The Agent Stack: AI Spend & ROI contains 17 companies across
3 categories.
This landscape covers startups that help companies understand and control the economics of AI systems and agents. These products show where AI spend comes from and whether that spend creates enough business value.
AI Cost Management
Software that helps companies understand and manage AI spending. It tracks usage costs and supports decisions that reduce unnecessary spending.
AgentCost
Website: agentcost.tech
AgentCost is an open source cost observability tool for AI systems that attributes LLM spend to the agent, workflow, project, or model that caused it.
Category: AI Cost Management.
Funding stage: Bootstrapped.
Founded: 2025.
Employees: 1-10.
Landscape fit: It fits AI cost attribution because the core product traces LLM calls to the specific agent or workflow that generated the cost and shows spend breakdowns by agent, project, and model.
Company timeline
-
2026-10-04:
HR snapshot
(hr)
. AgentCost identifies Kushagra Agrawal as founder and maintainer and offers direct founder contact. The supplied evidence does not establish headcount or active job openings.
Source: About AgentCost.
https://agentcost.tech/about
-
2026-10-04:
Product snapshot
(product)
. AgentCost describes an open-source LLM cost observability platform with a Python SDK, analytics dashboard and API, budget controls, reports, local analysis CLI, model catalog, MCP tools, and self-hosted deployment.
Source: About AgentCost.
https://agentcost.tech/about
-
2026-10-01:
Free cloud pricing launches
(gtm)
. AgentCost positioned its hosted cloud as free and said self hosting is optional.
Source: AgentCost pricing page.
https://agentcost.tech/pricing
-
2026-09-08:
Per-agent analytics pages released
(product)
. AgentCost added dedicated agent pages and an Agents list with spend changes, trends, cost per run information, and guardrail state.
Source: Changelog | AgentCost.
https://agentcost.tech/changelog
-
2026-09-05:
Per-agent guardrails and compliance view released
(product)
. AgentCost added a Guardrails page for per-agent boundaries, compliance checks, and breach notifications.
Source: Changelog | AgentCost.
https://agentcost.tech/changelog
-
2026-08-11:
AgentCost adds workflow tracing and pre-deployment cost analysis
(product)
. AgentCost released workflow-level cost tracing, cost per completed outcome, and a local pre-deployment analyzer that can flag expensive agent behavior before production.
Source: AgentCost changelog.
https://agentcost.tech/changelog
FinOps LLM
Website: finopsllm.com
FinOps LLM helps teams attribute and govern LLM and GenAI spend so they can trace costs back to the workloads, teams, or features that caused them.
Category: AI Cost Management.
Founded: 2026.
Country: United States.
Employees: 1-10.
Landscape fit: It fits AI cost attribution because the product is built to reconcile invoices and usage data, assign spend to the responsible workload, and make AI costs traceable at the request and team level.
Company timeline
-
2026-10-04:
Verified savings pricing
(gtm)
. FinOps LLM prices implementation and optimization as a share of verified monthly savings. The pricing page also sets a minimum managed engagement of $20,000 per month in LLM spend.
Source: Pricing · FinOps LLM.
https://finopsllm.com/pricing
-
2026-10-04:
HR snapshot
(hr)
. The careers page states that there are no open roles and no standing headcount plan. It invites targeted outreach and mentions possible contract or advisory work. The supplied evidence does not establish headcount.
Source: Careers · FinOps LLM · No Open Roles, Notes Welcome.
https://finopsllm.com/careers
-
2026-10-04:
Free audit motion
(gtm)
. FinOps LLM offers a free, human led audit that starts from read only billing and usage data. The booking flow positions the audit as the entry point to an implementation engagement.
Source: Book · FinOps LLM.
https://finopsllm.com/book
-
2026-10-04:
Product snapshot
(product)
. FinOps LLM describes managed AI cost management and LLM observability, including invoice reconciliation, token-level attribution, anomaly detection, chargeback/showback, routing, caching, and compression. Tidal Telemetry provides hosted LLM traces and usage reporting as a separate beta product. The website also hosts a public AI model release tracker.
Source: FinOps LLM - AI cost management for the era of tokens.
https://finopsllm.com
-
2026-10-04:
AI model release tracker shipped
(product)
. FinOps LLM launched a public page for tracking new GPT and Claude model releases. It is a distinct public tool rather than a general research article.
