AI development companies, as listed here, are software firms you can hire to design and build custom AI systems, such as language-model assistants, document automation or computer vision. They are not AI research labs or model providers.
Which AI development companies should you consider?
Start with the 20 companies below. Each links to public pages showing its AI work, 65 project pages and 1 portfolio index in all, covering chatbots and AI assistants; AI agents and orchestration; predictive models and analytics; and 5 other types of work. We publish a full count for 9 of these 20 portfolios: 13% to 83% AI work.
All 20 at a glance
Most common, at 11 of 20 companies: chatbots and AI assistantsA square is at least one listed project page of that type, and a number means more than one. Select a column heading to filter the list.
All 20 companies, A to Z, against the types of work their listed project pages show
1 of the 9 portfolios we counted in full is mostly AI workEach dot is one company: across, the public projects on its portfolio (log scale); up, the share of them that are AI work. Companies on the same dashed curve have the same number of AI projects: 5, 10, 20 and 30.
51020501002000%25%50%75%100%5102030half the portfolio↑ Share that is AI workProjects on the portfolio, log scale →AE StudioAgile InfowaysDevBrotherFlyapsGeniuseeNeotericSigma SolveStride ConsultingTechAhead
All 20 companies, A to Z, with headquarters, delivery locations, the counted share of public projects, case studies read, the most-named standards or methods, and project pages
Look up the type of work closest to your project below. Of the 20 companies on this AI list, the most common types include chatbots and AI assistants (11 companies); AI agents and orchestration (10); predictive models and analytics (9); and document, email and claims automation (8).
Chatbots and AI assistants
Show or hide the companies for Chatbots and AI assistants
Eleven of the 20 companies on this AI list have a public project in chatbots and AI assistants.
AE Studio: an AI Spanish conversation feature for Pimsleur and a pilot benefits qualification assistant for IncomeMax
DevBrother: AI assistants and document intake for insurance agencies
EvinceDev: an AI chatbot insurance app named careless and the MindFlow AI wellness companion app
Fingent: a conversational AI agent over call center data
LeewayHertz: a dealer support chatbot for URC and a RAG service desk chatbot for Rackspace
Master of Code Global: a GenAI FAQ assistant for insurance, a Teams knowledge assistant for a credit union and a conversational onboarding agent for wellness customers
Neoteric: a proof-of-concept GenAI maintenance assistant for ZF Group and an AI assistant for LeadStar
Xicom: an asset tracking app with an LLM assistant, a conversational AI financial planning platform and a mental wellness app with an AI assistant
AI agents and orchestration
Show or hide the companies for AI agents and orchestration
Ten of the 20 companies on this AI list have a public project in AI agents and orchestration.
DevBrother: an agent team for a multimodal content pipeline, portfolio index only
Devox Software: an agentic system for settlement reconciliation, an AI anomaly investigation workflow and AI visual inspection for automotive production lines
Geniusee: a literature data pipeline and research agent
Markovate: document extraction for insurance data collection
Sigma Solve: AI email processing in a freight management system, AI voice confirmation validation for a metals firm and an AI document analysis platform for EdTech
Spiral Scout: AI detection of personal data in handwritten letters
We read case studies for 8 of the 20 companies on this list, so the other 12 are not covered here. A case study that calls a system live is the company's own statement, not an audit. Ask each company which of those systems it still supports, and for a reference. In a random sample of 298 AI case studies from 181 software firms, 48% say they are live or in production.
Where do these companies deliver from?
All 20 companies on this AI list publish an office or delivery location in the United States. Other locations published most often are India (9), Poland (6) and Ukraine (5). 15 of the 20 publish more than one location.
Asia
Show or hide the companies for Asia
Nine of the 20 companies on this AI list deliver from Asia, all of them from India.
For a US buyer this is offshore work, with a working day many hours ahead of US time zones, so agree overlap hours.
What has each company published about its AI work?
