Valuation Analyst

Anthropic and OpenAI now sell implementation, not models. Within five weeks this summer, Anthropic and OpenAI each launched a business built around one idea: the model was never the hard part. Here is what that admission means if you do not have a Fortune 500 budget. On June 14, 2026, OpenAI announced a $150 million bet called the OpenAI Partner Network. Five weeks later, on July 15, Anthropic - alongside Blackstone and Hellman & Friedman - launched an entire standalone company called Ode with Anthropic. Different structures, different investors, same underlying admission from the two labs that make the world's leading AI models: the model was never the bottleneck. Getting it actually working inside a real business is. That is a remarkable thing for either company to say out loud, let alone both of them within the same month. It is also the clearest signal yet of where the money in enterprise AI is actually moving - and it has direct implications for any business that isn't large enough to hire what these two programs are built to sell. The announcements, in brief. Ode with Anthropic is a new enterprise AI services firm built on the team from Fractional AI, an applied-AI consultancy Anthropic acquired in May 2026. Its backers read like a sovereign wealth roll call: Blackstone, Hellman & Friedman, Goldman Sachs, General Atlantic, Leonard Green & Partners, Apollo Global Management, GIC, and Sequoia Capital. It's led by Fractional AI's co-founders - CEO Chris Taylor and CTO Eddie Siegel - and its pitch is that Claude alone doesn't transform a bank, hospital, or manufacturer; pairing Claude with dedicated human engineers who redesign the actual workflow does. As Taylor put it: "Our teams partner closely with CEOs to define and execute highest priority AI initiatives, driving transformation level impact." Anthropic's Garvan Doyle framed it as filling a gap Anthropic couldn't fill alone: "Ode was built to be that partner, adding to Anthropic's growing ecosystem helping enterprises put Claude to work." Baker Tilly signed on almost immediately, announcing a joint initiative with Ode to advance AI-enabled client service. The OpenAI Partner Network is structurally different but philosophically identical. It's a three-tier certification program - Select, Advanced, and Elite - with credential tracks in Codex, Cybersecurity, and AI Agents, plus a "Forward Deployed Experts" pilot for the hardest enterprise deployments. Founding and launch partners include Accenture, Bain, BCG, McKinsey, and PwC. OpenAI's stated goal is to train and certify 300,000 consultants by the end of 2026. The company's own framing of why the program exists is the most quotable part: "The limiting factor for seeing value from AI in the enterprise is no longer model capabilities" - the bottleneck now is use-case identification, workflow redesign, systems integration, and change management. Read that line again. It is worth sitting with what OpenAI just said in public, in its own announcement copy, about its own product. Not "our models got better." Not "here's our new benchmark score." The company that built GPT is telling the market that the thing it sells - model access - is no longer what determines whether an enterprise gets value from AI. What determines it is whether someone with real implementation experience sat down with the business, found the right use case, redesigned the workflow around it, and wired it into the existing systems without breaking anything. Anthropic, days apart, built an entire company on the same premise instead of just saying it. The line that matters "The limiting factor for seeing value from AI in the enterprise is no longer model capabilities." - OpenAI, announcing the Partner Network. Two of the best-funded AI labs on earth are now spending hundreds of millions of dollars on the belief that implementation, not intelligence, is what enterprise AI is short on. But look at who these are actually built for. Neither program was built with a 40-person business in mind, and neither pretends otherwise. Ode's backers are private equity and sovereign capital; its early public partnership is with Baker Tilly, a top-10 accounting and advisory firm. The OpenAI Partner Network's founding partners are Accenture, Bain, BCG, McKinsey, and PwC - five of the most expensive consultancies in the world, operating at day rates and minimum engagement sizes calibrated for Fortune 500 transformation budgets, not for a regional distributor or a 60-person clinic group. That's not a criticism of either program - they were designed to solve a real problem at the scale their backers operate at. But it does mean the actual mechanism both companies are betting on - a dedicated implementation partner who understands your workflows, redesigns them around AI, and integrates the result into your existing systems - is currently being delivered almost exclusively to companies large enough to sign a McKinsey-sized contract or a private-equity-backed services retainer. Everyone else is left with the model access, minus the part both labs just said actually matters. The principle is right. The packaging is wrong for most businesses. The underlying insight - that workflow redesign, systems integration, and hands-on implementation determine whether AI produces value, far more than which model sits underneath - is not an enterprise-only phenomenon. It is exactly as true for a 50-person logistics company as it is for a bank. A business that buys ChatGPT Enterprise seats or a Claude API key and hands them to staff with no redesigned workflow around them gets roughly the same disappointing result a Fortune 500 company would get doing the same thing - which is precisely the failure mode Ode and the Partner Network exist to prevent for their clients. The gap isn't the idea. It's that nobody built the SMB-scale version of it - a dedicated implementation partner, without the private-equity balance sheet, the Big 5 day rate, or the requirement that you route everything through one lab's proprietary platform. Realistic scenario A 55-person industrial equipment distributor gives its sales and service teams