Source: AI Model Release Tracker: New GPT & Claude Models · FinOps LLM.
https://finopsllm.com/tools/llm-release-tracker
-
2026-10-03:
No open roles
(hr)
. FinOps LLM says it has no open roles and no standing headcount plan. It still invites targeted outreach for relevant experience and possible contract or advisory work.
Source: Careers · FinOps LLM · No Open Roles, Notes Welcome.
https://finopsllm.com/careers
Onaro
Website: onaro.io
Onaro builds Meridian, software that tracks AI agent costs and assigns them to owners and workflows for finance teams.
Category: AI Cost Management.
Country: United States.
Employees: 11-50.
Landscape fit: Meridian attributes AI spending to agents, owners, and workflows, making the costs traceable to the work that generated them.
Company timeline
-
2026-10-01:
Meridian automates GPU capacity records
(product)
. Meridian began building hourly GPU capacity records daily and rebuilding recent hours with samples to include late-arriving data.
Source: Onaro.
https://www.onaro.io/changelog
-
2026-10-01:
Meridian adds self-serve ingest keys
(product)
. Organization administrators can create, view, and revoke Meridian ingest keys in Settings without contacting Onaro support.
Source: Onaro.
https://www.onaro.io/changelog
-
2026-10-01:
Onaro lists a Founding Account Executive opening
(hr)
. Onaro advertised a Founding Account Executive role and described it as the company's first sales hire.
Source: LinkedIn Jobs.
https://www.linkedin.com/jobs/view/founding-account-executive-at-onaro-io-4463512830
-
2026-09-01:
Meridian adds a GPU capacity feed
(product)
. The feed accepts NVIDIA DCGM metrics from Prometheus or OpenTelemetry and allocates GPU capacity costs across endpoints and agents.
Source: Onaro.
https://www.onaro.io/docs/guides/capacity-feed-meridian
-
2026-09-01:
Meridian adds LiteLLM gateway ingest
(product)
. A shipped integration sends LiteLLM Generic API callback logs to Meridian, mapping calls to OASA usage records and the spend ledger.
Source: Onaro.
https://www.onaro.io/docs/guides/litellm-meridian
-
2026-09-01:
Onaro publishes the OASA specification
(product)
. Onaro published the Open Agent Spend Attribution specification, an open record format for joining agent identity, tasks, cost objects, telemetry, billing, and settlement in an auditable ledger.
Source: Onaro.
https://www.onaro.io/spec
AI SpendOps
Website: aispendops.com
AI SpendOps builds a proxy based platform that tracks AI usage, attributes cost, enforces spend policies, and produces finance grade reporting for AI API calls.
Category: AI Cost Management.
Founded: 2026.
Country: United Kingdom.
Employees: 1-10.
Landscape fit: It fits runtime spend control because its proxy sits between applications and AI providers, and it can enforce policies, track usage, and alert on or limit budget use while requests are running.
Company timeline
-
2026-10-04:
HR snapshot
(hr)
. Companies House lists one active director for AI SPENDOPS LTD. This is not evidence of total employee count. No clear public hiring signal is established by the supplied evidence.
Source: AI SPENDOPS LTD people.
https://find-and-update.company-information.service.gov.uk/company/17046015/officers
-
2026-10-04:
Product snapshot
(product)
. AI SpendOps is a proxy-based AI spend management platform. Its official website and documentation describe usage tracking, cost attribution, policy enforcement, budget alerts, and reporting for AI API calls.
Source: official website.
https://www.aispendops.com
-
2026-10-01:
Self serve access announced
(gtm)
. AI SpendOps says the service is available to customers in the United States and worldwide, with self serve signup available.
Source: FAQ.
https://www.aispendops.com/faq
-
2026-09-01:
LLM gateway comparison published
(gtm)
. AI SpendOps published a comparison of LLM gateways and placed its product in the cost attribution and governance category.
Source: The Best LLM Gateways in 2026, Sorted by the Job You Need Done.
https://www.aispendops.com/blog/best-llm-gateways-2026
-
2026-08-01:
Vercel AI Gateway support added
(product)
. AI SpendOps announced support for Vercel AI Gateway. Teams can route traffic through AI SpendOps in front of the gateway for cost attribution, budgets, and governance while keeping the existing gateway key.
Source: AI SpendOps Now Supports Vercel AI Gateway.
https://www.aispendops.com/blog/vercel-ai-gateway-support
-
2026-08-01:
Budget alerts launched
(product)
. AI SpendOps announced budget alerts as live for all plans. Budgets can be scoped to an organisation, API key, provider, or dimension value, with email warnings at configured thresholds.
