I
Ishan Pandey
Guest
Gartner expects worldwide spending on AI to reach $2.67 trillion this year and $3.64 trillion in 2027, yet when IBM surveyed 2,000 chief executives it found that only a quarter of their AI initiatives had delivered the return they expected over the previous three years. That gap between how fast the money is moving and how little of it can be tied to results has become the defining question of enterprise AI in 2026, because the next round of budgets depends on knowing which projects deserve more investment and which should quietly stop.
Ascerta, the Bellevue company previously known as Pay-i, has raised an $18 million Series A to answer that question for large organisations. Its platform tracks what AI costs down to the individual user, team and use case, including sub-token charges, hidden fees and enterprise discounts that most tools never see, before connecting that spending to the business outcomes each initiative was meant to deliver, giving CIOs, CFOs and AI leaders a single system for deciding what to scale, what to fix and what to cut.
The round was led by Dell Technologies Capital, with Hitachi Ventures, BGV, Wipro Ventures and earlier investors participating, bringing Ascerta's total funding to $22.9 million. Dell Technologies Capital's managing director Raman Khanna, who speaks for the firm on the deal, spent 16 years as Stanford University's chief information officer before moving into venture capital. He has twice been named to the Forbes Midas List and counts data and security companies such as Alation, JFrog and Netskope among his current and past investments. A former CIO leading a round in a company that sells to CIOs is a strong signal about who Ascerta is built for.
The Series A comes 16 months after the company emerged from stealth as Pay-i with a $4.9 million seed round co-led by Fuse Partners and Tola Capital, when its focus was measuring the cost of AI applications. Customers quickly asked for more than a cost view, because they also needed to know how people were using AI, what agents and models were actually doing and whether any of it justified further investment, which is what drove the rebrand to Ascerta and the move into what the company calls Enterprise AI Management. The strategic investors in this round fit that expansion closely, since Wipro is both an investor through Wipro Ventures and a customer, while Hitachi and Dell sit at the heart of the enterprise infrastructure where much of this AI runs.
AI has become one of the fastest-growing lines in corporate spending anywhere, with Gartner's latest forecast putting worldwide AI spending at $1.79 trillion in 2025, rising 49% to $2.67 trillion in 2026 and another 36% to $3.64 trillion in 2027, with infrastructure taking the largest share while software, services and a fast-growing group of newer categories take a rising portion of the rest.
The composition of that spending matters for a company like Ascerta, because the fastest growth is happening in exactly the categories that are hardest to measure. Gartner expects spending on AI agents and assistants to grow about fourfold between 2025 and 2027, from $16.5 billion to $65.5 billion, with generative AI models growing at the same pace and AI cybersecurity more than tripling, while the market as a whole roughly doubles. Every one of those agents consumes tokens, triggers other services and does work whose value is spread across teams, so the more autonomous AI becomes, the harder it gets to see what any single dollar produced.
The smaller AI segments are growing fastest, with agents and assistants and generative AI models each set to grow about 4x between 2025 and 2027, against 2x for the market as a whole. Source: Gartner, September 2026.
A second force makes the accounting problem harder rather than easier, because AI keeps getting cheaper per unit while companies keep spending more in total. Andreessen Horowitz calculated that the price of language-model output at a fixed level of capability fell about 1,000 times in three years, from $60 per million tokens in 2021 to $0.06 in 2024, a decline of roughly 10 times a year. Over almost the same period, Menlo Ventures found that enterprise spending on generative AI grew from $1.7 billion in 2023 to $37 billion in 2025, about 22 times higher, which means falling prices have translated into far more usage rather than smaller bills, the pattern economists call Jevons' paradox.
On a common index with 2023 set to 1, the price of a fixed level of AI capability fell about tenfold a year while enterprise generative AI spending rose about 22-fold between 2023 and 2025. The 2025 price point extends a16z's trend and is not a measured price. Sources: a16z, Menlo Ventures; author's indexing.
That explosion of usage has not yet produced an equally clear record of returns, since IBM's 2025 CEO Study found that only 25% of AI initiatives had delivered their expected return and only 16% had been scaled across the enterprise, while 64% of chief executives admitted that the fear of falling behind pushes them to invest in some technologies before they understand the value. Gartner has gone further for agents specifically, predicting that more than 40% of agentic AI projects will be cancelled by the end of 2027 because of rising costs, unclear business value or weak risk controls, which is exactly the set of problems that better measurement is meant to catch early.
