Top AI Tools for Analyzing Emotional Tone

Top AI Tools for Analyzing Emotional Tone

Understanding emotional tone in conversations is now easier with AI tools. These tools analyze text, voice, and facial expressions to provide insights into emotions, sarcasm, and context. Whether you're improving customer support, monitoring team morale, or identifying manipulation in personal interactions, these tools offer various features to meet your needs. Here's a quick overview of the top options:

• Gaslighting Check: Detects manipulation in conversations with real-time text and voice analysis. Free and premium plans start at $9.99/month.
• Hume AI: Offers multimodal analysis of text, voice, and facial expressions with over 600 emotion tags. Free tier available; enterprise pricing on request.
• Imertiv AI: Tracks emotions across text, voice, and video with real-time insights. Custom pricing available.
• MoodMe Face AI: Focuses on facial expressions with 450 3D landmarks per frame. Processes data locally for privacy. Pricing tailored to needs.
• Symbl.ai: Combines text, audio, and video inputs with real-time sentiment tracking. Free tier for 1,000 minutes/month; custom pricing for advanced features.
• CallMiner: Designed for contact centers, analyzing emotional tone in high-volume interactions. Enterprise pricing based on call volume.
• iWeaver: Links emotions to causes in text, images, and more. Free tier available; Pro and enterprise plans offered.
• ScreenApp: Provides real-time voice emotion analysis with multimodal inputs. Free for three analyses; plans start at $19/month.
• Revoicer: Creates emotion-rich AI voiceovers in 40+ languages. Plans start at $27/month.
• Affectiva: Analyzes facial expressions, voice, and body language with enterprise-level accuracy. Pay-per-minute or custom licensing.

Each tool varies in features, privacy measures, and pricing. Choose one based on your goals, whether it's personal use, customer service, or large-scale business needs.

Quick Comparison

Choose the tool that aligns with your specific needs, whether for personal insights, content creation, or business applications.

::: @figure

{AI Emotional Tone Analysis Tools Comparison: Features, Privacy & Pricing} :::

Hume: The AI That Understands and Responds to Your Emotions

1. Gaslighting Check

Gaslighting Check is designed to help people spot manipulation tactics in personal conversations. It focuses on identifying patterns like blame-shifting, reality distortion, and emotional invalidation - key indicators of gaslighting. This tool is especially useful for those navigating difficult relationships.

Multimodal Analysis: Text and Voice

The platform uses advanced NLP and deep learning to analyze both text and voice in real time. It can flag manipulative phrases like "You're overreacting" and detect vocal cues such as changes in tone, pitch, or stress. By combining these insights, users get a fuller picture of the conversation's dynamics.

Real-Time Feedback

Gaslighting Check processes conversations as they happen, providing instant alerts for manipulative language or tone shifts. This feature allows users to identify gaslighting vs manipulation right away, rather than piecing it together after the fact.

Privacy and Security

To ensure user confidentiality, the platform uses end-to-end encryption and automatically deletes data after processing. This commitment to privacy makes it a trusted option for detecting emotional manipulation.

Pricing Options

Gaslighting Check offers three subscription plans:

• Free Plan ($0): Basic text analysis features.
• Premium Plan ($9.99/month): Includes voice analysis, history tracking, and detailed reports with actionable insights.
• Enterprise Plan: Custom pricing and features tailored to specific needs.

2. Hume AI

Hume AI takes emotional analysis to the next level by integrating facial expression recognition with tools like Gaslighting Check, which already analyze text and voice. Through a unified API, the platform evaluates facial, vocal, and textual expressions using seven specialized models. It identifies over 600 distinct emotional tags and voice traits [4], leveraging a multidimensional mapping system that goes beyond basic positive or negative labels [5].

Multimodal Analysis

This platform digs deep into human expression by analyzing textual sentiment, vocal tones (including prosody and non-verbal cues like laughter or sighs), and facial movements. It captures emotional signals across numerous dimensions [4][5]. Hume AI’s Dynamic Reaction model even tracks how these expressions evolve over time, combining visual and audio inputs for a more comprehensive understanding [4].

Real-Time Analysis Capabilities

Hume AI is built for real-time emotional manipulation detection applications, offering tools like the Empathic Voice Interface (EVI), which processes live speech with a 250-millisecond latency [4]. The platform supports real-time use cases through WebSocket and REST endpoints, ensuring seamless responsiveness during live interactions [5].

