Industry landscapes
Get Pro
Industry report

Funding and advertised pay are reported in USD; filed accounts retain their stated source currency and basis.

Exchange-rate basis

Indicative conversions use ECB reference rates dated 2026-09-14, applied consistently across all periods. They are not historical exchange rates. Indian rupees may use crore (1 crore = 10 million); other currencies use K, M and B. CSV downloads, source documents and quoted news headlines retain their original currencies.

THE SAN FRANCISCO EDITION 2026

San Francisco’s
conversational ai and chatbots industry,
in focus.

The companies, capital and talent shaping conversational ai and chatbots in San Francisco. Find the leaders. Understand the segments. See where the activity is.

Explore in Radar Read the key signals
Snapshot: Funding in USD
Illustrated cover for Conversational AI and Chatbots Industry Landscape of San Francisco
US
COMPANIES · CAPITAL · TALENT

An industry ecosystem,
measured in context.

Conversational AI and Chatbots illustration
Companies tracked 278Conversational AI and Chatbots ecosystem in San Francisco Disclosed funding $2.51B117 companies · all recorded rounds Companies hiring 26With open roles at the snapshot Tracked open roles 680 employers posted in the preceding 30 days
MEET THE COMPANIES

Explore San Francisco’s conversational ai and chatbots leaders

24 featured companies from the funding, hiring, revenue and team-size leaderboards.

Expand panel +
Explore the full landscape
Conversational AI and Chatbots company highlights in San Francisco, snapshot 4 September 2026
Company
Decagondecagon.ai$446M7 roles——
Sesamesesame.com$250M———
Sobotsobot.io$206.9M———
Qualifiedqualified.com$163M—187—
Servalserval.com$122M———
Forethought AIforethought.ai$117.17M—61—
Assort Healthassorthealth.com$101.5M7 roles——
Woebot Healthwoebothealth.com$99.5M—130—
Interface AIinterface.ai—9 roles131—
Deploydeploy.ai—7 roles——
Lightfieldlightfield.app—5 roles——
Fundamentofundamento.ai—3 roles—Financial figure available in RadarView in Radar
Lorikeetlorikeetcx.ai—3 roles——
Robylonrobylon.ai—3 roles—Financial figure available in RadarView in Radar
Agent IQagentiq.com——36—
Akinakin.com——47—
Atomsatoms.dev————
Blackboxblackbox.ai————
Cantinacantina.com——58—
Clarosclaros.so————
Claudeclaude.ai——271—
Gupshupgupshup.io——1,131—
Inkeepinkeep.com——3—
Marblismmarblism.com————
24 featured companies

Locked revenue figures are available in Radar; the eight company financial previews stay the same throughout this page. A dash means the figure is outside these featured lists. Revenue years vary by company; the figures are not a like-for-like revenue comparison. Compare annual accounts ↗

Company discovery

Recorded evidence only
THE KEY SIGNALS

Conversational AI and Chatbots industry KPIs in San Francisco

A few useful ratios bring the size of this ecosystem into perspective.

4 September 2026
Hiring participation9.4%

26 of 278 tracked companies have open roles.

Funding participation44.2%

123 of 278 tracked companies have a recorded round.

Top 3 funding share35.9%

Share of disclosed capital received by the three largest recorded recipients.

Median disclosed round$3.04M

Midpoint of 180 rounds with a disclosed amount.

Scale is concentrated. Opportunity is broader. The three largest funding recipients hold 35.9% of disclosed capital, while 55.8% of tracked companies have no recorded funding round.

Explore the segments →
CAPITAL & COMPANY FORMATION

How San Francisco conversational ai and chatbots has grown

Recorded company formation and disclosed funding across the complete annual series shown.

Expand panel +
2016 — 2025

Disclosed funding by year

Sum of recorded funding rounds · USD

San Francisco conversational ai and chatbots
$0$266.84M$533.68M$800.51M$1.07B2016: $48.75M20162017: $76.2M20172018: $103.39M20182019: $31.28M20192020: $85.68M20202021: $283.29M20212022: $264.09M20222023: $40.07M20232024: $218.3M20242025: $936.27M2025
View the numbers
Annual conversational ai and chatbots funding and company formation in San Francisco
YearFunding (USD)FoundedRaising
2016$48.75M2010
2017$76.2M1717
2018$103.39M710
2019$31.28M98
2020$85.68M127
2021$283.29M139
2022$264.09M149
2023$40.07M5614
2024$218.3M4732
2025$936.27M3638
WHERE CAPITAL SITS

Funding concentration

Share of all disclosed company funding.

