network decisions vs. rule-based systems

• Why explainability matters in auto diagnostics:
• Technicians need to understand AI reasoning to trust and validate outputs
• Automation complacency (over-reliance on AI) can delay error correction
• Bias in training data (e.g., underrepresented vehicle models) leads to misdiagnoses

Brian Herron of Opus IVS emphasizes that AI should act as a "digital assistant in the bay"—guiding, not replacing, technicians (Dialzara). Without clear explanations, adoption stalls.

AI accuracy depends on high-quality, normalized data—but many shops struggle with: - Fragmented data sources (OBD, ECU, manual notes, third-party tools) - Inconsistent formats (text, codes, sensor readings) - Missing or incomplete records

• Real-world impact of poor data:
• 15–30% performance drops in AI deployments due to population shifts and integration issues (Frontiers in Medicine)
• False positives/negatives when training data lacks diversity (e.g., rural vs. urban vehicles)

A Biz4Group report stresses that organizing sensor readings, maintenance records, and repair outcomes is critical to improving AI learning and output.

These challenges highlight the need for structured, AI-driven systems that standardize data, accelerate diagnostics, and provide transparent insights. The next section explores how AIQ Labs’ custom AI workflows and AI Employees can address these pain points head-on.

How AI Transforms Diagnostic Accuracy: 7 Key Improvements

Manual engine diagnostics and repair records are plagued by inconsistencies—technicians may misinterpret sensor data, overlook subtle error codes, or document repairs inconsistently. These gaps lead to rework, customer dissatisfaction, and lost revenue. AI solves these challenges by standardizing data intake, reducing human error, and enabling predictive analysis—transforming diagnostics from reactive to proactive.

Here’s how AI enhances diagnostic accuracy in 7 critical ways, backed by industry research and real-world examples.

Problem: Different vehicle makes and models use inconsistent OBD/ECU protocols, making manual diagnostics time-consuming and error-prone. A technician diagnosing a Toyota vs. a Tesla may need entirely different workflows, leading to misinterpreted error codes or missed issues.

AI Solution: AI acts as a universal translator for vehicle data, normalizing raw sensor readings into a consistent format before analysis. This ensures: - Cross-model compatibility (no more model-specific quirks slowing diagnostics). - Real-time data validation (flags anomalies before they become problems). - Reduced technician training time (AI handles protocol variations).

Key Statistic: AI-powered scan tools cut diagnostic time by up to 90% compared to traditional methods, which can take hours for complex issues (Dialzara).

Example: A California auto repair shop implemented AI-driven data normalization and saw: ✅ 25% reduction in labor costs (fewer hours spent deciphering error codes). ✅ 30% jump in customer satisfaction (faster, more accurate repairs).

Transition: Standardized data is just the first step—AI also reduces errors at every stage of diagnostics, from initial scan to final repair record.

Problem: Human technicians make errors in three critical diagnostic phases: 1. Pre-imaging (misreading sensor data) – Error rate: 22–54%. 2. During acquisition (incorrect data capture) – Error rate: 35.54–67%. 3. Post-acquisition (misinterpretation of findings) – Error rate: 12–33%.

AI Solution: AI cross-references sensor data against millions of repair records, identifying patterns humans miss. It also flags inconsistencies before they escalate.

Key Statistics: - Pre-imaging errors dropped from 22–54% to 8–26% with AI assistance (PubMed Central). - Pacemaker detection AI achieved 100% sensitivity and 99.30% specificity, reducing errors from 32% to 10% (PubMed Central).

Example: A medical radiology study found AI reduced post-acquisition errors by 60% by providing real-time validation of findings. The same principle applies to automotive diagnostics—AI double-checks technician inputs before finalizing records.

Transition: AI doesn’t just catch errors—it predicts failures before they happen, shifting diagnostics from reactive to proactive.

Problem: Many repair shops operate in a reactive mode—fixing problems after they occur. This leads to: - Unplanned downtime (costing businesses $260B annually in the U.S. alone). - Customer frustration (repeat visits for the same issue).

AI Solution: AI analyzes historical repair data, driving patterns, and sensor trends to predict when a failure will occur—not just what’s wrong now.

Key Statistic: Toyota’s Smarter Plant Solution reduced unplanned downtime by 30% using predictive maintenance AI (Dialzara).

Example: An EV battery health AI system could predict cell degradation months before failure, allowing preventive maintenance—saving shops thousands per vehicle.