Privacy and Data Security

Security and privacy are top priorities for Hume AI. The platform is HIPAA compliant and features a "Zero data retention" policy for the EVI API, meaning it doesn’t store chat histories, transcripts, or recordings [6]. For the Expression Measurement API, any uploaded image, audio, or video files are deleted immediately after analysis [6]. Users can also opt out of contributing anonymized data for model training, and by default, customer-submitted data is not used for this purpose [6]. Organizations dealing with health information can request Business Associate Agreements by reaching out to legal@hume.ai [6].

Pricing Tiers

Hume AI offers a free tier to help users get started with expression analysis [4]. The company is revising its pricing structure as of early 2026 to provide additional resources, with the latest rates available on their pricing page [4][5].

3. Imertiv AI

Imertiv AI brings together video, audio, text, and image data to analyze emotions in a comprehensive way. It uses Facial Emotion Recognition (FER) to monitor facial expressions frame by frame, while Speech Emotion Recognition (SER) evaluates vocal elements like tone, pitch, and rhythm to detect emotional cues. Its text analysis stands out by identifying nearly 29 different emotions, including curiosity, confusion, and admiration [7][12].

This tool, like others in its category, employs a multimodal approach to create a detailed emotional profile.

Multimodal Analysis

Imertiv AI assesses emotional intensity and polarity across all input types. For example, during a podcast analysis, it identified "neutral" as the dominant emotion at 63.19%, while also detecting subtler feelings like realization [8][12]. The system's speaker diarization feature separates voices in group settings, enabling it to track each participant’s emotional journey [9][11].

Real-Time Analysis Capabilities

With its multimodal framework, Imertiv AI delivers real-time emotional insights. The platform visualizes emotional changes using Valence-Arousal model graphs, which update dynamically as the content unfolds [9][11]. It also includes an LLM-powered "AI Insights" feature, allowing users to interact with the AI and ask detailed questions about the emotional context of a specific video or audio file [8][10]. A live emotion tracking feature is on the horizon, promising even more immediate feedback [11][3].

Privacy and Data Security

To ensure secure usage, Imertiv AI employs API keys and tokens for authorized access. Additionally, expert psychologists review the data to minimize biases and refine emotional triggers [8][3]. In a 2026 benchmark test of emotion AI tools, Imertiv AI demonstrated a 40% success rate in accurately identifying dominant emotions from facial images [3].

Pricing Tiers

Imertiv AI offers a free version, giving users the chance to explore its analysis capabilities [3]. For more advanced needs, the platform provides Emotion APIs tailored for Video, Audio, Text, and Image analysis. Custom pricing for high-volume usage is available upon request [7][9][11]. These features highlight the platform's role in shaping the future of AI-driven emotional analysis.

4. MoodMe Face AI

MoodMe Face AI focuses on facial analysis through its Vision Edge AI, identifying seven key emotions: happiness, sadness, surprise, anger, fear, disgust, and neutrality. It tracks 450 3D facial landmarks per frame, allowing it to detect subtle microexpressions that reveal authentic emotions. With a face recognition accuracy rate exceeding 99%, the platform has been adopted by over 160 organizations, including FIFA, Gucci, and Stanford University [13][14].

Multimodal Analysis

MoodMe goes beyond facial recognition by incorporating multimodal analysis in specific tools. For example, the MoodMirror Journal combines facial expression data with text and speech analysis to provide a deeper understanding of emotional states [15]. The system assigns per-frame percentage scores (ranging from 0–100%) for each emotion, enabling it to identify complex emotional shifts and blends - like 62% happiness mixed with 28% surprise [13].

Real-Time Analysis Capabilities

MoodMe delivers real-time insights with millisecond-level latency, maintaining accuracy even when the user's head is tilted up to 30 degrees [13][15]. All processing happens entirely on the user’s device, whether it’s iOS, Android, or Web-based. This technology is increasingly used to detect emotional manipulation in chats and digital interactions. This eliminates the need for internet connectivity or cloud-based processing, ensuring immediate results without delays [13][16].

"MoodMe helped us refine our messaging, grow our audiences, and amplify our mission to create real, positive change." - Guillaume Cabane, Unity Customer [13][15]

Privacy and Data Security

MoodMe prioritizes privacy with a "Privacy First" design. No facial images or data are sent to the cloud, and all analysis is handled locally on the device. Faces are not stored, even temporarily, after processing [13][16]. This setup adheres to GDPR and other global privacy standards, making it a reliable choice for industries with strict data regulations [13][14]. The AI models are trained on diverse datasets to ensure fair performance across various skin tones, age groups, and lighting conditions [13][16].

Pricing Tiers

MoodMe does not offer a free version of its Emotion Detection Engine [3]. Pricing is tailored based on SDK integration requirements for different platforms. Interested users can request a demo or quote directly through their website [13][15].

5. Symbl.ai

Symbl.ai takes sentiment detection to the next level by combining text, audio, and video inputs using advanced multimodal techniques.

The platform measures tone with a polarity score ranging from –1.0 (negative) to +1.0 (positive), categorizing sentiment as positive, neutral, or negative. It works with both recorded files and live streaming data, making it adaptable for a range of applications [17].

Multimodal Analysis

Symbl.ai uses aspect-based sentiment analysis to break conversations into specific topics, assigning sentiment to each aspect instead of providing a single overall score. This method helps identify which parts of a discussion evoke positive or negative reactions. By leveraging deep learning models, including the Nebula LLM, Symbl.ai can even pick up on subtle cues like sarcasm and irony [18][19]. This makes it a strong choice for live interaction scenarios.

Real-Time Analysis Capabilities

For live conversations, Symbl.ai processes sentiment sentence-by-sentence through WebSockets and its Streaming API. This allows for real-time applications like agent coaching, compliance monitoring, and identifying at-risk calls. Additionally, the platform tracks key conversation metrics, including speaker ratio, talk time, silence duration, speaking pace, and speaker overlap [17].

"Symbl.ai's Sentiment API is flexible, extensible and user-friendly so you can easily transcribe and capture sentiments in real-time over WebSockets." – Eric Giannini, Lead Developer Evangelist & Advocate, Symbl.ai [20]

Privacy and Data Security

Symbl.ai emphasizes privacy by analyzing conversation transcripts instead of raw voice audio, ensuring personal identity remains protected. The platform also provides automated compliance alerts during live conversations to help identify policy violations and minimize legal risks [19][20].

Pricing Tiers

Developers can start for free with a plan that includes 1,000 minutes of processing per month - no credit card required [21]. For enterprise-level deployments and access to advanced features like the Nebula LLM, users can reach out to the sales team for tailored solutions [19].

6. CallMiner

CallMiner takes conversation intelligence to the next level by combining acoustic and linguistic analyses to deliver insights for high-volume contact centers and customer teams. It’s designed to capture emotional tone across various communication channels, offering a deeper understanding of customer and agent interactions [22].

Multimodal Analysis

CallMiner’s approach to emotional tone is rooted in both voice and text analysis. For voice interactions, it evaluates acoustic cues like stress levels, pitch changes, and speech rate to gauge emotional intensity [22]. On the text side - covering chat, email, and social media - it uses Natural Language Processing (NLP) to analyze word patterns, context, and even emojis [22]. Its Emotion Solution Suite moves beyond simple sentiment scores that track emotional manipulation by pinpointing emotions such as frustration or joy, delivering actionable insights into how both customers and agents are feeling [22].

"Language is complex, and as a process for quantifying and scoring language, sentiment analysis is equally complex... customer analytics software vendors must consider acoustic measurements (the rate of speech, stress in a caller's voice, and changes in stress signals) in the context of the conversation." – The Team at CallMiner [22]

Real-Time Analysis Capabilities

CallMiner processes conversations as they happen, enabling real-time coaching, immediate escalation alerts, and proactive responses powered by adaptive machine learning. This feature is particularly useful in identifying at-risk customers who may need extra attention, ensuring that urgent situations are handled promptly [22].

Privacy and Data Security

Data security is a top priority for CallMiner. It operates with a PCI-certified hosted model and includes automatic redaction of sensitive information to maintain compliance with industry regulations [23]. Additionally, it flags compliance-related issues during its analysis, ensuring interactions meet required standards - making it a strong choice for large-scale operations [24].

Pricing Tiers

CallMiner is available as a licensed SaaS solution tailored for enterprise-level deployments. Pricing is customized based on call volume and specific feature needs, with details available upon request. This setup is designed to deliver measurable business improvements compared to in-house solutions. Interested organizations can reach out to CallMiner’s sales team for a personalized quote [22].

7. iWeaver

iWeaver stands out among sentiment analysis tools by blending data processing with advanced reasoning. Unlike many tools that stop at tagging emotions, iWeaver connects emotional tone with a broader knowledge base, offering more meaningful insights [1].

Multimodal Analysis

iWeaver can handle a wide range of input formats, including social media comments, e-commerce reviews, emails, PDFs, and even images extracted through OCR. Its batch processing feature allows it to analyze multiple files at once, making it ideal for large datasets [25]. With its advanced natural language processing (NLP) capabilities, iWeaver achieves over 90% accuracy in detecting sarcasm and mixed emotions [1].

"iWeaver is not just a sentiment tagger; it is an AI-native synthesis engine." – iWeaver Blog [1]

One of its standout features, "Why-Behind-the-Emotion", links data across multiple documents to uncover root causes of emotions. For example, it can identify that user frustration stems from a specific UI update [1]. This functionality ensures that iWeaver provides actionable insights at the speed demanded by real-time scenarios.

Real-Time Analysis Capabilities

iWeaver excels in real-time sentiment analysis. Its online tool processes up to 5,000 characters per session in the free tier, delivering results instantly. For enterprise users, it offers automated monitoring of brand-related keywords and an interactive AI chat interface that provides contextual responses in seconds [25].

Privacy and Data Security

In addition to its analytical power, iWeaver places a strong emphasis on data security. The platform uses end-to-end encryption, secure transmission protocols, protected cloud storage, and regular security audits. Strict access controls and automatic data deletion policies ensure that sensitive information is handled responsibly and not stored longer than necessary [25].

Pricing Tiers

iWeaver offers a free version for basic analysis, requiring no signup, while Pro and custom Enterprise plans cater to users needing deeper insights and large-scale monitoring [1] [25]. With a user rating of 4.6/5 from 2,878 votes, it’s a well-regarded, budget-friendly choice for sentiment analysis [25].

8. Voice Emotion Analysis by ScreenApp

ScreenApp's Voice Emotion Analysis tool provides real-time emotional feedback from audio and live video, making it a go-to solution for anyone needing instant insights during conversations. With over 3 million users [27] and an impressive 85–90% accuracy rate, it identifies six core emotions: happiness, sadness, anger, fear, surprise, and neutral [26].

Multimodal Analysis

This tool doesn’t just stop at voice - it combines voice tone, facial expressions, and text sentiment to offer a well-rounded emotional understanding. It evaluates vocal elements like tone, pitch changes, rhythm, stress levels, and emotional intensity [26]. Add to that its ability to analyze facial expressions from video frames and text sentiment from transcripts [27][30].

One standout feature is the "Ask AI" function, which lets users query recordings with specific questions, such as, "What frustrations were most commonly mentioned?" The answers come with timestamps for easy reference [30]. It also includes speaker diarization, a feature that identifies individual speakers, tracks their talk-time ratios, and highlights interruption patterns [28][29]. The platform supports a range of file formats, including MP3, WAV, M4A, MP4, MOV, AVI, and WebM [26][29].

Real-Time Analysis Capabilities

Speed is another strength. ScreenApp processes files within 30–60 seconds and even supports live streams. With the help of GPT-4 and Claude, the tool achieves over 95% accuracy in its analyses [29][32]. This aligns with the broader trend of leveraging multimodal analysis for precise, real-time emotional detection.

Privacy and Data Security

Your privacy is a priority. All audio and video uploads are encrypted during transmission and processing [26][27][32]. Files are securely handled and deleted after analysis, ensuring no data is stored permanently without user consent [26][27]. The platform is accessible across desktop, tablet, and mobile devices, with seamless cloud synchronization [29].

Pricing Tiers

ScreenApp offers 3 free analyses for video, audio, or documents, complete with full AI insights [29]. For more extensive use, the Growth plan is available at $19/month (billed annually) or $30/month (billed monthly), offering unlimited analysis and 600 AI credits per year. The Business plan costs $34/month (billed annually) or $69/month (billed monthly), featuring unlimited credits and API access [29]. With a 4.8/5 star rating, users appreciate its robust free tier and professional-grade features [31].

9. Revoicer

Revoicer uses Neural-Emotion technology to create voiceovers that convey a range of emotions. Users can choose from tones like friendly, cheerful, sad, angry, whispering, shouting, terrified, or excited, offering plenty of flexibility for different projects[34].

An impressive 97% of listeners reportedly can't tell the difference between Revoicer's AI-generated voiceovers and human speech[34]. The platform supports over 40 languages and various English accents, including American, British, Australian, Indian, and Irish. However, its emotional tone options are currently limited to 11 specific English voices. Depending on the subscription tier, users can access between 80 and 250+ voices[56, 58].

Pricing Tiers

Revoicer has a straightforward pricing structure with three subscription options:

• Standard Plan: A one-time $67 payment or $27–$37 per month. It includes 100+ voices and 600,000 characters per month[34].
• Pro Plan: Costs $47 per month, offering 250+ voices, 1,000,000 characters per month, and access to a background music library[34].
• Agency Plan: Priced at $127 per month, this plan includes 4,000,000 characters per month and commercial agency rights[34].

All plans come with a 60-day money-back guarantee, giving users some peace of mind when trying the service.

Revoicer has received mixed reviews. It earned an 8/10 rating from Appscribed[33], but G2 users gave it a 2.5 out of 5, citing issues with achieving perfect emotional synthesis and the need for a constant internet connection[35]. Other criticisms include the lack of a free trial and limited emotional tone options for non-English languages[52, 53, 57]. Despite these challenges, Revoicer stands out for its focus on emotion-rich voice generation and its tiered pricing that caters to different needs.

10. Affectiva

Affectiva stands out among AI tools with its extensive data resources and advanced multimodal analysis capabilities. With a database of 17.4 million face videos and 8 billion facial frames from 90 countries, it has earned the trust of 90% of the world's largest advertisers and around 26% of Global Fortune 500 companies [37][38]. This makes it a go-to solution for enterprise-level emotion and behavior analysis.

Multimodal Analysis (Facial and Voice)

Affectiva's technology dives deep into understanding emotions and cognitive states by analyzing facial expressions, voice tone, head gestures, and body language [38]. Its facial analysis can identify subtle details, like the intensity and duration of a smile, across hundreds of facial cues [38]. On the vocal side, it examines "vocal bursts" and "speech prosody", which include elements like tone, pitch, and rhythm [36][3]. Beyond emotions, it can assess cognitive states such as alertness, drowsiness, and distraction - features that are particularly beneficial for automotive safety systems and academic research [36][38]. When paired with the iMotions platform, Affectiva's emotion AI integrates with other biosensors, such as eye tracking, GSR, EEG, and survey data, to deliver a full-spectrum analysis of human behavior [36][39].

Real-Time Analysis Capabilities

Affectiva excels in real-time analysis, processing voice inputs in about 300 milliseconds - on par with human response times [3]. This speed enables applications like monitoring driver alertness and detecting distractions to improve road safety [36]. It also provides real-time audience feedback during media testing, making it a powerful tool for understanding consumer engagement [36]. These capabilities reinforce Affectiva's position as a leader in emotion and behavior analysis for enterprises.

Privacy and Data Security

Affectiva places a strong emphasis on privacy, ensuring data collection is only done with informed consent [38]. The company states:

"Data will never be collected without informed consent" [40].

All video data is anonymized before being used for research, and personal identifiers are stored separately from data derived from facial and vocal analysis [40]. The Affectiva Facial Coding SDK can operate locally on Windows or Linux, removing the need for constant internet or cloud connectivity [41]. The platform complies with GDPR and CCPA standards, employing measures like SSL encryption to protect user data [40].

Pricing Tiers

Affectiva offers flexible pricing tailored to specific sectors, including Academia, Consumer Insights, and Human Factors [36]. The Facial Coding API operates on a pay-per-minute model, so users are charged only for the time they use [41]. Academic licenses are available with annual renewal options, while commercial and development licenses require direct inquiries [41]. This range of pricing options accommodates everyone from individual researchers to large-scale enterprises and universities conducting behavior research [36].

Feature Comparison Table

Choosing an emotional tone analysis tool can feel overwhelming, especially when you’re juggling priorities like real-time feedback, privacy, or pricing. To help you decide, here’s a side-by-side look at some key players and what they bring to the table:

Each tool shines in its own way. Gaslighting Check stands out for detecting manipulative patterns with both text and voice capabilities, paired with flexible pricing. If privacy is your top priority, MoodMe Face AI processes data directly on your device. For advanced multimodal analysis, Hume AI, Imertiv AI, and Affectiva provide a range of options, from facial expressions to body language. Meanwhile, iWeaver takes a unique approach by linking emotions to their causes, offering a free tier for individual users.

The pricing models are just as varied as the features. Some tools, like Gaslighting Check and iWeaver, provide free options for basic use, while others, like Affectiva, cater to specialized needs with pay-per-minute or custom licensing plans. Whether you’re an individual user or an enterprise, this table highlights how each tool aligns with different priorities for analysis, privacy, and cost.

Conclusion

Pick the tool that aligns best with your specific needs, whether you're seeking personal insights or enterprise-level analytics. For individual researchers or knowledge workers, tools like iWeaver and Hume AI are great for analyzing emotional nuances without spending a fortune - both even offer free versions to get you started. If you're a content creator aiming for realistic voiceovers, Revoicer provides emotionally rich text-to-speech in over 40 languages. On the other hand, Gaslighting Check is ideal for identifying emotional manipulation in personal conversations, with pricing options to suit different budgets.

For businesses, scalability is key. Smaller marketing teams can benefit from social listening tools that monitor brand sentiment in real time, while larger enterprises need platforms capable of processing millions of interactions across channels like voice, chat, and email. Specialized solutions for contact centers, such as live coaching prompts during calls, help manage customer frustration effectively. Advanced systems can even detect emotional invalidation to improve communication dynamics. If you're a developer building custom applications, API-based tools allow you to integrate emotion detection directly into your software.

"Look at sentiment, reviews, mentions, and the tone of what's being said. I pay attention to patterns in feedback and how people respond over time. These signals help you understand what's working - and what needs to change!" - Phil Pallen, Brand Strategist [2]

This insight highlights the importance of testing tools under real-world conditions. Use challenging phrases like "I'm not happy" to see how well tools handle sarcasm or negation. If you don’t have a data science team, focus on no-code platforms rather than developer-heavy APIs. For privacy-sensitive projects, prioritize tools that process data locally instead of relying on cloud storage.

Ultimately, the right tool will match your data needs, technical capabilities, and privacy requirements. Check the comparison table above for a quick overview of each tool's features. Whether you're analyzing personal relationships or managing customer experiences on a large scale, these tools offer diverse ways to understand emotional tone.

FAQs

How do AI tools detect emotional tone in conversations?

AI tools are designed to pick up on emotional tones by examining vocal details like pitch, tone, speed, volume, and stress - often in real-time. They rely on technologies such as Speech-to-Text (STT), Natural Language Processing (NLP), and Machine Learning (ML) to process conversations and detect emotional cues or recurring patterns.

By focusing on how something is said rather than just the words, these tools can reveal subtle emotional changes or even manipulative behaviors, like mismatched tones or veiled aggression. At the same time, they prioritize user privacy with strong encryption and data protection systems, offering a secure way to analyze communication dynamics as they happen.

How do AI tools ensure user privacy when analyzing emotional tone?

AI tools that assess emotional tone often prioritize privacy to protect sensitive user information. Many employ end-to-end encryption, ensuring conversations remain secure and inaccessible to unauthorized parties. Some also incorporate automatic data deletion, minimizing the risk of misuse by erasing information once it’s no longer needed.

Beyond these safeguards, certain tools emphasize transparency by openly detailing their data practices. They explain what information is collected, how it’s used, and ensure users give explicit consent. These efforts not only protect sensitive emotional data but also help establish trust with users.

What is the best AI tool for identifying emotional manipulation in conversations?

Gaslighting Check stands out as a top-tier AI tool for identifying emotional manipulation in conversations. Using advanced text and voice analysis, it detects tactics like gaslighting, blame-shifting, and emotional invalidation as they happen.

What makes it so effective? It offers features like real-time audio recording, detailed conversation history tracking, and in-depth reports. These tools give users a clear understanding of manipulative behaviors within their interactions.

On top of that, Gaslighting Check takes privacy seriously. With encrypted data and automatic deletion policies, your personal information stays secure while you focus on gaining clarity.