Source: Budget Alerts Are Now Live in AI SpendOps.
https://www.aispendops.com/blog/budget-alerts-now-live
Tokenomy
Website: tokenomy.ai
Tokenomy builds runtime FinOps software for AI systems. It meters AI usage, enforces budgets and policies before spend happens, and helps teams attribute and control model cost as systems run.
Category: AI Cost Management.
Founded: 2025.
Country: United States.
Employees: 1-10.
Landscape fit: It fits runtime spend control because the product is explicitly built to enforce budgets, routing, and circuit breakers in the request path while AI is running.
Company timeline
-
2026-10-04:
HR snapshot
(hr)
. The official about page names Mahesh Lambe as founder. No employee count or verifiable active hiring openings are established by the supplied fetched text.
Source: About & Team.
https://www.tokenomy.ai/about
-
2026-10-04:
Product snapshot
(product)
. Tokenomy describes an AI spend control and runtime FinOps platform for LLM and agent workloads. The supplied official materials describe metering, budget controls, routing, caching, chargeback, invoice runs, API access, and research and free-tool offerings.
Source: Tokenomy homepage.
https://www.tokenomy.ai
-
2026-09-01:
Tokenomy publishes AI economics academy
(product)
. Tokenomy launches an academy surface for AI economics education.
Source: AI Economics Academy.
https://tokenomy.ai/academy
-
2026-09-01:
Tokenomy publishes security posture
(product)
. Tokenomy adds a trust page that presents its security and compliance posture for the runtime product.
Source: Trust, security and compliance.
https://www.tokenomy.ai/trust
-
2026-09-01:
Tokenomy opens public pricing
(product)
. Tokenomy publishes paid pricing tiers on its official pricing page.
Source: Pricing page.
https://www.tokenomy.ai/pricing
-
2026-09-01:
Tokenomy adds API access
(product)
. Tokenomy’s API access page describes a key for live and historical model prices and authentication for the metering runtime.
Source: API Access.
https://www.tokenomy.ai/api-access
FinChop
Website: finchop.io
FinChop is a private beta platform that reduces AI spend by routing each task to the lowest cost model that can still meet quality needs and by showing ROI across workflows.
Category: AI Cost Management.
Country: Unknown.
Employees: 1-10.
Landscape fit: It fits this category because it changes how AI work is executed to lower cost, mainly by routing requests to cheaper models when acceptable results are likely.
Company timeline
-
2026-10-04:
HR snapshot
(hr)
. The supplied homepage evidence provides no direct hiring or team signal.
Source: FinChop homepage.
https://finchop.io
-
2026-10-04:
Product snapshot
(product)
. The supplied homepage evidence describes enterprise AI workflow ROI visibility and cost-oriented model routing, including cloud and self-hosted models.
Source: FinChop homepage.
https://finchop.io
Tokenistt
Website: tokenistt.com
Tokenistt is an AI FinOps platform that helps enterprises track, govern, and reduce LLM spend by routing requests across providers and optimizing execution paths.
Category: AI Cost Management.
Funding stage: Pre Seed.
Country: India.
Employees: 1-10.
Landscape fit: It fits cost optimization and routing because the product explicitly chooses cheaper model and provider paths, caches or reroutes requests, and reduces AI spend when a lower cost path can meet the quality bar.
Company timeline
-
2026-10-04:
HR snapshot
(hr)
. The supplied careers-page summary describes Tokenistt as a small, remote-first, pre-seed team, with five people on the founding team and two open roles. This is current-state evidence rather than a dated hiring event.
Source: Tokenistt careers page.
https://tokenistt.com/careers
-
2026-10-04:
Product snapshot
(product)
. Tokenistt describes an AI FinOps control plane for enterprise LLM spend. Official machine-readable company materials list monitoring, cost management, governance, routing, optimization, cache intelligence, and developer-tool integrations.
Source: Tokenistt homepage.
https://tokenistt.com
-
2026-07-01:
YC application announced
(gtm)
. Tokenistt said it had applied to Y Combinator Summer 2026.
Source: Tokenistt company post.
https://www.linkedin.com/posts/tokenistt_weve-been-a-little-quiet-over-the-last-few-activity-7459357857356963840-ch6b
-
2026-07-01:
Early workflow rollout and waitlist interest
(gtm)
. Tokenistt said it was rolling the product into early workflows and reported more than 300 developers on its waitlist in under 24 hours. The post also said teams were reaching out inbound.