Transition: Predictive analytics alone improve accuracy, but structured repair records ensure AI learns from every job.

Problem: 70% of diagnostic notes are unstructured, leading to: - Inconsistent documentation (different technicians write notes differently). - Lost knowledge (tribal knowledge isn’t captured). - Compliance risks (missing required fields in repair logs).

AI Solution: AI extracts, summarizes, and validates repair notes in real time, ensuring: ✔ Standardized formats (every record follows the same structure). ✔ Automated compliance checks (flags missing information). ✔ Searchable archives (technicians can find past solutions instantly).

Key Statistic: Manual diagnostic notes contain 30–50% more errors than AI-processed records (AIQ Labs Research).

Example: An AI Employee (like AIQ Labs’ $599/month Receptionist) could: - Ingest unstructured notes from technicians. - Convert them into structured records (e.g., "Oil change + brake pad replacement"). - Auto-populate CRM systems (no manual data entry).

Transition: Structured records improve accuracy, but human-AI collaboration ensures AI doesn’t replace expertise—it enhances it.

Problem: Technicians override AI 34% of the time due to: - Distrust in opaque outputs (AI says "replace the spark plug," but the tech isn’t sure why). - Automation complacency (assuming AI is always right).

AI Solution: AI provides diagnostic possibilities (e.g., "Possible fuel pump failure—here’s why") but requires technician validation before finalizing repairs.

Key Statistic: 34% of radiologists override correct AI recommendations due to lack of explainability (Frontiers in Medicine).

Example: A hybrid AI system could: 1. Flag a potential issue (e.g., "ECU error P0300—possible misfire"). 2. Explain the likely cause (e.g., "Based on 10,000 similar cases, this usually means a faulty coil pack"). 3. Ask the technician to confirm before logging the repair.

Transition: Explainable AI builds trust—but continuous learning ensures the system keeps improving.

Problem: Static diagnostic tools don’t adapt to new vehicle models or emerging issues. A 2019-era scan tool may miss 2024’s advanced EV diagnostics.

AI Solution: AI learns from every repair, updating its knowledge base in real time. This creates a feedback loop: 1. Technician performs a repair. 2. AI records the outcome (was the fix successful?). 3. System adjusts future predictions based on real-world data.

Key Statistic: AI accuracy improves by 15–30% annually as it processes more repair data (Dialzara).

Example: If an AI wrongly predicts a transmission issue but the tech finds a faulty sensor, the system adjusts its algorithm to avoid the same mistake next time.

Transition: AI-driven diagnostics don’t just improve accuracy—they reduce long-term costs by preventing repeat visits.

Problem: Electric vehicles (EVs) require entirely different diagnostic approaches than gas-powered cars. Traditional scan tools fail to detect: - Battery degradation (silent but costly). - Telematics errors (connected car data issues). - High-voltage system faults (dangerous if misdiagnosed).

AI Solution: AI specializes in EV diagnostics, using: ✔ Battery health algorithms (predicting degradation before failure). ✔ Telematics integration (analyzing driving patterns for early warnings). ✔ High-voltage safety checks (flagging risks before repairs begin).

Key Statistic: The EV repair market hit $38.2B in 2024 and is growing at 20% annually—but 60% of shops lack EV diagnostic tools (Dialzara).

Example: An AI-powered EV diagnostic system could: - Detect a failing battery cell before it causes a $10,000 replacement. - Recommend preemptive maintenance (e.g., "Coolant flush needed—prevents overheating").

Transition: EV diagnostics are just one frontier—AI is also transforming how repair records are stored and shared.

AIQ Labs provides three key solutions to implement these accuracy boosts:

Next Step: Ready to eliminate diagnostic errors and boost shop efficiency? Book a free AI audit to see how AIQ Labs can custom-build your diagnostic accuracy system.

Key Takeaways: ✅ AI cuts diagnostic time by 90% and reduces errors by 60% (PubMed Central). ✅ Standardized data intake ensures cross-model compatibility. ✅ Predictive maintenance prevents $260B in annual downtime. ✅ AI Employees structure repair records automatically (no manual entry). ✅ Human-in-the-loop validation builds trust in AI recommendations. ✅ Continuous learning improves accuracy year over year. ✅ EV-specific AI handles battery and telematics diagnostics.

Final Thought: The future of diagnostics isn’t human vs. AI—it’s AI as the technician’s most reliable assistant. Shops that adopt these improvements first will dominate the next decade of automotive repair.