ChatGPT and Claude accounts after reading that AI "boosts productivity." Six months later, usage has flatlined to a handful of people using it to draft emails - because nobody redesigned how a quote gets built, how a service ticket gets triaged, or how parts inventory gets reconciled around what the models can actually do. The models were never the problem. There was no implementation partner translating capability into a rebuilt workflow - the exact gap Ode and the OpenAI Partner Network exist to close for companies fifty times this one's size. What to ask before you sign anything. Whether you're evaluating an enterprise-scale program or a smaller implementation partner, the same questions separate a real engagement from a licensing deal with a consulting label on it: * Does anyone actually redesign its workflow, or are Wizeb just getting access to a model and a training session? * Who owns the infrastructure and the data - Wizeb, or the platform Wizeb is integrating with? * Is pricing predictable, or does it scale unpredictably with usage the way per-seat and per-token consulting retainers often do? * Are Wizeb locked into one lab's ecosystem, or can the implementation move with Wizeb if its tooling changes? * Is there a dedicated person accountable for the outcome, or is this a certification badge on top of a self-service product? Where Wizot Agent Studio comes in. Wizot Agent Studio is Wizeb's answer to exactly this gap - the same principle Ode and the OpenAI Partner Network are built on, sized and priced for businesses that will never sign a Big 5 contract. It deploys custom AI agents on your own infrastructure, not inside a vendor's proprietary platform - your data never leaves your network, which matters as much for a healthcare group under HIPAA as it does for a bank under Ode. Pricing is a fixed retainer, not a per-token or per-seat charge that turns into a surprise bill the moment usage climbs. And critically, it comes with the part both labs just spent hundreds of millions of dollars proving matters most: dedicated implementation work - discovery, workflow redesign, and deployment tailored to how your business actually operates, not a generic chatbot dropped on top of it. Anthropic and OpenAI just told the market, in public, that the model was never the hard part. If you're running a business that's never going to be Ode's or McKinsey's next client, that doesn't mean the lesson doesn't apply to you - it means you need someone building the implementation layer at your scale instead of theirs. Get a free Wizot Agent Studio consultation Wizeb scopes and builds private, custom AI agents on your own infrastructure - fixed pricing, no vendor lock-in, no per-token surprises. Visit wizeb.com/wizot to talk through what an implementation partner actually looks like at your scale. Ready to act on this? Wizeb build exactly what this article is about. Tell Wizeb about your situation - Wizeb'll come back with a realistic assessment. More to read

Baker Tilly launches its own AI platform for accountancy and advisory. June 16, 2026 Consultancy.nl Baker Tilly has launched its own AI platform on which a growing suite of specialized AI agents runs. With the platform, the business services provider wants to deploy AI on a larger scale within its services. According to Baker Tilly, with about 1,000 employees one of the larger accounting and advisory organizations in the Netherlands, the introduction of the platform marks the transition from experimenting with artificial intelligence to broader application within the organization. In recent months, various applications have been developed, tested, and validated in practice. 'AI is developing rapidly and is structurally changing our sector. For us, it is clear that we do not want to follow, but to take the lead ourselves,' says CEO Ronald Hoeksel. 'By reducing repetitive tasks, more room is created for craftsmanship, judgment, and client contact. This makes our work better and more attractive.' On building its own AI platform, he says: 'The foundations have been laid. This enables us to apply AI at scale and demonstrably strengthen our services.' Developed in-house. The new platform has been developed by Baker Tilly itself. While many organizations use generic AI solutions, the firm has chosen its own environment. The platform is now used by hundreds of professionals and combines internal and external data sources. It supports both general research activities and specialized AI agents for specific fields. With the new platform, Baker Tilly aims to maintain more control over the application of AI while ensuring quality, security, and compliance. An important part of the development is a suite of AI agents designed for tasks within the various service lines of the firm. These applications combine technology, expertise, and data to support analyses and assessments. Various applications. Examples of the new AI agents include FiscAI, which helps analyze tax issues and prepare advice based on laws and regulations, and a management report review agent that checks reporting against relevant guidelines. Baker Tilly has also developed an AI agent for tax position and optimization. This analyzes the tax position of clients and checks corporate income tax calculations. Furthermore, a going concern analysis agent has been developed. This application systematically maps risks, assumptions, and uncertainties and supports auditors in going concern assessments by systematically analyzing information and testing assumptions for consistency. According to Baker Tilly, this should contribute to better-founded and more consistently reached judgments. The next step in digitalization. Hoeksel emphasizes that the AI platform is still in its infancy. 'This is not an end point, but a starting point. We have laid a foundation that allows us to develop quickly. By combining technology, data, and expertise, we are building an organization that remains ahead and continuously adds value for clients.' Baker Tilly has thirteen offices in the Netherlands. The organization focuses on small and medium-sized enterprises, family businesses, and the public sector.