Each profile links the public pages that show the company's AI work, with the date we captured each one, and says where the company may fit and what to check before hiring it. The 20 profiles run A to Z. Search or filter to narrow them.
No companies selected
Maximum three companies. Remove one to add another.
Your comparison
Rates are not listed, because few companies publish current terms.
83%29 of 35 public projects, counted on its portfolio, Nov 6, 2026
Case studies read
159 say the product is live
Most-named methods
LLM in 5 case studies, speech in 4, AI agent in 3, predictive model in 3
Applied AI studio in the Los Angeles area whose public work includes an eight-model airline pricing system, AI Spanish conversation practice for Pimsleur, a retrieval-based benefits qualification assistant and neural decoder work for a brain-computer interface maker.
Eight-model revenue management system giving flight-level pricing recommendations, with automated, validated hand-off into the airline's live pricing system
Project page · captured · Predictive models and analytics
AI benefits qualification assistant using retrieval over a disability rights handbook, with Salesforce transcripts for advisor review; piloted with one partner
Project page · captured · Chatbots and AI assistants
Self-supervised neural decoder training, real-time session analysis tools, automated multi-session analytics and Rust changes so the MoveAgain brain-computer interface platform can run the new models
Project page · captured · Predictive models and analytics
Show or hide fit, checks and technology for AE Studio
Where AE Studio may fit
Operators with high-volume pricing decisions that want predictive models connected to a live pricing system
Product and service teams adding a conversational AI feature, such as voice practice or a document-grounded assistant with human advisor review
Check before hiring AE Studio
The site gives a Marina del Rey, California address and says the firm is based in Los Angeles, but it names no delivery locations, so ask where your team would be based
Education technology is the largest sector in the case-study index, with 9 of its 33 entries, and at least three of those name the same client, Alpha School; ask for references outside education if that is not your sector
Technology focus
Predictive pricing models on Python, Celery and Snowflake
GPT-4o mini voice pipelines with speech-to-text and text-to-speech
Retrieval-augmented generation with Salesforce integration
Logistics platform for e-commerce and 3PL operators in India with live GPS tracking, predictive analytics for demand and delivery timelines, and order workflow automation
Project page · captured · Predictive models and analytics
Show or hide fit, checks and technology for Agile Infoways
Where Agile Infoways may fit
Startups building a platform with applied AI or automation features
Products combining AI with web or mobile delivery
Check before hiring Agile Infoways
AI depth varies by project
Published case-study titles provide limited evaluation detail
We have not published a count of this company's whole portfolio, so no share is shown.
Software engineering firm with public AI work covering agentic financial reconciliation, an AI investigation layer on a legacy .NET system, and computer-vision inspection with an investigating agent.
We have not published a count of this company's whole portfolio, so no share is shown.
Software development firm with public AI work spanning an insurtech app with an AI chatbot, a retail analytics dashboard using ChatGPT, and an AI emotional wellness companion app.
We have not published a count of this company's whole portfolio, so no share is shown.
Custom software firm with public AI work spanning an agentic email ticketing system, call center quality scoring with Azure OpenAI, and a conversational agent for marketing analysis.
45%5 of 11 public projects, counted on its portfolio, Nov 6, 2026
Case studies read
54 say the product is live
Most-named methods
NLP in 2 case studies, LLM in 1
AI and machine learning engineering firm whose public work shows an RNN-based drone navigation model with an automated telemetry pipeline and MLOps architecture for an aerospace and defense startup.