Source: Tokenistt early traction post.
https://www.linkedin.com/posts/tokenistt_tokenistt-ai-finops-platform-for-enterprise-activity-7455111394712129536-3nuq
-
2026-07-01:
Execution layer build update
(product)
. Tokenistt said it was building an MCP based execution layer and improving real time cost estimation before prompts are sent. It also said it had started testing cache return on investment.
Source: Tokenistt build update.
https://www.linkedin.com/posts/tokenistt_starting-the-week-with-a-quick-build-update-activity-7456897984006602753-UA4Z
AI Financial ROI & Unit Economics
Software that compares AI costs with financial outcomes. It helps companies understand whether AI work generates enough economic value to justify its cost.
Husk
Website: usehusk.com
Husk is a finance platform that helps startups and growing teams manage company spending, cash flow, and AI spend economics with cards, controls, and real time insights.
Category: AI Financial ROI & Unit Economics.
Funding stage: Pre Seed.
Founded: 2023.
Country: Belgium.
Employees: 2-10.
Landscape fit: It fits AI ROI and unit economics because the product is built to show what spend is driving value, compare costs against outcomes, and help teams control or justify AI related spending.
Funding rounds
-
Pre-seed
announced 2025-05-27
: 1000000 EUR
; investors: Techstars, Birdhouse Ventures, NewSchool.vc, angel investors
; source: Husk
; https://www.usehusk.com/blog/husk-raises-eu1m-pre-seed-to-expand-its-growth
Company timeline
-
2026-10-04:
HR snapshot
(hr)
. Husk's LinkedIn company page reports 2–10 employees. Its careers page says there are no open roles right now.
Source: Husk LinkedIn company page.
https://www.linkedin.com/company/usehusk
-
2026-10-04:
Product snapshot
(product)
. Husk's current public product scope includes corporate debit and virtual cards, card and spending controls, expense management, transaction and cash-flow insights, and AI spend visibility. Bill pay is shown as coming soon.
Source: Husk home page.
https://www.usehusk.com
-
2025-12-01:
WhatsApp receipt matching
(product)
. Husk launched a receipt matching flow through WhatsApp. Users can send a receipt photo to Husk, which identifies the transaction and posts the matched expense to the dashboard.
Source: Husk blog.
https://www.usehusk.com/blog/whatsapp-receipt-matching
-
2025-05-27:
Husk raises €1M pre-seed
(funding)
. Husk closed a €1M pre-seed round to expand its financial platform for European startups.
Source: Husk.
https://www.usehusk.com/blog/husk-raises-eu1m-pre-seed-to-expand-its-growth
-
2025-02-01:
First dedicated sales role
(hr)
. Husk posted a Brussels based Sales Development Representative role and described it as its first dedicated sales representative position.
Source: Husk sales development representative job.
https://be.linkedin.com/jobs/view/sales-development-representative-at-husk-4299973833
-
2025-02-01:
First customers and sales hire
(gtm)
. Husk said it had launched its product and onboarded its first customers while hiring its first dedicated sales representative.
Source: Husk sales development representative job.
https://be.linkedin.com/jobs/view/sales-development-representative-at-husk-4299973833
Cogs'z
Website: cogsz.com
Cogs'z is a cloud software platform that helps SaaS and AI companies track profitability, allocate costs, and measure AI spend against business outcomes.
Category: AI Financial ROI & Unit Economics.
Founded: 2022.
Country: United States.
Employees: 1-10.
Landscape fit: It fits this category because it is software for measuring AI spend and unit economics, which directly supports AI ROI and margin analysis for customers.
Company timeline
-
2026-10-04:
HR snapshot
(hr)
. The fetched evidence does not verify employee count or current recruiting activity.
Source: HR source.
-
2026-10-04:
Product snapshot
(product)
. Cogs'z describes a cloud software platform for SaaS and AI businesses focused on unit-level profitability and margin improvement. Its published scope includes cost allocation, customer and product profitability, pricing and quoting, contract management, vendor spend, and AI spend visibility and control.
Source: Cogs'z | The Future of Profitability - Cogs'z.
https://cogsz.com
-
2026-09-04:
Profitability platform pages expanded
(product)
. Cogs'z expanded its official product pages to describe customer and product profitability capabilities for SaaS and AI businesses.