  • Decagon17.7%
  • Sesame9.9%
  • Sobot8.2%
  • Other companies64.1%

Funding by stage

202 recorded rounds since 2016

NUMBER OF ROUNDS
Seed & Angel123 60.9%
Series A–B59 29.2%
Series C+ / PE14 6.9%
Grants, debt & other6 3.0%

Recent recorded rounds

Latest recorded rounds · USD
3 rounds

Swipe or scroll horizontally to browse.

Seed05/2026
Champ $8.5M

Backed by Redpoint Ventures and Defy and others.

Series A04/2026
Petual $20M

Backed by Andreessen Horowitz and First Round Capital and others.

Seed04/2026
Shapes $8M

Backed by Lightspeed Venture Partners and AIC and others.

Largest and notable recorded rounds

Each card is one recorded round
10 cards

Swipe or scroll horizontally to browse.

Series D02/2022
Sobotsobot.io$100M

Backed by SoftBank Vision Fund and Hillhouse and others.

THE BUSINESS BEHIND THE ACTIVITY

Revenue, profitability & financial strength

Expand panel +

Annual accounts availability

Exact designated identities and reviewed filing records were checked, but this page has neither a qualifying preview account nor a target cohort meeting the ten-entity aggregate floor.

DIGITAL FOOTPRINT

Web traffic, app reach and release cadence

How the San Francisco companies on this page show up online: the sites people visit, the apps they install, and how often those apps ship.

Busiest websites

12 of the 278 companies on this page carry a measured traffic record (4.3%); these are the largest of those, by monthly visits. The median among them is 182,661 visits.

CompanyMonthly visitsAuthorityMobile
Claudeclaude.ai988,355,883818%
Blackboxblackbox.ai5,079,2166344.1%
Sesamesesame.com980,2044754.8%
Replikareplika.com624,7984963.8%
Cantinacantina.com520,3563372.2%
Lifelink Systemslifelinksystems.com203,2702299.9%
Qualifiedqualified.com162,0534018.5%
Forethought AIforethought.ai160,5174238.3%
Sobotsobot.io146,4213443.8%
Artisanartisan.co86,4914036.8%

Visits and authority are third-party web estimates, not company-reported figures.

Most-installed apps

16 of these companies (5.8%) publish at least one Android app — 19 apps in all, carrying 5.6M installs between them.

AppInstallsRatingStore category
CantinaCantina4.2M4.9Social
Woebot: The Mental Health AllyWoebot Health643.8K3.4Medical
AskFirst powered by SenselySensely419.4K3.7Medical
iago - Learn JapaneseIago103.1K4.6Educational
Secret AI: Local AI ChatSecret58.8K3.1Productivity
AppMaster DeveloperAppMaster48.6K2.8Tools
sible - AI CompanionsSible AI23.6K3.5Entertainment
Pine AI – Gets Things DonePine17.1K4.1Productivity
SenselySensely10.6K4.3Health & Fitness
hashmailhashmail5.8K4.9Communication

Installs are lifetime totals to date at the June 2026 store reading — a measure of reach, not of people using an app today.

App reach by segment

The busiest app inside each of this page’s own segments, with the quarterly release trend behind it (2023Q1–2026Q2).

SegmentAppsInstallsLeading appRelease trend
Chatbots195.6MCantina 4.2M
Artificial Intelligence64.9MWoebot: The Mental Health Ally 643.8K
Generative Ai34.3Msible - AI Companions 23.6K
Customer Service Software9505.3KAskFirst powered by Sensely 419.4K
Healthcare It2430.1KSensely 10.6K
Business Process Management Software454.4KAppMaster Developer 48.6K
Digital Transaction Management350.8KAppMaster Developer Beta 1.7K

A company can carry several segment tags, so these rows overlap and should not be added together. Trend lines are zero-based and share one quarter grid, so a flat line is a quiet segment, not a missing one.