Of the AI initiatives CEOs reported over the past three years, 25% delivered the expected return and 16% were scaled across the enterprise. Source: IBM Institute for Business Value, 2025 CEO Study of 2,000 CEOs in 33 countries.
The finance function has responded by treating AI as its most urgent new responsibility, which shows clearly in the work of the FinOps Foundation. Its members manage cloud and technology costs for large organisations, with its latest survey finding that the share of practitioners managing AI spend rose from 31% in 2024 to 63% in 2025 and 98% in 2026. FinOps for AI is now the top forward-looking priority among those members, while AI value management is the skill they most want to add to their teams. Traditional FinOps tools can show what cloud and AI cost, yet they were not built to show what that spending does for the business, a gap that widens as agents take on more of the work.
The share of FinOps practitioners managing AI spend rose from 31% to 98% in two years, faster than any other category they oversee. Source: FinOps Foundation, State of FinOps 2026 (1,192 respondents) and prior editions via CIO Dive.
Ascerta connects to the AI already running inside an organisation and deploys alongside existing systems, covering homegrown applications as well as the enterprise tools most companies now use, including Microsoft's Copilot suite, Amazon Bedrock AgentCore, Salesforce Agentforce and coding agents such as GitHub Copilot, Claude Code and Codex. From there it follows AI through three layers, starting with how people use it, moving to the work that AI actually performs and ending with the outcomes that work drives, so that every model call can be tied to a specific use case and every use case to the business metric it was meant to move.
Three products put that model to work across an organisation's AI estate, starting with Atlas, which measures value, adoption and return on investment from a single workflow up to the full portfolio, showing which initiatives create value, which need fixing and which should be cut. Forge shows how engineering teams use coding agents and helps turn that adoption into measurable productivity, which matters as coding tools become one of the largest AI line items in most technology budgets. Convoy serves organisations that provision their own AI capacity, helping them consolidate workloads onto what they already pay for and add new use cases without disrupting production.
The detail in the cost layer is what separates the platform from a dashboard of token counts. Ascerta tracks sub-token costs, hidden fees and negotiated enterprise discounts at the level of individual model calls while measuring adoption by person, team and tool so leaders can see who is getting real results and help everyone else build the same fluency. That combination lets a company put a hard dollar value on an AI-powered feature, recover money lost to failed agent runs, duplicate projects and unsanctioned "shadow AI" before deciding where its next dollar is most likely to pay off.
Ascerta works with customers including Atos, Wipro and several global insurance carriers, alongside partners such as Microsoft, AWS, IBM, Slalom and Trace3, with further engagements at AWS's Generative AI Innovation Center. Across that base the company reports average improvements of 47% in return on AI initiatives, 24% in the time it takes to launch an agent and 86% in wasted AI spend. The examples behind those averages are concrete, since one customer discovered through Ascerta that agent runs averaging $0.40 were occasionally spiking to $70, while a global insurance carrier saved about $3 million by consolidating its AI capacity, according to GeekWire.
Ascerta customers report on average 86% less wasted AI spend, 47% higher ROI on AI initiatives and 24% faster agent launches. Company-reported figures. Sources: Ascerta; agent cost example via GeekWire.
Atos offers the clearest picture of how a large deployment works, since the French technology group runs what it calls Sovereign Agentic Studios, an operating model for moving agentic AI from pilot to production at global scale, with its group chief technology officer and chief AI officer Florin Rotar crediting Ascerta with giving Atos the visibility and control it needs to scale those initiatives on the basis of measurable business value rather than technical capability alone. Rotar also sits among Ascerta's advisers, alongside former Microsoft chief financial officer John Connors and McKinsey senior partner Lari Hämäläinen, who leads the firm's global work on generative AI and agents.
Ascerta was founded in 2024 by three Microsoft veterans who watched AI economics break at enterprise scale from the inside. Chief executive David Tepper spent 19 years at Microsoft and led generative AI strategy for internal use across Azure, chief technology officer Doron Holan spent 27 years there, including as a Windows architect, where he designed the hyperscale throttling infrastructure that handles hundreds of billions of requests a day, while chief operating officer Erik Winters brings experience running startups. The company is still small, with 15 employees today and plans to reach about 40 by the end of 2027.