Source: Customer Product Profitability - Cogs'z.
https://cogsz.com/customer-product-profitability
-
2026-09-01:
Platform modules clarified
(product)
. Cogs'z clarified the platform structure on its how it works page.
Source: How Cogs'z works.
https://cogsz.com/how-it-works
-
2026-05-01:
Company post invites demos
(gtm)
. Cogs'z posted about its profitability platform on LinkedIn and invited viewers to book a demo.
Source: Cogs'z post on SaaS profitability.
https://www.linkedin.com/posts/cogsz_saas-profitability-financeleadership-activity-7384977331494211584-l6rU
Floe
Website: floefinance.com
Floe is software that tracks AI spend against the revenue or outcome it produced, so teams can see margin, enforce budgets, and bill by task or client.
Category: AI Financial ROI & Unit Economics.
Funding stage: Pre Seed.
Founded: 2025.
Country: United States.
Employees: 2-10.
Landscape fit: It fits AI ROI and unit economics because it directly compares AI costs with the business outcomes and margin they generate.
Funding rounds
-
pre-seed
announced 2026-04-30
: 1000000 USD
; source: Wellfound
; https://wellfound.com/company/floe-labs/funding
Company timeline
-
2026-10-06:
Floe is hiring across engineering and growth
(hr)
. Floe is recruiting for financial data infrastructure, developer relations and growth roles as it expands the product and go-to-market team.
Source: Wellfound.
https://wellfound.com/company/floe-labs
-
2026-10-06:
Early product usage reported
(gtm)
. Floe's founding engineer reports 400 users and 746 wallets onboarded across the product, alongside live production use of the spend infrastructure.
Source: Raj Bhensdadiya.
https://www.rajbhensdadiya.com/projects
-
2026-10-06:
HR snapshot
(hr)
. Wellfound's supplied summary lists Floe Labs at 1–10 employees, and a supplied job-board summary describes a Data / ML Engineer opening. The available evidence suggests a small team and a technical hiring signal.
Source: Wellfound Floe Labs people page.
https://wellfound.com/company/floe-labs/people
-
2026-10-06:
Product snapshot
(product)
. Floe presents software and infrastructure for measuring AI-agent costs, associating spend with tasks and customers, applying budgets, and routing agent payments. Its public materials also describe billing from actual usage and working-capital products.
Source: Floe official website.
https://www.floefinance.com
-
2026-09-01:
Open source guardrails launch
(product)
. Floe published floe-guard as a local spend meter and budget gate for agent workflows.
Source: floe-guard GitHub repository.
https://github.com/Floe-Labs/floe-guard
-
2026-07-01:
Voice metering metrics are published
(product)
. Floe published measured voice agent overhead and timing data for its metering approach.
Source: What Floe costs a voice agent in millisecond methodology, comparison, and why the tail matters more than the median.
https://www.floefinance.com/resources/what-floe-costs-a-voice-agent-in-millisecond-methodology-comparison-and-why-the-tail-matters-more-than-the-median.
CogsLayer
Website: cogslayer.com
CogsLayer helps companies track AI spending, adoption, and unit economics so they can assess costs and identify savings.
Category: AI Financial ROI & Unit Economics.
Employees: 1-10.
Landscape fit: It measures AI spend by team, employee, and customer, and provides unit economics and savings tracking to help companies evaluate AI’s business value.
Spendline
Website: spendline.ai
Spendline helps B2B software companies track AI costs by customer and workflow, compare costs with revenue, and enforce spending budgets.
Category: AI Financial ROI & Unit Economics.
Founded: 2026.
Country: United States.
Employees: 2-10.
Landscape fit: It connects AI usage costs with customer revenue and budget controls, helping companies assess whether AI powered workflows justify their cost.
Company timeline
-
2026-10-12:
Spendline publishes a one-week AI cost-to-serve audit guide
(gtm)
. Spendline published a free step-by-step guide for measuring AI cost per customer over one week, including fields to log, queries, a worked example, and common errors.
Source: Spendline Guides.
https://www.spendline.ai/guides
-
2026-10-01:
Spendline lists a product engineering internship
(hr)
. A LinkedIn Jobs listing seeks a part-time product engineering intern in India to work with the founders on product development and engineering.
Source: LinkedIn Jobs.
https://in.linkedin.com/jobs/view/product-engineering-intern-at-spendline-4483335684
AI Productivity & Impact
Software that measures how AI affects work performed by employees. It connects AI use to observed output so companies can assess productivity gains.