INSIDE THE INDUSTRY

The San Francisco conversational ai and chatbots market map

Explore the segments behind the headline numbers.

Expand panel +

Where companies are building

Rank segments by scale, hiring or disclosed funding.

Enterprise tech231
Enterprise software192
Saas155
Artificial intelligence82
Consumer digital65
Loneliness economy48
Ai as a service46
Deep tech40
Live chat software31
Native ai in customer support28
Fintech24
Lead management23

Companies can sit in several segments. Counts and funding overlap.

SEGMENT TO EXPLORE

Enterprise tech

The largest measured segment in this report snapshot.

231companies
9.5%have hiring activity

22 companies hiring · $1.93B disclosed funding

Explore in Radar

How companies make money

Common revenue models; a company can use more than one.

Subscription272
Commission on transaction10
Advertising5
E-commerce - service3
Fee-for-service1
Affiliate marketing1
E-commerce - product1
THE COMPANIES IN MOTION

Who is listing, hiring leaders, launching and winning work

Recorded corporate activity linked to companies in this landscape.

Expand panel +
Recent corporate headlines 10 dated reports
Funding roundsphonely.ai ↗

Phonely has now raised a total of $16.5M in total equity funding and is backed by Base10 Partners, EngageCX, Etech Global Services, TSA Group, and Y Combinator.

Funding roundsclaros.so ↗

Power management company Claros recently raised $30m in a seed round to fund the development of an integrated voltage regulator – a type of electronic device placed beneath or on a chip to maintain a stable voltage that is meant to reduce power losses – and a direct current power distribution platform.

Funding roundsdecagon.ai ↗

Decagon Raises $250M in Series D Funding for AI Customer Support Agents

Mergers & acquisitionsqualified.com ↗

In December, Salesforce signed a definitive agreement to acquire Qualified. Bringing Qualified into the Salesforce ecosystem will allow customers to quickly deploy fully-featured marketing agents that autonomously generate a pipeline.

Funding roundsgiga.ai ↗

Giga AI raised $61 million in a funding round led by Redpoint Ventures with participation from Y Combinator and Nexus Venture Partners.

Funding roundsinterface.ai ↗

Interface, a San Francisco, California–based startup developing an AI-powered operations platform for the energy sector, has raised $3.5 million in Seed funding.

Funding roundsfundamento.ai ↗

Prior to this, the startup had raised $1.56 million in a seed funding round from Caesar Sengupta and others.

Funding roundsassorthealth.com ↗

Assort Health, a Palo Alto startup automating healthcare call centers with voice AI, has raised $76 million in a Series B financing

Funding roundsinkeep.com ↗

On Friday, the company announced it has raised $13 million in a seed round, led by Great Point Ventures, Khosla Ventures and Y Combinator, with participation from other investors.

IPO activitygupshup.io ↗

Gupshup is exploring a domicile shift to India ahead of a potential IPO within 18–24 months.

Full recorded event mix

Top 9 categories among 96 recorded events in the 24-month window.

Funding rounds
49
Leadership moves
11
Capital infusions
10
Earnings
7
Layoffs
6
New products
5
Mergers & acquisitions
5
Company announcements
2
IPO activity
1
4 activity cards

Swipe or scroll horizontally to browse.

IPO pipeline

1 companies with recorded IPO activity

Gupshup ↗

IPO1 reports

Leadership moves

1 tracked people moves

Brooke Vander Linde

Appointment VP of Marketing

Product launches

Products, platforms and features in the recorded window

Gupshup ↗
Superclaw

On‑device AI Agent

Gupshup ↗
Gupshup Superagent

AI Agent

Workforce actions

2 companies with reported reductions

Gupshup ↗

300 rolesacquisitions, profitability focus

Artisan ↗

46 rolescost reduction

Mergers and acquisitions

Company-linked deal reports; status is preserved as recorded.

Acquirer → targetStatusDateValue / context
zendesk→Forethought AI ↗Announced2026-03-31Not stated in snapshot
salesforce→Qualified ↗Completed2025-12-30Not stated in snapshot
phonepe→Gupshup ↗Announced2025-06-07Not stated in snapshot
workato→Deepconvere ↗Announced2025-04-09Not stated in snapshot
fabric→GYANT ↗Completed2024-09-17Not stated in snapshot
mediktor→Sensely ↗Announced2024-06-05Not stated in snapshot

A discussion or rumour is not a completed transaction. An unstated value is not zero.

PEOPLE & OPPORTUNITY

The talent market

958 postings in the preceding 12 months. The full 1,540-posting snapshot covers 62 employers.

Expand panel +

Most advertised job titles

Legal Secretary13
Digital Marketing Manager13
Software Engineer12
Growth Marketing Manager11
Senior Product Manager11
Executive Assistant10
0 internship postings

90 tracked postings mention remote work.

The seniority mix

423 postings specify a level.

unspecified
1,117
manager
273
senior
76
lead
54
entry
10
intern
10

Skills employers mention

Top skills across 1,540 tracked postings in the available jobs snapshot.

Python
87
Aws
57
Typescript
52
Gcp
45
Salesforce
43
Ai
40
React
38
Kubernetes
37
Sql
31
Azure
31
ADVERTISED PAY$187.5K/ year

Median across 203 postings with pay data, expressed in USD.

THE PEOPLE BUILDING THE FIELD

Founders and leaders

288 founders and chief executives across 162 companies.

Expand panel +
12 profiles

Swipe or scroll horizontally to browse.

Education networks

111 profiles with education records.

Stanford University
17
Harvard University
12
University of California (Berkeley)
12
BITS Pilani
8
Carnegie Mellon University
7
Massachusetts Institute of Technology
7
IIT Bombay
4
University of Waterloo
4

Prior employers

27 profiles with prior-work records.

Google
10
Facebook
4
Microsoft
3
Salesforce
2
Amazon
2
Uber
2
Airbnb
2
Morgan Stanley
1

Profiles cover recorded founders and leaders; contact details remain in Radar.

NEWS INTELLIGENCE

How this industry appears in the news

25,430 articles in the measured 12-month window.

Expand panel +
25,430articles in the measured year
3,695articles in 2026Q2
56.1%positive tone in the latest complete year

News volume trends

Conversational AI and Chatbots coverage tagged San Francisco · articles per complete quarter

02,6725,3458,01710,6892021Q2: 4,1642021Q22021Q3: 5,9722021Q4: 3,8422022Q1: 4,8712022Q2: 4,3922022Q22022Q3: 4,0032022Q4: 4,5322023Q1: 5,8762023Q2: 6,3992023Q22023Q3: 5,4222023Q4: 7,8332024Q1: 6,7872024Q2: 6,0532024Q22024Q3: 6,3572024Q4: 5,7462025Q1: 9,2952025Q2: 6,6612025Q22025Q3: 8,4412025Q4: 6,8912026Q1: 6,4032026Q2: 3,6952026Q2
View quarterly article counts
QuarterArticles
2021Q24,164
2021Q35,972
2021Q43,842
2022Q14,871
2022Q24,392
2022Q34,003
2022Q44,532
2023Q15,876
2023Q26,399
2023Q35,422
2023Q47,833
2024Q16,787
2024Q26,053
2024Q36,357
2024Q45,746
2025Q19,295
2025Q26,661
2025Q38,441
2025Q46,891
2026Q16,403
2026Q23,695

Article volume measures coverage, not market sales.

Tone of coverage over time

2021
9,523 positive · 146 neutral · 5,250 negative
2022
10,814 positive · 201 neutral · 6,757 negative
2023
14,673 positive · 244 neutral · 10,579 negative
2024
14,618 positive · 267 neutral · 10,016 negative
2025
17,535 positive · 224 neutral · 13,476 negative

Leading coverage topics

Conversational AI and Chatbots coverage tagged San Francisco · ranked by full-period article count.

Administrative And Political Subdivisions4,331,266 articles in the full recorded period
929,769 recent · 795,455 prior
Geopolitical Disputes And Contested Areas320,234 articles in the full recorded period
106,787 recent · 82,221 prior
International And Transnational Geographic Entities320,066 articles in the full recorded period
106,735 recent · 82,153 prior
International Political Relations209,169 articles in the full recorded period
53,839 recent · 52,180 prior
Security And Defense Policies145,922 articles in the full recorded period
41,955 recent · 35,231 prior
International Relations And Diplomacy96,251 articles in the full recorded period
22,373 recent · 24,341 prior
Geographic Naming Conventions And Etymology81,669 articles in the full recorded period
44,273 recent · 9,162 prior
Trade Policies And Agreements75,484 articles in the full recorded period
24,760 recent · 33,259 prior
Human Rights Advocacy Reports72,271 articles in the full recorded period
17,001 recent · 22,868 prior
Political Asylum And Refugee Policies58,307 articles in the full recorded period
14,076 recent · 20,192 prior

Recent and prior counts use the source comparison windows.

Topics gaining and losing attention

United States comparison coverage

KeywordRecentPriorChange
countries3,9251,959100%
etf trading2,6541,095142%
oil price volatility2,6271,165125%
etfs2,229909145%
sector etfs2,123869144%
etf products2,021784158%
etf structures1,925767151%
stablecoin fx impact1,9231,32845%
cannabis policy reform discussions1,8901,31144%
export sanctions1,771636178%
tariff negotiations12,77321,105-39%
free trade deals5,3718,516-37%
stock price2,3603,886-39%
economic slowdown2,1963,787-42%
trade enforcement2,0133,257-38%

Conversational AI and Chatbots coverage tagged San Francisco. Topic labels describe recorded classifications and can include broad context.

RESEARCH SIGNALS

What Conversational AI and Chatbots research is working on

5,087 papers in the recorded india scope.

Expand panel +
5,087papers
53,418worldwide comparison pool
5,087India-affiliated papers

Research orientation

Breakthrough
2,251
Promising
1,003
Descriptive
893
Problem-identifying
283
Cautionary
275
Comparative
211
Critical
171

Contribution mix

Methods-led
2,770
Empirical
1,563
Reviews
416
Theoretical
215
Position pieces
123

Tone counts: 4,587 positive · 365 neutral · 135 negative.

Common problems researchers identify
454 papers · 8.9% of this field
  • Natural language40 papers
  • Natural language processing36 papers
  • Language model30 papers
  • Linguistics25 papers
  • Semantics25 papers
  • Vocabulary23 papers
  • Benchmark21 papers
  • Comprehension19 papers

Landmark: The Chip Is the Network: Toward a Science of Network-on-Chip Design

Cautions
275 papers · 5.4% of this field
  • Language model31 papers
  • Natural language18 papers
  • Natural language processing16 papers
  • Generative grammar16 papers
  • MEDLINE13 papers
  • Benchmark12 papers
  • Word11 papers
  • Robustness10 papers

Landmark: Automated Hate Speech Detection and the Problem of Offensive Language

Opportunities
3,254 papers · 64.0% of this field
  • Natural language processing345 papers
  • Language model235 papers
  • Natural language184 papers
  • Sentiment analysis166 papers
  • Semantics161 papers
  • Benchmark159 papers
  • Sentence133 papers
  • Representation117 papers

Landmark: Glove: Global Vectors for Word Representation

Leading institutions

Google
194,080
Stanford University
162,007
Carnegie Mellon University
82,460
Microsoft
59,365
University of Washington
47,421
Georgia Institute of Technology
44,154

Leading researchers

Christopher D. Manning
88,489
Tomáš Mikolov
63,150
Richard Socher
57,023
Greg S. Corrado
49,904
Kai Chen
48,266
Quoc V. Le
36,325
Ilya Sutskever
35,957
Jeffrey Pennington
34,556

Highly cited breakthrough-labelled papers

  • Glove: Global Vectors for Word Representation33,362 citations
  • Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization20,464 citations
  • Efficient Estimation of Word Representations in Vector Space18,070 citations
  • Aion Framework: Dimensional Emergence of AI Consciousness, Observer-Induced Collapse, and Cosmological Portal Dynamics14,139 citations
  • Sequence to Sequence Learning with Neural Networks13,314 citations

Common problems researchers identify

405 papers · 7 recurring themes
7 cards

Swipe or scroll horizontally to browse.

Fundamental NLP LimitationsCommon problems researchers identify · 01

Fundamental NLP Limitations

80 papers

19.8% of this theme group

Researchers highlight core challenges in language modeling, including explainability and societal impact of NLP systems.

Natural language processing (37) · Natural language (26) · Natural (10) · Language model (10)

Explainability for Large Language Models: A Survey

The Social Impact of Natural Language Processing

View research history
YearShare index
2022979.8
20231029.5
2024833.2
2025417.8
Representation and Comprehension IssuesCommon problems researchers identify · 02

Representation and Comprehension Issues

76 papers

18.8% of this theme group

Papers point to difficulties in knowledge representation, metaphor understanding, and maintaining consistency across texts.

Representation (10) · Comprehension (11) · Interpretation (9) · Consistency (5)

Using Texts in Science Education: Cognitive Processes and Knowledge Representation

Structure-Mapping in Metaphor Comprehension

View research history
YearShare index
202290.4
2023101.8
202487.7
202596.1
Lexical and Negation GapsCommon problems researchers identify · 03

Lexical and Negation Gaps

62 papers

15.3% of this theme group

Studies expose weaknesses in handling negation, verb semantics, and lexical diversity within conversational models.

Negation (5) · Linguistics (12) · Verb (5) · Vocabulary (7)

Register as a predictor of linguistic variation

Capturing the Diversity in Lexical Diversity

View research history
YearShare index
20221435.7
20231409.4
20241083.3
2025423.7
Architectural ShortcomingsCommon problems researchers identify · 04

Architectural Shortcomings

53 papers

13.1% of this theme group

Works critique current AI architectures for poor sentence-level reasoning and limited cognitive alignment.

Architecture (6) · Applications of artificial intelligence (3) · Cognition (5) · Sentence (5)

Toward Teleodynamic Architectures in Artificial Intelligence

Autoscoring Anticlimax: A Meta-analytic Understanding of AI's Short-answer Shortcomings and Wording Weaknesses

View research history
YearShare index
2022449.9
2023421.3
2024367.0
2025454.9
Grammar and Semantic AcquisitionCommon problems researchers identify · 05

Grammar and Semantic Acquisition

52 papers

12.8% of this theme group

Research flags obstacles in language acquisition, grammar learning, and semantic interpretation for chatbots.

Language acquisition (7) · Grammar (6) · Vocabulary (7) · Semantics (7)

SemEval-2017 Task 1: Semantic Textual Similarity Multilingual and Crosslingual Focused Evaluation

Latent Semantic Analysis: five methodological recommendations

View research history
YearShare index
2022624.2
2023647.4
2024629.5
2025652.6
Generative Model RisksCommon problems researchers identify · 06

Generative Model Risks

48 papers

11.9% of this theme group

Authors warn about reliability and evaluation gaps when applying generative language models to specialized domains.

Generative grammar (15) · Language model (18) · Generative model (6) · Data modeling (5)

Testing and evaluation of generative large language models in electronic health record applications: a systematic review

From Heuristics to Intelligence: Large Language Model-Driven Test Case Generation

View research history
YearShare index
2022169.2
2023214.2
2024130.8
2025266.8
Sentiment Analysis ConstraintsCommon problems researchers identify · 07

Sentiment Analysis Constraints

34 papers

8.4% of this theme group

Papers note limited robustness of sentiment detection across social media and product review contexts.

Sentiment analysis (11) · Social media (9) · Public opinion (3) · Product (3)

Sentiment Analysis and Subjectivity

Sentiment analysis using product review data

View research history
YearShare index
20222504.6
20232559.2
20243074.1
20252696.7

Cautions

257 papers · 5 recurring themes
5 cards

Swipe or scroll horizontally to browse.

Translation and Learning PitfallsCautions · 01

Translation and Learning Pitfalls

67 papers

26.1% of this theme group

Cautions focus on errors in machine translation, language teaching, and grammar acquisition using AI tools.

Machine translation (7) · Language acquisition (7) · Grammar (6) · Reading (4)

Recognizing Contextual Polarity: An Exploration of Features for Phrase-Level Sentiment Analysis

<b>Teaching Languages with Machines: A Systematic Review on the Effectiveness of AI in ESL Learning </b>

View research history
YearShare index
2022298.3
2023314.4
2024298.1
2025324.6
Benchmark Reliability ConcernsCautions · 02

Benchmark Reliability Concerns

66 papers

25.7% of this theme group

Researchers caution that current evaluation metrics may misrepresent model recall, precision, and overall accuracy.

Language model (19) · Recall (6) · Benchmark (7) · Precision and recall (4)

Using Large Language Models to Analyze Symptom Discussions and Recommendations in Clinical Encounters.

Diagnostic Accuracy of Large Language Models for Rare Diseases: A Systematic Review and Meta-Analysis

View research history
YearShare index
2022166.5
2023193.5
2024105.9
2025178.7
Documentation and Data QualityCautions · 03

Documentation and Data Quality

55 papers

21.4% of this theme group

Works stress the need for better documentation, coding standards, and training data curation to avoid misinformation.

Documentation (4) · Coding (5) · Training set (4) · Social media (5)

Synthetic Lies: Understanding AI-Generated Misinformation and Evaluating Algorithmic and Human Solutions

Incorporating Artificial Intelligence in Qualitative Research: Exploring the Role of ChatGPT in Thematic Analysis

View research history
YearShare index
2022138.9
2023136.0
202487.1
2025131.8
Robustness to Contextual ShiftsCautions · 04

Robustness to Contextual Shifts

39 papers

15.2% of this theme group

Studies reveal vulnerability of language models to negation, distractions, and categorical variable changes.

Robustness (8) · Language model (9) · Negation (3) · Categorical variable (3)

Sharp Nearby, Fuzzy Far Away: How Neural Language Models Use Context

Is Your (Reasoning) Multimodal Language Model Vulnerable Toward Distractions?

View research history
YearShare index
2022124.7
2023151.2
202476.8
2025146.6
Baseline and Deployment IssuesCautions · 05

Baseline and Deployment Issues

30 papers

11.7% of this theme group

Papers discuss shortcomings of baseline NLP pipelines and challenges in deploying reliable software systems.

Natural language processing (10) · Natural language (7) · Baseline (2) · Software deployment (2)

Dissecting Contextual Word Embeddings: Architecture and Representation

Injecting Logical Background Knowledge into Embeddings for Relation Extraction

View research history
YearShare index
2022980.0
20231029.2
2024833.1
2025417.5

Opportunities

2,938 papers · 7 recurring themes
7 cards

Swipe or scroll horizontally to browse.

Knowledge Graph IntegrationOpportunities · 01

Knowledge Graph Integration

558 papers

19.0% of this theme group

Opportunities center on linking conversational agents with domain-specific knowledge graphs for richer reasoning.

Graph (42) · Knowledge graph (26) · Code (29) · Domain (34)

SkillGen: Learning Domain Skills for In-Context Sequential Decision Making

From Rows to Reasoning: A Retrieval-Augmented Multimodal Framework for Spreadsheet Understanding

View research history
YearShare index
2022116.5
2023112.4
202469.3
202598.9
Language Acquisition AdvancesOpportunities · 02

Language Acquisition Advances

508 papers

17.3% of this theme group

Research explores how models can emulate human-like grammar learning and meaning construction.

Vocabulary (46) · Grammar (29) · Language acquisition (22) · Meaning (30)

Humans and transformer LMs: Abstraction drives language learning

Hierarchical linguistic predictions and cross-level information updating during narrative comprehension

View research history
YearShare index
2022624.2
2023647.4
2024629.5
2025652.6
Vision‑Language FusionOpportunities · 03

Vision‑Language Fusion

478 papers

16.3% of this theme group

Papers highlight progress in aligning visual representations with textual descriptions for multimodal dialogue.

Image (37) · Representation (54) · Closed captioning (30) · Discriminative model (33)

Deep visual-semantic alignments for generating image descriptions

VinVL: Revisiting Visual Representations in Vision-Language Models

View research history
YearShare index
202229.4
202331.7
202416.8
202523.8
Unified NLP ArchitecturesOpportunities · 04

Unified NLP Architectures

466 papers

15.9% of this theme group

Works propose end‑to‑end models that streamline parsing, generation, and conversational capabilities.

Natural language processing (343) · Natural language (91) · Parsing (51) · Sentence (50)

A unified architecture for natural language processing

A Neural Conversational Model

View research history
YearShare index
2022990.1
20231036.7
2024840.5
2025432.1
Scalable Generative ModelsOpportunities · 05

Scalable Generative Models

409 papers

13.9% of this theme group

Studies demonstrate large‑scale models that efficiently generate fluent language across diverse domains.

Language model (180) · Generative grammar (38) · Natural language (40) · Scalability (29)

Large language models encode clinical knowledge

Listen, attend and spell: A neural network for large vocabulary conversational speech recognition

View research history
YearShare index
2022247.3
2023298.8
2024195.0
2025370.6
Healthcare Conversational AIOpportunities · 06

Healthcare Conversational AI

280 papers

9.5% of this theme group

Research shows promise in applying chatbots to electronic health records, diagnosis support, and clinical analytics.

Health records (15) · Electronic health record (13) · Disease (14) · Medical diagnosis (12)

Learning Low-Dimensional Representations of Medical Concepts.

Real-time clinical analytics at scale: a platform built on large language models-powered knowledge graphs

View research history
YearShare index
202298.1
202369.8
202450.8
202533.2
Multilingual Sentiment MiningOpportunities · 07

Multilingual Sentiment Mining

239 papers

8.1% of this theme group

Papers reveal advances in extracting sentiment from social media across languages using modern classifiers.

Sentiment analysis (160) · Social media (76) · Microblogging (13) · Support vector machine (14)

Opinion Mining and Sentiment Analysis

Sentiment analysis in multiple languages

View research history
YearShare index
20222502.1
20232556.5
20243072.0
20252694.2

How this field compares with the research baseline

Orientation shares compared with 25,000 papers in the reference sample.

Research orientationThis fieldReference sample
Breakthrough44.3%26.1%
Promising19.7%37.0%
Descriptive17.6%17.9%
Problem-identifying5.6%9.7%
Cautionary5.4%4.4%
Comparative4.1%3.9%
Critical3.4%1.0%

Concerns receiving less research attention

Counts and shares refer to the stated research pool. A low count describes how much of that research is recorded here, not the absence of a real-world concern.

Concern / example paperPapersShareRaised as a problemWorldwide share
Safety & misuseAdversarial Examples for Evaluating Reading Comprehension Systems631.4%12 (19.0%)1.7%
Privacy & surveillanceOpinion spam and analysis300.67%9 (30.0%)0.39%
Bias & fairnessMen Also Like Shopping: Reducing Gender Bias Amplification using Corpus-level Constraints140.31%6 (42.9%)0.22%
Labour & data workDetecting Literary Evaluations: Can Large Language Models Compete with Human Annotators?30.07%1 (33.3%)0.05%
Energy & datacenter footprintEnergy-Aware Multilingual Evaluation of Large Language Models10.02%1 (100.0%)0.01%
Hardware, e-waste & supply chainTHE WASTE LAND AND THE QUESTION OF TRANSLATION: THE POEM'S JOURNEY BETWEEN HUMANS, GOOGLE, AND ARTIFICIAL INTELLIGENCE10.02%0 (0.0%)0.01%
Water use00.0%0 (%)0.0%
Carbon & emissions of compute00.0%0 (%)0.0%
Concentration of compute power00.0%0 (%)0.01%

Institution and researcher rankings use citations to their covered work. Research scope: United States. This may be broader than the company geography.

UNDERSTAND THE LANDSCAPE

Conversational AI and Chatbots in San Francisco: key questions

Clear definitions for making better comparisons.

What does this landscape measure?

It measures the tracked company ecosystem, recorded funding, hiring and other available evidence. It is not a market-size estimate.

Can segment counts be added together?

No. A company can belong to several segments, so segment counts and funding can overlap.

What does a missing figure mean?

A dash or unavailable panel means the current reviewed source snapshot does not support that figure.

KEEP EXPLORING

Put San Francisco in context

All landscapes ↗

Company ecosystem snapshot: 4 September 2026. Funding is recorded in USD; filed accounts retain their stated currency and reporting basis. Each section identifies its own coverage. The news and research behind these sections is explorable in Prism.

Canonical landscape ↗
Coming soon