Tepper's argument is that the market is full of vanity metrics, because companies can count tokens, lines of generated code and agent runs yet still struggle to say what AI has done for the business. His view is that each organisation needs insight specific to its own people, business and use cases, together with tools built to prevent waste and optimise aggressively for value, since the companies which learn to measure AI this way will be the ones that win the next phase of adoption. The rebrand from Pay-i to Ascerta reflects that shift from tracking cost to managing the entire value of an enterprise's AI investment.
Dell Technologies Capital sees Ascerta building the system of record for AI value creation, a phrase that places it alongside the systems enterprises already rely on for finance, customers and people. Khanna's view is that most enterprises are moving beyond broad experimentation and concentrating their investment on what delivers measurable business value, which creates demand for exactly the visibility and rigour Ascerta provides, with the firm partnering with the company to define the Enterprise AI Management category.
The market data supports that view from several directions at once, because AI budgets are on course to roughly double in two years, the fastest-growing segments are the agents and models whose value is hardest to trace, unit prices keep falling in a way that pushes total usage higher and almost every FinOps team now has AI spend on its agenda. Categories that begin as cost tools often grow into systems of record once they hold the data that finance and operations teams both depend on, while Ascerta already sits between the AI tools, the finance function and the business units that fund them.
Ascerta will use the Series A to scale its platform and its go-to-market team while extending integrations to every major enterprise AI tool, building on coverage that it says already includes nearly all of them. The larger plan is to turn its research on AI value into products that improve returns directly, so the platform moves from measuring what AI is worth toward actively raising that value as enterprises run thousands of models, agents and workflows at once.
The direction of enterprise AI makes that ambition timely, because spending is heading toward $3.6 trillion a year, agents are multiplying inside every major software platform and boards are asking for evidence that the money is working. With $18 million from Dell Technologies Capital, strategic backers in Hitachi and Wipro, a customer base that includes Atos and global insurers and a founding team that ran AI at Microsoft's scale, Ascerta enters that moment with a strong claim to become the place where enterprises settle the question every AI budget eventually raises, which is what all of this is actually worth.
Don’t forget to like and share the story!
Vested Interest Disclosure: HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR.
Ascerta, the Bellevue company previously known as Pay-i, has raised an $18 million Series A to answer that question for large organisations. Its platform tracks what AI costs down to the individual user, team and use case, including sub-token charges, hidden fees and enterprise discounts that most tools never see, before connecting that spending to the business outcomes each initiative was meant to deliver, giving CIOs, CFOs and AI leaders a single system for deciding what to scale, what to fix and what to cut.
The Round
The round was led by Dell Technologies Capital, with Hitachi Ventures, BGV, Wipro Ventures and earlier investors participating, bringing Ascerta's total funding to $22.9 million. Dell Technologies Capital's managing director Raman Khanna, who speaks for the firm on the deal, spent 16 years as Stanford University's chief information officer before moving into venture capital. He has twice been named to the Forbes Midas List and counts data and security companies such as Alation, JFrog and Netskope among his current and past investments. A former CIO leading a round in a company that sells to CIOs is a strong signal about who Ascerta is built for.
The Series A comes 16 months after the company emerged from stealth as Pay-i with a $4.9 million seed round co-led by Fuse Partners and Tola Capital, when its focus was measuring the cost of AI applications. Customers quickly asked for more than a cost view, because they also needed to know how people were using AI, what agents and models were actually doing and whether any of it justified further investment, which is what drove the rebrand to Ascerta and the move into what the company calls Enterprise AI Management. The strategic investors in this round fit that expansion closely, since Wipro is both an investor through Wipro Ventures and a customer, while Hitachi and Dell sit at the heart of the enterprise infrastructure where much of this AI runs.
The Enterprise AI Budget
AI has become one of the fastest-growing lines in corporate spending anywhere, with Gartner's latest forecast putting worldwide AI spending at $1.79 trillion in 2025, rising 49% to $2.67 trillion in 2026 and another 36% to $3.64 trillion in 2027, with infrastructure taking the largest share while software, services and a fast-growing group of newer categories take a rising portion of the rest.