Larridin
Website: larridin.com
Larridin helps enterprises measure AI spend against productivity and workflow outcomes to show where AI is creating value.
Category: AI Productivity & Impact.
Funding stage: Series A.
Founded: 2024.
Country: United States.
Employees: 20.
Landscape fit: It directly compares AI costs with the work and business value produced, helping companies assess AI return on investment.
Funding rounds
-
Series A
announced 2026-11-02
: 30000000 USD
; investors: Pelion Ventures, Andreessen Horowitz (a16z), Gradient Ventures, GFT, W23, Thayer, DG Daiwa Ventures, Alumni Ventures, Cougar Capital
; source: Larridin
; https://larridin.com/press-releases/larridin-raises-30-million-in-series-a-to-prove-ais-impact-across-the-human-and-agent-workforce
-
Seed
announced 2025-04-02
: 17000000 USD
; investors: Andreessen Horowitz (a16z), Gradient Ventures, Bloomberg, Haystack, Homebrew
; source: Larridin
; https://larridin.com/blog/the-larridin-story
Company timeline
-
2026-11-02:
Larridin reports 50 enterprise customers
(gtm)
. CEO Russ Fradin told Business Insider that Larridin had grown from one enterprise customer in December 2025 to 50 at the time of the report. The company also reported measuring more than 2 billion AI sessions and tracking more than 82,000 workflows.
Source: The Next Web.
https://thenextweb.com/news/larridin-30m-series-a-ai-spending-measurement
-
2026-11-02:
Series A
(funding)
. Amount: USD 30M. Investors: Pelion Ventures, Andreessen Horowitz (a16z), Gradient Ventures, GFT, W23, Thayer, DG Daiwa Ventures, Alumni Ventures, Cougar Capital
Source: Larridin.
https://larridin.com/press-releases/larridin-raises-30-million-in-series-a-to-prove-ais-impact-across-the-human-and-agent-workforce
-
2026-11-01:
Larridin lists two active roles
(hr)
. The a16z jobs page listed an AI Native Software Engineer role and a B2B Marketing Expert role for Larridin in the San Francisco Bay Area.
Source: Andreessen Horowitz Jobs.
https://jobs.a16z.com/jobs/larridin?skills=product+design
-
2026-09-30:
Larridin products become available on Google Cloud Marketplace
(gtm)
. Larridin announced that its products are available through Google Cloud Marketplace, giving enterprise customers another route to access the platform.
Source: Larridin.
https://larridin.com/press-releases/larridin-announces-a-new-collaboration-with-google-cloud-to-help-enterprises-measure-and-maximize-the-impact-of-ai
-
2026-06-23:
Larridin launches Token Spend & Insights
(product)
. Larridin announced a product for enterprise AI spending visibility, with cost information and usage analysis across employees, teams, projects, and departments.
Source: Larridin.
https://larridin.com/press-releases/new-larridin-token-spend-insights-solves-enterprise-ai-budget-crisis-as-spending-surges-13x
-
2026-03-01:
Larridin updates its measurement dashboard
(product)
. The mid-March 2026 software release added dashboard visibility into AI adoption and organized adoption metrics into reach and depth.
Source: Larridin.
https://larridin.com/blog/larridin-release-update-mid-march-2026
Ascerta
Website: ascerta.com
Ascerta makes software that measures AI cost, adoption, and business value so enterprises can tell which AI initiatives are worth scaling.
Category: AI Productivity & Impact.
Funding stage: Series A.
Founded: 2024.
Country: United States.
Employees: 15.
Landscape fit: It fits AI ROI and unit economics because the product is built to connect AI usage and cost to measurable business value, so teams can judge whether an agent or workflow earns its keep.
Funding rounds
-
Series A
announced 2026-09-30
: 18100000 USD
; investors: Dell Technologies Capital, Hitachi Ventures, Wipro, BGV, Gaia Ventures, FUSE, Tola Capital
; source: The Journey to Series A | Ascerta
; https://ascerta.com/resources/blogs/the-journey-to-series-a
-
Seed
announced 2025-05-01
: 4900000 USD
; source: Ascerta raises $18M Series A to help enterprises maximize AI ROI
; https://ascerta.com/resources/press/ascerta-raises-18m-series-a
Company timeline
-
2026-10-04:
HR snapshot
(hr)
. No sufficiently verified employee-count or recruiting signal is available from the supplied evidence.