Shopify dropshipping store for 10,000+ products with a custom AI model that cleans supplier data and generates product listings, plus MLOps infrastructure
Project page · captured · AI features in business software
iOS bid-and-ask marketplace for sneakers and streetwear with AI models that generate product descriptions and enhance images, plus price collection from similar platforms
Project page · captured · AI features in business software
Show or hide fit, checks and technology for Flyaps
Where Flyaps may fit
Startups turning research documentation into a working ML model and MVP
Teams needing automated data pipelines and MLOps around a custom model
Check before hiring Flyaps
The three cited projects are a drone navigation model for aerospace and defense and two e-commerce product-listing systems, so confirm relevant experience in your domain
Accuracy and funding figures on the page come from client announcements and are not independently verified
21%15 of 73 public projects, counted on its portfolio, Nov 6, 2026
Case studies read
158 say the product is live
Most-named methods
chatbot in 5 case studies, NLP in 4, AI agent in 3, LLM in 3, predictive model in 3
AI and machine learning development company with public work spanning an AI recruitment platform, a scientific literature data pipeline with a research agent, and an AI candidate screening workflow.
We have not published a count of this company's whole portfolio, so no share is shown.
AI development company, acquired by The Hackett Group, with public LLM work spanning a dealer-support chatbot, an advertisement-generation SaaS platform, and a retrieval-based service-desk chatbot built on a ServiceNow knowledge base.
Cashierless checkout system for Canadian grocery stores with image-recognition product scanning, RFID and IoT inventory tracking, a smart cart and in-store navigation
AI inventory management system with demand forecasting, cluster analysis, automated stock-level alerts and a chatbot, replacing a manual dashboard for a retail client
Project page · captured · Predictive models and analytics
Show or hide fit, checks and technology for Markovate
Where Markovate may fit
Applied AI products tied to a defined workflow
Computer vision and document-processing systems
Check before hiring Markovate
Many outcome figures are vendor-published and need reference validation
Confirm data access and evaluation ownership before scope
Microsoft-native conversational AI knowledge assistant in Teams, using Power Automate AI and connected to SharePoint libraries, so credit union staff can find policy answers
Project page · captured · Chatbots and AI assistants
29%14 of 48 public projects, counted on its portfolio, Nov 6, 2026
Case studies read
143 say the product is live
Most-named methods
predictive model in 7 case studies, LLM in 5, recommendation in 3
Generative AI and custom software firm with work covering a multi-tenant GenAI SaaS platform and a maintenance assistant proof of concept for an automotive supplier.
Pilot and rollout of churn-prediction and product-recommendation models with a 360-degree customer view and retention suggestions for front-line staff at a Polish telecom
Project page · captured · Predictive models and analytics
AI chatbot inside the LeadStar affiliate platform answering publisher questions through retrieval-augmented generation over documentation and blog content stored as vector embeddings; MVP phase
Project page · captured · Chatbots and AI assistants
Show or hide fit, checks and technology for Neoteric
Where Neoteric may fit
Startups building generative AI SaaS products from prototype to launch
Enterprises validating a generative AI use case through workshops and a proof of concept
Check before hiring Neoteric
Case pages date from 2023, and the ZF Group assistant was a proof of concept sent to one plant for testing; confirm production experience
Outcome claims are reported by the firm and its clients
NLP chatbot on the Zephyr 7B open-source LLM that answers insurance inquiries, calculates risk profiles and generates policies, integrated with the insurer's CRM
Project page · captured · Chatbots and AI assistants
Computer vision system on a Jetson Nano with underwater cameras that counts fish moving upstream and downstream, identifies species and sends data to the cloud
We have not published a count of this company's whole portfolio, so no share is shown.
Software development firm with public AI work including an NLP-based AI interviewer for technical recruiting and deep learning for orthodontic treatment planning.
Modular agentic AI platform on microservices handling customer engagement and lead handoff for automotive dealerships, with internal marketing queries and merged dealership data sources
Project page · captured · AI agents and orchestration
21%8 of 38 public projects, counted on its portfolio, Nov 6, 2026
Case studies read
86 say the product is live
Most-named methods
LLM in 3 case studies, chatbot in 2, recommendation in 2
Enterprise AI and software engineering firm with public work including an AI-native patient access platform for NHS GP practices and an AI-driven freight operations platform.
Salesforce-integrated platform that transcribes customer voice confirmations with Whisper on Azure, extracts entities, validates required responses, and raises severity-based alerts on dashboards
Project page · captured · Document, email and claims automation
Cloud-native AI document analysis platform using ChatGPT-4.0 for conversational analysis and summaries, Pinecone semantic search, LangChain orchestration, and role and subscription management
Project page · captured · Document, email and claims automation
Show or hide fit, checks and technology for Sigma Solve
Where Sigma Solve may fit
Digital health companies building AI-assisted patient triage and booking with clinical system integration
Logistics operators replacing disconnected tools with a custom platform and AI-driven automation
Check before hiring Sigma Solve
Of Sigma Solve's 38 public projects, 8 are AI work, and the four cited span healthcare, freight, finance and education, so ask which AI projects the proposed team delivered.
None of the four cited cases names its client, and outcomes are vendor-reported, so ask for references
We have not published a count of this company's whole portfolio, so no share is shown.
Software development agency with public AI work including multi-agent orchestration on Temporal for an e-commerce support platform and an agentic runtime architecture for a banking platform.
Multi-agent conversational CRM system on Wippy that queries Salesforce account data in natural language and applies sales playbook logic, with a multi-tenant agent registry
Project page · captured · AI agents and orchestration
Show or hide fit, checks and technology for Spiral Scout
Where Spiral Scout may fit
Teams building multi-agent systems that need durable workflow orchestration on Temporal
Fintech and banking platforms designing AI agents with compliance and audit controls
Check before hiring Spiral Scout
The cited pages center on agent orchestration and architecture rather than model development, and the Staq page lists design outputs rather than a launched product
Outcome figures on the Gorgias page are reported by the firm; ask the client to confirm them
Rearchitected the Ava patient assistant as a LangGraph LLM agent on Claude, with retries, confidence-based escalation to human reviewers and a custom evaluation suite
Project page · captured · Chatbots and AI assistants
Agentic AI SMS system that answers patient inquiries from a healthcare corpus, routing queries to suitable Claude models, with staff reviewing responses before delivery
Project page · captured · AI agents and orchestration
13%10 of 78 public projects, counted on its portfolio, Nov 6, 2026
Case studies read
106 say the product is live
Most-named methods
AI agent in 3 case studies, chatbot in 3, LLM in 3
Enterprise AI and product engineering firm with public work including a Model Context Protocol layer that makes a legacy construction ERP queryable by AI agents.
Flutter and Node.js news app using OpenAI models for article summaries, translation into Indian languages, audio briefs, Junior Mode content curation and a story chatbot
Project page · captured · Generative AI products
Show or hide fit, checks and technology for TechAhead
Where TechAhead may fit
Companies with legacy ERP systems that want AI agents to query live business data safely
Teams exposing existing business systems to AI agents without changing the underlying software
Check before hiring TechAhead
One cited case describes an integration layer for third-party AI agents rather than custom model development
Case page figures are reported by the firm; request references for production AI work
We have not published a count of this company's whole portfolio, so no share is shown.
Digital engineering company with public AI work including an asset tracking platform with an LLM and RAG assistant, an AI financial planning platform, and an AI-powered mental health app.
We have not published a count of this company's whole portfolio, so no share is shown.
Product engineering firm with public AI work covering insurance claims automation, machine learning analytics on health plan claims data, and a diagnostic imaging platform that uses image-analysis models.
Standalone diagnostic imaging platform with a DICOM viewer, CNN model integration for image analysis, clinician review workflows and PACS interoperability
Project page · captured · Computer vision
Show or hide fit, checks and technology for Zymr Systems
Where Zymr Systems may fit
Insurers automating claims intake, fraud checks, and settlement workflows with AI
Health plans consolidating fragmented claims data for machine learning analytics
Check before hiring Zymr Systems
All three case pages leave the client unnamed, so references must be requested directly
Outcome figures on all three pages are reported by the firm, and none of the pages names the languages, frameworks or cloud services used
Technology focus
Machine learning models
AI document processing
Cloud-based analytics
Fraud and risk scoring
None of the companies on this list match those filters. A suitable company may still exist outside this list.