The composition of that spending matters for a company like Ascerta, because the fastest growth is happening in exactly the categories that are hardest to measure. Gartner expects spending on AI agents and assistants to grow about fourfold between 2025 and 2027, from $16.5 billion to $65.5 billion, with generative AI models growing at the same pace and AI cybersecurity more than tripling, while the market as a whole roughly doubles. Every one of those agents consumes tokens, triggers other services and does work whose value is spread across teams, so the more autonomous AI becomes, the harder it gets to see what any single dollar produced.
The smaller AI segments are growing fastest, with agents and assistants and generative AI models each set to grow about 4x between 2025 and 2027, against 2x for the market as a whole. Source: Gartner, September 2026.
The AI Value Gap
A second force makes the accounting problem harder rather than easier, because AI keeps getting cheaper per unit while companies keep spending more in total. Andreessen Horowitz calculated that the price of language-model output at a fixed level of capability fell about 1,000 times in three years, from $60 per million tokens in 2021 to $0.06 in 2024, a decline of roughly 10 times a year. Over almost the same period, Menlo Ventures found that enterprise spending on generative AI grew from $1.7 billion in 2023 to $37 billion in 2025, about 22 times higher, which means falling prices have translated into far more usage rather than smaller bills, the pattern economists call Jevons' paradox.
On a common index with 2023 set to 1, the price of a fixed level of AI capability fell about tenfold a year while enterprise generative AI spending rose about 22-fold between 2023 and 2025. The 2025 price point extends a16z's trend and is not a measured price. Sources: a16z, Menlo Ventures; author's indexing.
That explosion of usage has not yet produced an equally clear record of returns, since IBM's 2025 CEO Study found that only 25% of AI initiatives had delivered their expected return and only 16% had been scaled across the enterprise, while 64% of chief executives admitted that the fear of falling behind pushes them to invest in some technologies before they understand the value. Gartner has gone further for agents specifically, predicting that more than 40% of agentic AI projects will be cancelled by the end of 2027 because of rising costs, unclear business value or weak risk controls, which is exactly the set of problems that better measurement is meant to catch early.
Of the AI initiatives CEOs reported over the past three years, 25% delivered the expected return and 16% were scaled across the enterprise. Source: IBM Institute for Business Value, 2025 CEO Study of 2,000 CEOs in 33 countries.
The finance function has responded by treating AI as its most urgent new responsibility, which shows clearly in the work of the FinOps Foundation. Its members manage cloud and technology costs for large organisations, with its latest survey finding that the share of practitioners managing AI spend rose from 31% in 2024 to 63% in 2025 and 98% in 2026. FinOps for AI is now the top forward-looking priority among those members, while AI value management is the skill they most want to add to their teams. Traditional FinOps tools can show what cloud and AI cost, yet they were not built to show what that spending does for the business, a gap that widens as agents take on more of the work.
The share of FinOps practitioners managing AI spend rose from 31% to 98% in two years, faster than any other category they oversee. Source: FinOps Foundation, State of FinOps 2026 (1,192 respondents) and prior editions via CIO Dive.
How Ascerta Works
Ascerta connects to the AI already running inside an organisation and deploys alongside existing systems, covering homegrown applications as well as the enterprise tools most companies now use, including Microsoft's Copilot suite, Amazon Bedrock AgentCore, Salesforce Agentforce and coding agents such as GitHub Copilot, Claude Code and Codex. From there it follows AI through three layers, starting with how people use it, moving to the work that AI actually performs and ending with the outcomes that work drives, so that every model call can be tied to a specific use case and every use case to the business metric it was meant to move.
Three products put that model to work across an organisation's AI estate, starting with Atlas, which measures value, adoption and return on investment from a single workflow up to the full portfolio, showing which initiatives create value, which need fixing and which should be cut. Forge shows how engineering teams use coding agents and helps turn that adoption into measurable productivity, which matters as coding tools become one of the largest AI line items in most technology budgets. Convoy serves organisations that provision their own AI capacity, helping them consolidate workloads onto what they already pay for and add new use cases without disrupting production.
The detail in the cost layer is what separates the platform from a dashboard of token counts. Ascerta tracks sub-token costs, hidden fees and negotiated enterprise discounts at the level of individual model calls while measuring adoption by person, team and tool so leaders can see who is getting real results and help everyone else build the same fluency. That combination lets a company put a hard dollar value on an AI-powered feature, recover money lost to failed agent runs, duplicate projects and unsanctioned "shadow AI" before deciding where its next dollar is most likely to pay off.