Source: HR source.
-
2026-10-04:
Product snapshot
(product)
. Ascerta presents an enterprise platform for analyzing AI costs and activity and relating AI use cases to business value. Its materials describe use-case and KPI tracking, AI adoption and capacity analysis, and coding-agent activity insights.
Source: Ascerta home page.
https://ascerta.com
-
2026-09-30:
Customer savings case
(gtm)
. GeekWire reported a customer example in which Ascerta helped an insurer cut capacity costs.
Source: GeekWire.
https://www.geekwire.com/2026/pay-i-rebrands-as-ascerta-and-raises-18m-to-help-enterprises-track-ai-business-value
-
2026-09-30:
Series A
(funding)
. Amount: USD 18.1M. Investors: Dell Technologies Capital, Hitachi Ventures, Wipro, BGV, Gaia Ventures, FUSE, Tola Capital
Source: The Journey to Series A | Ascerta.
https://ascerta.com/resources/blogs/the-journey-to-series-a
-
2026-09-17:
Pay-i becomes Ascerta
(product)
. The company said it changed its name from Pay-i to Ascerta and broadened its platform beyond AI cost management.
Source: We're Now Ascerta | The Next Chapter for Pay-i.
https://ascerta.com/resources/blogs/pay-i-is-now-ascerta
-
2025-05-01:
Seed
(funding)
. Amount: USD 4.9M
Source: Ascerta raises $18M Series A to help enterprises maximize AI ROI.
https://ascerta.com/resources/press/ascerta-raises-18m-series-a
Weave
Website: weaveos.com
Weave measures how AI use affects engineering output, quality, and cost, helping companies assess AI return and route coding tasks to cost efficient models.
Category: AI Productivity & Impact.
Funding stage: Series A.
Founded: 2024.
Country: United States.
Employees: 16.
Landscape fit: It directly connects AI token spend with engineering output and quality to help teams judge whether AI costs are producing value. ([weaveos.com](https://weaveos.com/))
Funding rounds
-
Series A
announced 2026-07-28
: 13500000 USD
; investors: Standard Capital, Y Combinator, Moonfire, Burst Capital, IrregEx, the Agent Fund
; source: Weave
; https://weaveos.com/blog/weave-raises-13.5m-series-a-to-build-engineering-and-token-intelligence
-
Seed
announced 2025-07-21
: 4200000 USD
; investors: Moonfire, Burst Capital, Y Combinator
; source: Weave
; https://weaveos.com/blog/announcing-our-4-2m-seed-round
Company timeline
-
2026-11-01:
Weave open sourced Checks
(product)
. Weave Checks applies Markdown defined standards to pull requests and leaves review comments when a standard is not met.
Source: Adam Cohen on LinkedIn.
https://www.linkedin.com/in/adam-b-cohen
-
2026-11-01:
Zach Garrett joined Weave as founding product engineer
(hr)
. Andrew Churchill announced Zach Garrett's arrival at Weave as a founding product engineer.
Source: Andrew Churchill on LinkedIn, reposted by Adam Cohen.
https://www.linkedin.com/in/adam-b-cohen
-
2026-11-01:
Weave listed ten active roles
(hr)
. Weave's Y Combinator jobs page showed ten roles with application links across commercial, product, and engineering functions.
Source: Y Combinator.
https://www.ycombinator.com/companies/weave-3/jobs
-
2026-09-17:
Bytes AI reported higher engineering output after expanding Weave access
(gtm)
. Bytes AI said engineering output rose about 30% in six weeks after it opened Weave profiles to its whole engineering team. The story reports an observed outcome, not proof that Weave caused it.
Source: Weave Customer Stories.
https://weaveos.com/customers
-
2026-09-09:
Weave released Router 2.0
(product)
. Router 2.0 added cache aware switching, escalation to a frontier model, and routing across multiple provider subscriptions.
Source: Weave.
https://weaveos.com/blog/introducing-weave-router-2-0
-
2026-09-01:
Weave reported Router usage across 52 installations
(gtm)
. Weave's Q3 report says Router handled about 1.62 million requests across 52 installations and 234,000 agent sessions in September 2026.
Source: Weave Research.
https://weaveos.com/research/q3-2026
Mentlio
Website: mentlio.com
Mentlio measures AI spend against engineering output and helps teams reduce token costs through intelligent model routing and workflow analysis.
Category: AI Productivity & Impact.
Funding stage: Acquired.
Founded: 2026.