How do you choose an AI development company?
Choose an AI development company by matching one of its documented projects to your use case, then checking how much evidence it has, whether that work runs in production, and who owns evaluation and support. Shortlist two or three firms and compare written proposals, not demos.
How much public evidence does each company have?
Of the 20 companies on this AI list, 19 link two or more project pages and 1 links one. A single case study is one instance of the work as the company reports it, so before you shortlist, ask for a second reference from a project like yours. Across the 870 software firms whose public pages we captured on Nov 5 and 6, 2026, 434 show at least one AI project. Of those, 77 have five or more public projects in total, with 30% or more of them in AI. Where we publish a count of a firm's whole portfolio, its card shows the AI share.
What should you verify beyond an AI demo?
Ask about data access, evaluation, failure handling, latency and cost constraints, security, and monitoring ownership. Where a firm delivers from does not show what it costs, and this page publishes no performance benchmarks.
Who owns evaluation and production support?
If a public page does not say who runs evaluation or production support, ask the company directly: a type-of-work label on this page says what was built, not who runs it. The lines on each card about where a company may fit are editorial judgments, not verified performance scores.
What should you ask an AI development company before shortlisting it?
Public case studies rarely show who owns evaluation, production support, or the cost of running a model, so ask each company these four questions before you shortlist it, and expect the answers in its written proposal.
What data would they need, and who owns the evaluation criteria?
Which of their AI projects run in production today, rather than as a pilot or demo?
How are failures, latency, cost, and security handled after launch?
Who owns handover, monitoring, and production support?
How were these companies selected?
A company is listed only when its own public pages name a relevant AI project, on a project page or a portfolio index; an industry label alone is not enough. Peter Korpak, founder of 100Signals, reviews every listing.
This list is not exhaustive, and each company appears on one industry page only. A company missing here may be on the fintech or healthcare page, may not have been reviewed yet, or may not have met our criteria.
We started from 721 firms that list AI work. Of those, 245 had a portfolio entry that looked like AI work, we confirmed a project page showing AI work for 128, and 20 are on our shortlist so far.
These counts are a minimum: a firm we did not count may still have relevant work on its site.
The market figures on this page describe the 870 software firms whose public pages we captured on Nov 5 and 6, 2026, not the whole market. The sample percentages above come from a random draw of 298 case studies per industry on Nov 6, 2026 and describe only those case studies.
Many lists on this topic are published by companies that sell to the same buyers who read them. Of the 13 such lists we read on Nov 5, 2026 for healthcare, fintech and AI development searches, 12 listed their own publisher first. 100Signals is not listed here.
Each card shows the date its listing was last reviewed. 100Signals is not itself a software vendor in these categories: it is a demand generation agency for software development firms that sells marketing services to those firms and may have commercial relationships that are separate from this list.
Peter Korpak started as a financial analyst at Credit Suisse and Aviva Investors and has spent 8 years building pipeline for software development agencies. About Peter
What do the project sources prove?
A project page describes one engagement; a portfolio index only names projects, so it supports a narrower claim. Neither is independent proof of delivery, and a standard named in a case study is not a certification. Capture dates show when a page was saved, not when the work happened, and outcome figures a company reports about itself are not presented here as measured results.
How can a listing be corrected?
If a listed company needs a factual correction, email peter@100signals.com. Each correction is checked against the company's own pages, and a review date changes only when the whole listing is reviewed again.
Can Peter Korpak review your shortlist?
Yes, if you want a second opinion. Peter Korpak at 100Signals reads shortlists sent through this optional form and replies by email. The companies you select are not contacted automatically.