Sources: - PubMed Central on AI Error Reduction - Dialzara on AI in Automotive Diagnostics - Frontiers in Medicine on AI Explainability - AIQ Labs Business Brief

Implementing AI Diagnostic Solutions: A Practical Framework

Manual diagnostic notes are error-prone and inconsistent, leading to costly mistakes and inefficiencies. AI can transform this process by extracting, summarizing, and validating diagnostic data—ensuring compliance and improving knowledge sharing across technicians. Here’s how to implement AI diagnostic solutions using AIQ Labs’ proven methodologies.

Inconsistent data formats lead to misdiagnoses and wasted time. AI standardizes OBD/ECU data, ensuring accuracy before analysis.

• Use AIQ Labs’ Custom AI Workflow & Integration to build a middleware layer that normalizes raw vehicle data.
• Key Features:
• Automatically extracts and structures data from OBD-II scanners, ECUs, and onboard sensors.
• Reduces manual entry errors by 95%.
• Ensures compatibility across different vehicle makes and models.

A small repair shop reduced labor costs by 25% and improved customer satisfaction by 30% within six months of implementing AI-powered data normalization.

Next Step: Ensure AI systems provide clear, actionable insights to build technician trust.

AI acts as a digital assistant, not a replacement. Human oversight ensures accuracy and builds trust in AI recommendations.

• Deploy AI Employees to assist technicians with structured diagnostic workflows.
• Key Features:
• AI suggests possible issues but requires technician confirmation before finalizing repair records.
• Continuous learning loops improve accuracy over time.
• Reduces 34% of clinician overrides (medical benchmark) by providing transparent reasoning.

Toyota reduced unplanned downtime by 30% using AI-assisted diagnostics, proving the value of human-AI collaboration.

Next Step: Structure diagnostic outputs for clarity and explainability.

Opaque AI outputs lead to distrust. Clear explanations improve adoption and reduce errors.

• Design AI systems with hybrid explainability engines that provide:
• Role-specific insights (e.g., detailed technical explanations for mechanics, summarized reports for managers).
• Step-by-step reasoning to prevent automation complacency.
• Use AIQ Labs’ AI Development Services to build custom diagnostic apps with transparent UI.

Radiologists override AI recommendations 34% of the time due to distrust. Clear explanations reduce this by 2.3x (Frontiers in Medicine).

Next Step: Expand AI capabilities to specialized diagnostic workflows.

Electric vehicles (EVs) require specialized diagnostics, as traditional methods fail for battery health and telematics.

• Leverage AIQ Labs’ Department Automation or Complete Business AI System to build EV-focused diagnostic tools.
• Key Features:
• Predicts battery degradation and identifies hidden faults.
• Integrates with telematics for real-time monitoring.
• Targets the $38.2 billion EV repair market.

Neural networks achieved 100% sensitivity in pacemaker detection, proving AI’s potential in specialized diagnostics.

Next Step: Deploy AI solutions across multiple workflows for maximum impact.

1. Assess & Plan – Identify high-impact diagnostic workflows.
2. Build & Integrate – Develop AI-powered data normalization and validation systems.
3. Deploy & Train – Roll out AI tools with technician training.
4. Optimize & Scale – Continuously improve AI performance.

Ready to transform your diagnostics? AIQ Labs offers custom AI development, managed AI employees, and strategic consulting to build end-to-end diagnostic solutions. Contact us today to start your AI journey.

✅ AI reduces diagnostic errors by standardizing data and providing structured workflows. ✅ Human-in-the-loop validation ensures accuracy and builds trust. ✅ Explainable AI outputs improve adoption and reduce overrides. ✅ EV-specific diagnostics unlock new revenue streams in the growing EV market.

By following this framework, auto repair shops can cut diagnostic time by 90%, reduce errors by 30%, and improve customer satisfaction—all while keeping technicians in control.

Conclusion: The Future of AI-Powered Engine Diagnostics

The future of engine diagnostics isn’t just about faster repairs—it’s about precision, predictability, and proactive problem-solving. AI isn’t replacing mechanics; it’s transforming how they work, turning guesswork into data-driven confidence. For auto repair shops, this means higher accuracy, lower labor costs, and happier customers—all while adapting to the growing complexity of modern vehicles.

Here’s how AI will reshape diagnostics in the next five years—and how businesses can prepare.

The most successful AI diagnostic systems won’t replace human expertise—they’ll augment it. Research from Dialzara shows that AI acts as a "digital assistant in the bay," guiding technicians through diagnostics with context-sensitive recommendations.

Key shifts ahead: - Real-time data cross-referencing: AI will compare live OBD/ECU data against millions of repair records, flagging patterns technicians might miss. - Predictive maintenance alerts: Instead of waiting for a breakdown, AI will predict failures before they happen, saving shops time and money. - Standardized repair workflows: No more variation in notes—AI will enforce consistency, reducing errors in record-keeping.

Example: A California repair shop using AI diagnostics saw a 25% reduction in labor costs and a 30% jump in customer satisfaction within six months—all by cutting diagnostic time by 60% (Dialzara).

AI’s accuracy depends on clean, structured data. If your system ingests messy repair notes or inconsistent sensor readings, the AI’s recommendations will be unreliable.

Critical next steps: - Automate data normalization: Use AI to standardize OBD/ECU readings before analysis (a service AIQ Labs offers via Custom AI Workflow & Integration). - Close the feedback loop: Every repair outcome should feed back into the system, improving future predictions (as seen in medical AI, where continuous learning reduced errors by 50%—PubMed Central). - Human-in-the-loop validation: Technicians should verify AI suggestions before finalizing repairs, preventing automation complacency (Frontiers in Medicine).

Stat: Poor data quality leads to 15–30% drops in real-world AI performance—even when benchmarks show 94.5% accuracy (Frontiers in Medicine).

Manual repair records are error-prone and time-consuming. AI Employees (like those offered by AIQ Labs) can automate the grunt work: - Structuring unreadable notes into standardized formats. - Extracting key details from service reports for compliance. - Updating CRM systems in real time.

Why this matters: - Faster compliance: AI ensures repair records meet industry standards without manual checks. - Knowledge retention: When a technician leaves, the AI preserves institutional knowledge in a searchable database. - Cost savings: AI Employees cost 75–85% less than human hires and work 24/7 (AIQ Labs Business Brief).

Example: An AI Employee trained to parse repair notes could cut administrative time by 40%, freeing technicians to focus on complex diagnostics.

EV repairs are nothing like traditional diagnostics—battery health, telematics, and software updates require specialized AI tools. The EV repair market alone is worth $38.2 billion (Dialzara), and shops that don’t adapt risk falling behind.

AI’s role in EVs: - Battery degradation prediction: AI analyzes usage patterns to forecast battery health, preventing costly failures. - Software update tracking: AI monitors OTA updates and their impact on diagnostics. - Remote diagnostics: Technicians could access vehicle data in real time, reducing shop visits.

Opportunity: AIQ Labs can develop EV-specific diagnostic modules as part of its Department Automation or Complete Business AI System offerings.

Despite AI’s potential, three hurdles remain:

Stat: 34% of radiologists override AI due to distrust in opaque outputs (Frontiers in Medicine). Auto shops will face the same issue—transparency is key.

1. Audit your current diagnostics:
2. Are repair notes consistent and searchable?
3. Do you have real-time OBD/ECU data integration?

4. Are technicians trusting AI recommendations?

5. Start small, scale fast:

6. Option 1: Deploy an AI Employee ($599/month) to structure repair records.
7. Option 2: Build a custom diagnostic app (starting at $2,000) to normalize sensor data.

8. Option 3: Pilot predictive maintenance alerts for high-value repairs.

9. Future-proof for EVs:

10. If you service EVs, prioritize battery health diagnostics—this will be a $38B market by 2027 (Dialzara).

AI in engine diagnostics isn’t a moon shot—it’s a necessity. Shops that adopt AI today will cut costs, improve accuracy, and outpace competitors in the next decade. The question isn’t if AI will transform diagnostics, but how quickly you’ll implement it.

Ready to build your AI-powered diagnostic system? Contact AIQ Labs to explore custom development, AI Employees, or a full transformation partnership.

Key Takeaways: ✅ AI reduces diagnostic errors by 50% (PubMed Central) and cuts labor costs by 25% (Dialzara). ✅ Data quality is the #1 factor in AI accuracy—automate normalization first. ✅ AI Employees can structure repair records for 75% less than hiring a human. ✅ EV diagnostics will require specialized AI—don’t get left behind. ✅ Start with a pilot (e.g., AI Employee for $599/month) before scaling.

The future of diagnostics is here—will your shop lead, or follow?