Country: United States.
Employees: unknown.
Landscape fit: It directly compares AI costs with engineering output and productivity to help companies assess and improve AI return on investment.
Company timeline
-
2026-10-09:
Mentlio founders join Factory
(hr)
. AlphaSignal reports that Mentlio founders Ashank Shah and Ahmet Demirbas joined Factory following its acquisition of Mentlio. Factory's announcement names both founders and says Mentlio's team will bring its work into Factory.
Source: AlphaSignal.
https://alphasignal.ai/news/factory-acquires-mentlio-to-cut-coding-agent-costs-by-30
-
2026-10-09:
Factory announces agreement to acquire Mentlio
(acquisition)
. Factory announced that Mentlio was joining its autonomous software engineering platform. Neither source discloses deal terms, and the supplied evidence does not confirm closing.. Acquirer: Factory. Status: announced
Source: Factory.
https://factory.com/news/mentlio
-
2026-09-01:
Mentlio updates its AI ROI guide with Delivery Points methodology
(product)
. Mentlio's guide describes Delivery Points as a normalized unit of shipped engineering work scored from repository evidence, and explains that AI attribution is observational rather than proof that AI caused each delivery. The guide also reports a frozen Terminal-Bench 2.1 replay, not a production study.
Source: Mentlio.
https://www.mentlio.com/guides/ai-roi-software-engineering
-
2026-07-01:
Mentlio publishes SWE-Bench Pro model-routing replay
(product)
. Mentlio published a replay benchmark reporting that Route retained 96.32% route sufficiency and lowered average model cost by 20.66% versus Always Fable. The company states that this was a replay, not a live production trial.
Source: Mentlio.
https://www.mentlio.com/research/fable-era-routing
Nebuly
Website: nebuly.com
Nebuly measures how enterprise AI agents affect work and calculates the value they create against their operating costs.
Category: AI Productivity & Impact.
Funding stage: Series A.
Founded: 2022.
Country: Italy.
Employees: 11-50.
Landscape fit: It analyzes AI agent use and task outcomes to estimate time saved, team proficiency, and financial return, helping companies assess AI’s impact on work.
Funding rounds
-
Series A
announced 2026-11-12
: 21000000 USD
; investors: Planven, Exor Ventures, Tenacity Capital, TGVG Capital, Fastweb+Vodafone via Founders Factory, CDP Venture Capital SGR
; source: Nebuly
; https://www.nebuly.com/blog/nebuly-raises-21m-series-a
-
Pre-seed
announced 2022-03-01
; investors: Club degli Investitori, Vento, Exor Ventures, 3LB Seed Capital
; source: Caplight
; https://www.caplight.com/company/nebuly
Company timeline
-
2026-11-12:
Series A
(funding)
. Amount: USD 21M. Investors: Planven, Exor Ventures, Tenacity Capital, TGVG Capital, Fastweb+Vodafone via Founders Factory, CDP Venture Capital SGR
Source: Nebuly.
https://www.nebuly.com/blog/nebuly-raises-21m-series-a
-
2026-11-12:
Nebuly advertises hiring across functions
(hr)
. Nebuly said it was hiring in go-to-market, solutions and engineering across New York, London, Germany and Italy. Active postings include a US East Sales Engineer and a Turin Data Scientist (Delivery).
Source: Nebuly.
https://www.nebuly.com/blog/nebuly-raises-21m-series-a
-
2026-11-12:
Nebuly reports multiple six-figure enterprise contracts
(gtm)
. Nebuly reported multiple six-figure contracts with large enterprises and named Oura, Iveco Group, Fastweb+Vodafone, CNH and Zeiss among its customers. The announcement reports its commercial position, not when the contracts began.
Source: Nebuly.
https://www.nebuly.com/blog/nebuly-raises-21m-series-a
-
2026-09-16:
Nebuly describes its AI Proficiency Score
(product)
. Nebuly described a score that uses anonymized interaction traces to assess how employees work with AI. The company presented its current method separately from broader dimensions it says are on its roadmap.
Source: Nebuly.
https://www.nebuly.com/blog/from-better-prompts-to-better-operators-the-road-ahead-for-ai-proficiency-measurement
-
2022-03-01:
Undisclosed pre-seed
(funding)
. Amount undisclosed. Investors: Club degli Investitori, Vento, Exor Ventures, 3LB Seed Capital
Source: Caplight.
https://www.caplight.com/company/nebuly