Customer Deployments
Ascerta works with customers including Atos, Wipro and several global insurance carriers, alongside partners such as Microsoft, AWS, IBM, Slalom and Trace3, with further engagements at AWS's Generative AI Innovation Center. Across that base the company reports average improvements of 47% in return on AI initiatives, 24% in the time it takes to launch an agent and 86% in wasted AI spend. The examples behind those averages are concrete, since one customer discovered through Ascerta that agent runs averaging $0.40 were occasionally spiking to $70, while a global insurance carrier saved about $3 million by consolidating its AI capacity, according to GeekWire.
Ascerta customers report on average 86% less wasted AI spend, 47% higher ROI on AI initiatives and 24% faster agent launches. Company-reported figures. Sources: Ascerta; agent cost example via GeekWire.
Atos offers the clearest picture of how a large deployment works, since the French technology group runs what it calls Sovereign Agentic Studios, an operating model for moving agentic AI from pilot to production at global scale, with its group chief technology officer and chief AI officer Florin Rotar crediting Ascerta with giving Atos the visibility and control it needs to scale those initiatives on the basis of measurable business value rather than technical capability alone. Rotar also sits among Ascerta's advisers, alongside former Microsoft chief financial officer John Connors and McKinsey senior partner Lari Hämäläinen, who leads the firm's global work on generative AI and agents.
Founder Thesis
Ascerta was founded in 2024 by three Microsoft veterans who watched AI economics break at enterprise scale from the inside. Chief executive David Tepper spent 19 years at Microsoft and led generative AI strategy for internal use across Azure, chief technology officer Doron Holan spent 27 years there, including as a Windows architect, where he designed the hyperscale throttling infrastructure that handles hundreds of billions of requests a day, while chief operating officer Erik Winters brings experience running startups. The company is still small, with 15 employees today and plans to reach about 40 by the end of 2027.
Tepper's argument is that the market is full of vanity metrics, because companies can count tokens, lines of generated code and agent runs yet still struggle to say what AI has done for the business. His view is that each organisation needs insight specific to its own people, business and use cases, together with tools built to prevent waste and optimise aggressively for value, since the companies which learn to measure AI this way will be the ones that win the next phase of adoption. The rebrand from Pay-i to Ascerta reflects that shift from tracking cost to managing the entire value of an enterprise's AI investment.
Investor Thesis
Dell Technologies Capital sees Ascerta building the system of record for AI value creation, a phrase that places it alongside the systems enterprises already rely on for finance, customers and people. Khanna's view is that most enterprises are moving beyond broad experimentation and concentrating their investment on what delivers measurable business value, which creates demand for exactly the visibility and rigour Ascerta provides, with the firm partnering with the company to define the Enterprise AI Management category.
The market data supports that view from several directions at once, because AI budgets are on course to roughly double in two years, the fastest-growing segments are the agents and models whose value is hardest to trace, unit prices keep falling in a way that pushes total usage higher and almost every FinOps team now has AI spend on its agenda. Categories that begin as cost tools often grow into systems of record once they hold the data that finance and operations teams both depend on, while Ascerta already sits between the AI tools, the finance function and the business units that fund them.
Enterprise AI Management Roadmap
Ascerta will use the Series A to scale its platform and its go-to-market team while extending integrations to every major enterprise AI tool, building on coverage that it says already includes nearly all of them. The larger plan is to turn its research on AI value into products that improve returns directly, so the platform moves from measuring what AI is worth toward actively raising that value as enterprises run thousands of models, agents and workflows at once.
The direction of enterprise AI makes that ambition timely, because spending is heading toward $3.6 trillion a year, agents are multiplying inside every major software platform and boards are asking for evidence that the money is working. With $18 million from Dell Technologies Capital, strategic backers in Hitachi and Wipro, a customer base that includes Atos and global insurers and a founding team that ran AI at Microsoft's scale, Ascerta enters that moment with a strong claim to become the place where enterprises settle the question every AI budget eventually raises, which is what all of this is actually worth.
Don’t forget to like and share the story!
Vested Interest Disclosure: HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR.