Zeitgeist #5Saturday, August 29, 2026
Underlying Desire
The deeper force here is the desire to remove drag from work without removing control. People do not actually want another conversational assistant, they want relief from repetitive decisions, paperwork, follow-ups, and context switching, while still feeling confident that nothing important will go off the rails. AI agents promise a rare combination: speed plus delegation plus oversight. That is appealing because modern knowledge work is full of tiny obligations that are expensive for humans and too risky to fully automate with old software. Agents are attractive because they offer the fantasy of a capable junior operator who never sleeps, but the real demand is trust, accountability, and leverage.
Key Evidence
McKinsey’s 2026 State of AI survey says 44% of respondents now report AI is scaling across their enterprise, up from 38% a year earlier, signaling a shift from pilots to org-wide deployment. OpenAI’s August 12, 2026 enterprise signals note says agentic usage is spreading beyond software engineering into broader knowledge-work functions, expanding the addressable market. Contentstack’s 2026 Agentic Enterprise report says organizations are spending 35% more on AI-related software and tooling in 2026 versus 2025, showing that buyers are funding the layer.
Why Now
Three things changed at once: enterprises got past the novelty phase, agent usage broadened beyond engineering, and budgets started moving into AI infrastructure and tooling. Microsoft’s 2026 Work Trend Index showing nearly one in five software and technology firms already using agents is a sign that the category has enough installed base for adjacent products to integrate into real environments. The result is a new buying window for workflow-specific tools, because companies now want agents that fit inside existing systems instead of experimental assistants on the side.
AI platform for legal and professional services teams that runs research, document review, contract analysis, and other legal workflows. ([harvey.ai](https://www.harvey.ai/blog/harvey-raises-growth-round-at-dollar11-billion-valuation-co-led-by-gic-and-sequoia?utm_source=openai))
Outcome: Raised $200 million in March 2026 at an $11 billion valuation, and says it now works with more than 700 customers and is used by over 50% of the AmLaw 100. ([harvey.ai](https://www.harvey.ai/blog/harvey-raises-growth-round-at-dollar11-billion-valuation-co-led-by-gic-and-sequoia?utm_source=openai))
Enterprise AI agent platform for customer service and customer engagement workflows across chat, voice, email, and messaging. ([techcrunch.com](https://techcrunch.com/2025/09/04/bret-taylors-sierra-raises-350m-at-a-10b-valuation/?utm_source=openai))
Outcome: Raised $950 million in May 2026 at a valuation above $15 billion, and earlier reported it had passed $150 million in annual recurring revenue. It also says it has more than 40% of the Fortune 50 as customers. ([techcrunch.com](https://techcrunch.com/2026/05/04/sierra-raises-950m-as-the-race-to-own-enterprise-ai-gets-serious/?utm_source=openai))
Workflow Guard is a control plane for enterprise AI agents that lets operations, security, and business leaders define permissions, approval steps, logging, and rollback rules across critical workflows. The target customer is any mid-market or enterprise team deploying agents into finance, support, sales ops, procurement, or HR, where one bad action can create compliance or customer headaches. It would work because the biggest blocker to adoption is not model quality, it is trust, visibility, and control, and companies will pay for a system that makes agents safe enough to use in production.
Agent Playbooks is a vertical workflow builder that ships prebuilt agent templates for specific functions like customer support triage, invoice processing, sales follow-up, and internal knowledge retrieval. The target customer is an operations leader at a SaaS company or services business that wants measurable automation without hiring a full AI engineering team. It would work because most enterprises do not want to design agents from scratch, they want a fast path to a business outcome, with their data, tools, and approval flow already wired in.
Underlying Desire
At the core, this trend is about trust under uncertainty. Companies want to use powerful AI without feeling exposed, regulators want accountability without reading everyone’s source code, and users want systems that are useful but not mysterious. The deeper human need is control: a way to benefit from automation while reducing the fear that something important is happening invisibly, unfairly, or unlawfully.
Key Evidence
The European Commission says enforcement powers for the AI Act’s AI Office and national authorities apply from August 2, 2026, making compliance an operational requirement, not a future planning exercise. The EU AI Omnibus entered into force on July 27, 2026, according to the European Commission, which extends timelines while increasing oversight pressure. Axios reported on August 28, 2026 that Europe’s AI Act has moved into enforcement, and that Anthropic will add watermarking to future Claude outputs to align with the law.
Why Now
The shift is happening because enforcement has arrived, not just legislation. Once regulators can actually act, companies need evidence, workflows, and product changes, not legal memos. At the same time, the rules are still moving, which makes manual compliance brittle. That creates demand for software that can adapt quickly as obligations change across model providers, deployment types, and jurisdictions.
AI governance platform that helps enterprises discover, assess, govern, monitor, and report on AI systems, agents, and vendors.
Outcome: Credo AI raised $5.5 million at launch, then $12.8 million in Series A, then another $21 million in 2024. In its 2025 review, the company said it had 2x year-over-year revenue growth, 150% growth in enterprise customers, 70% faster AI use-case reviews, and 60% less manual AI compliance work for customers. ([techcrunch.com](https://techcrunch.com/2021/10/19/credo-ai-launches-out-of-stealth-with-5-5-million-to-help-companies-with-ethical-ai/?utm_source=openai))
Enterprise AI governance platform for discovering, testing, monitoring, and enforcing compliance across AI systems.
Outcome: Holistic AI says more than 200 enterprise AI use cases are governed on its platform, with 100+ automated red-team attack tests and a stated 50% AI risk reduction. It also lists customers such as Unilever, Michelin, Adecco, and other Fortune 500 companies, plus ISO 27001 and SOC 2 Type 2 compliance work in its trust center. ([go.holisticai.com](https://go.holisticai.com/?utm_source=openai))
ComplyPilot is a compliance SaaS for startups and mid-market software companies that use AI in products, hiring, support, or decision-making. It automatically maps models, prompts, vendors, datasets, and human review steps into a live AI Act readiness dashboard, then generates audit trails, policy checklists, and exportable evidence packs for legal and procurement teams. It would work because most teams do not need a giant consulting engagement, they need a system that keeps them continuously ready as rules change and customers ask for proof.
PolicyGrid is a platform for AI vendors and enterprise buyers that converts regulatory text into machine-readable controls. Users upload their internal policies, model cards, and deployment details, and the product outputs required tasks, missing evidence, risk flags, and jurisdiction-specific obligations across the EU and other markets. This works because companies are drowning in overlapping rules and need one control layer that can translate legal language into product actions without starting from scratch every quarter.
Underlying Desire
At the core, this trend is about permission and protection. Users, parents, regulators, and platforms all want the internet to preserve access to useful services while drawing a harder line around vulnerability, exploitation, and legal liability. Age verification is the mechanism that turns vague trust into a machine-readable boundary.
Key Evidence
Ofcom said in July 2026 that age checks are being deployed at unprecedented scale in the UK, with many online services now required to use highly effective age assurance under the Online Safety Act, according to Ofcom. The FTC said in February 2026 it would not bring COPPA enforcement actions against certain operators using age-verification technologies solely to determine user age, according to the FTC. Meta said in May 2026 it expanded AI-based age assurance and launched it on Instagram in the U.S., Australia, Canada, and the UK, according to Meta.
Why Now
The regulatory floor changed in 2026: the UK turned age assurance into a practical requirement, while the FTC signaled less fear of using age-verification tech in the U.S. At the same time, major platforms began shipping AI-based checks themselves, which lowers user friction and makes age assurance feel like standard product design instead of a niche compliance workaround.
A digital identity and age assurance company that lets platforms verify age with selfies, documents, and reusable credentials.
Outcome: Yoti says it completed over 1 billion age checks globally, had 13 million app installs globally, and reported 2025 revenue of £29.0 million, up 62% from 2024. It also said it raised £12.5 million from HSBC. ([yoti.com](https://www.yoti.com/blog/yoti-completes-1-billion-age-checks/?utm_source=openai))
A global compliance and age assurance platform for games, social apps, and consumer AI.
Outcome: k-ID says it raised a $45 million Series A in 2024, bringing total funding to $51 million. The company also says it operates across more than 200 jurisdictions and was identified as a leading provider in Australia’s Age Assurance Technology Trial. ([prnewswire.com](https://www.prnewswire.com/news-releases/k-id-closes-45-million-series-a-from-andreessen-horowitz-and-lightspeed-venture-partners-to-set-a-new-global-benchmark-for-age-appropriate-gaming-experiences-302180961.html?utm_source=openai))
AgeLayer is a compliance-first age assurance platform for consumer apps, marketplaces, and community products that need to verify user age without wrecking conversion. It would combine document checks, selfie-based age estimation, parental consent flows, audit logs, and country-specific policy rules into a single API and dashboard. This works because the pain is not just proving age, it is proving you followed the right process in the right jurisdiction, and most teams do not want to build that legal and technical stack from scratch.
AgeCheck API is a developer tool for product teams that need a fast, embeddable age gate across web and mobile apps. Instead of shipping one-off vendor integrations, teams could call a single API that returns an age confidence score, recommended action, and jurisdiction-specific compliance decision. This would appeal to startups and mid-market platforms that need to launch quickly, because the main bottleneck is not technical sophistication, it is reducing implementation time and legal ambiguity.
Underlying Desire
Under the surface, this trend is about control. People do not just want to lose weight, they want to feel stable again while their appetite, routines, identity, and social life are being rewired. GLP-1s create a new kind of dependency: the medication changes behavior, but users still need help making that change sustainable, understandable, and socially acceptable. The deeper desire is to make transformation feel manageable instead of chaotic, and to do it without feeling broken, isolated, or constantly self-monitoring.
Key Evidence
PwC’s 2026 GLP-1 consumer trends report says 80% of current users and 74% of lapsed users are assembling their own routines with supporting products, a strong sign of unmet demand for integrated software and services. BCG’s August 2026 research found 70% of GLP-1 users reported behavior changes among people they live with, showing this is a household-level shift, not just an individual health issue. Consumer Collective’s August 28, 2026 report found GLP-1 users are spending nearly 4x more on fitness than other consumers, which points to a recurring budget category around coaching, tracking, and adherence tools.
Why Now
Two things changed: adoption got big enough to create a real consumer ecosystem, and the side effects of that adoption are now visible in daily life. The market is moving from prescription access to post-prescription support, and the current patchwork of nutrition advice, fitness apps, and telehealth follow-up is not built for this new workflow. The timing is especially good for software because users are already spending, but they are spending across disconnected products.
A gut health nutrition brand that sells prebiotic fiber, bars, and GLP-1 support products.
Outcome: Raised $41.2M across 5 funding rounds, including a Series B in February 2025; the company said it had grown 172% year over year and expanded to more than 5,000 brick-and-mortar stores. ([cbinsights.com](https://www.cbinsights.com/company/muinq/financials?utm_source=openai))
A behavior-change and weight management platform that now offers a GLP-1 companion program.
Outcome: Noom says nearly 80% of Microdose GLP-1Rx users stayed engaged for 4 or more weeks, and its January 2026 analysis covered 14,210 GLP-1Rx members; the company also says members using its GLP-1 Companion lost 25% more weight at week 40 in a March 2026 overview. ([noom.com](https://www.noom.com/blog/weight-management/2025-was-the-breakout-year-for-nooms-glp1-companion/?utm_source=openai))
A daily operating system for GLP-1 users that combines meal planning, protein goals, side effect logging, habit nudges, and lightweight coach support. The target customer is the millions of people on GLP-1s who are improvising with spreadsheets, notes apps, and scattered advice. It would work because the highest-friction part of the journey is not the injection, it is the constant day-to-day decision making that follows. The product can expand into subscription coaching, grocery integrations, and employer benefits.
A shared household planning tool for families and couples where one person is on a GLP-1. It coordinates grocery lists, meal preferences, shared fitness goals, and simple prompts for how the rest of the household can adapt without turning dinner into a negotiation. The customer is not just the user of the medication, but everyone affected by the behavioral shift around them. It would work because BCG’s research suggests GLP-1s are changing behavior inside the home, which creates a real need for coordination software.
Underlying Desire
At the core, this trend is about reducing uncertainty. Businesses want to keep getting paid, avoid fines, and stop depending on humans to interpret constantly changing rules. The deeper desire is control: control over cash flow, control over compliance risk, and control over a process that is now being rewritten by governments instead of product teams.
Key Evidence
Belgium made structured Peppol e-invoicing compulsory on January 1, 2026, according to the Belgian federal finance administration. France’s phased e-invoicing rollout starts September 1, 2026, according to the French tax authority. Spain’s Verifactu timeline has been moved to 2027, according to Spain’s Agencia Tributaria, showing that the mandates are spreading, not stalling.
Why Now
What changed is that the deadlines are no longer abstract. Multiple major European markets have moved from policy debate to implementation dates, which forces software buyers to act now instead of waiting. At the same time, the EU’s VAT in the Digital Age program is pushing structured digital reporting into the mainstream, which makes this a platform shift, not a one-off compliance update.
AI tax infrastructure for global businesses, including e-invoicing, tax ID validation, tax engine, and automated returns.
Outcome: Raised $110 million in Series C on May 28, 2026, and says it powers e-invoicing for 1M+ sellers and processes 500M transactions annually.
API-first e-invoicing and digital tax compliance platform that connects businesses to tax authorities and Peppol-style networks.
Outcome: Closed a €4 million round led by P101 on May 4, 2026, and says it has managed 70M+ invoices since 2019, with 600+ direct clients and 120K+ people reached through its solutions.
A compliance middleware layer that plugs into existing billing systems like Stripe, NetSuite, QuickBooks, and custom ERP stacks, then converts invoices into the correct country-specific structured format. The target customer is mid-market finance teams selling across Europe, especially companies that do not want to rip out their current stack just to satisfy Belgium, France, or Spain. It works because the hardest part of e-invoicing is not generating an invoice, it is translating one invoice into many regulatory dialects and proving it was submitted correctly.
A validation and monitoring tool for accounting and ops teams that checks invoices before submission, flags schema errors, tracks mandate changes by country, and creates an audit trail for every rejected or corrected invoice. This is a strong wedge for companies with small finance teams that cannot afford compliance specialists in every market. It works because once e-invoicing becomes mandatory, the pain shifts from creation to error handling, exception management, and ongoing rule updates.
Underlying Desire
At the core, this trend is about trust plus speed. Businesses and consumers want money that moves instantly, globally, and cheaply, but they also want the reassurance that a real institution stands behind it. Stablecoin rails promise the emotional appeal of modern software, instant access and control, wrapped in the psychological safety of regulated finance.
Key Evidence
The FDIC proposed rules on April 7, 2026 to implement the GENIUS Act for permitted payment stablecoin issuers, according to the FDIC. The OCC issued a request for comments in February 2026 on the GENIUS Act framework, according to the OCC, which suggests banks and fintechs need operational tooling now. In the EU, MiCA is fully in force and a 2026 summary says the final compliance deadline for crypto-asset service providers was July 1, 2026, according to Teroxx.
Why Now
What changed is that regulation became concrete instead of conceptual. The U.S. moved from legislation to agency rulemaking, while the EU already forced the market through MiCA compliance deadlines. That combination turns stablecoins from a speculative asset class into an enterprise payments and compliance problem that software can solve.
Circle issues USDC and builds regulated infrastructure for payments, treasury, and settlement on public blockchains. ([investor.circle.com](https://investor.circle.com/resources/investor-faqs/default.aspx?utm_source=openai))
Outcome: Q2 2026: USDC in circulation was $73.3 billion, USDC onchain transaction volume was $14.8 trillion, total revenue and reserve income was $701 million, and Circle reported 175 financial institutions enrolled in CPN. Circle also went public on June 5, 2025 and received OCC approval to establish Circle National Trust. ([circle.com](https://www.circle.com/pressroom/circle-reports-second-quarter-2026-results?utm_source=openai))
BVNK provides stablecoin payments infrastructure for businesses, including payments, wallets, treasury, and stablecoin issuance. ([bvnk.com](https://www.bvnk.com/about-us?utm_source=openai))
Outcome: BVNK said it was processing $30 billion in annualized stablecoin payment volume in December 2025, up 2.3x year over year, with 2.8 million transactions. It raised $50 million in Series B funding in December 2024, received a strategic investment from Visa in May 2025, and in March 2026 entered into an agreement to be acquired by Mastercard. ([bvnk.com](https://www.bvnk.com/blog/stablecoins-core-financial-infrastructure-2025?utm_source=openai))
A compliance and operations SaaS for banks, fintechs, and licensed stablecoin issuers that need to manage issuance, redemption, reserve attestations, wallet permissions, and monitoring in one place. The product would sit between core banking systems, custodians, and blockchain rails, giving operators a single dashboard for policy controls, approval workflows, audit logs, and regulator-ready reporting. It would work because the hardest part of launching a stablecoin is not token creation, it is everything around it: controls, exceptions, evidence, and reconciliation.
A rules engine and API layer that lets fintechs embed stablecoin transfers without building their own compliance stack. The product would screen wallets, enforce jurisdiction-specific policies, route transactions through approved custodians, and generate the logs banks need for audits and regulator exams. It would sell to payment companies, neobanks, and cross-border platforms that want stablecoin speed but cannot afford to assemble legal and technical infrastructure from scratch.
Underlying Desire
At the deepest level, this trend is about control. Developers, utilities, and AI companies all want certainty in a world where power access feels increasingly random, political, and slow. The human desire underneath is simple: reduce uncertainty, compress waiting time, and turn a fragile physical constraint into something predictable enough to plan around and profit from.
Key Evidence
A 2026 U.S. buildout report tracked 1,547 data center facilities and 537 data-center-related legislative items across 45 states plus the federal level, according to datacenters.builtfor.ai. Newmark's 2026 U.S. data center market outlook says there is a 160-gigawatt pipeline under construction or announced, with energy constraints and delayed grid connections slowing projects, according to Newmark. Bloom Energy's 2026 mid-year pulse found that 61% of developers would bring their own power if the grid is unavailable, according to Bloom Energy.
Why Now
What changed is that AI demand finally outran the grid's ability to respond. The old assumption that a good site plus enough capital would get you built is breaking, because interconnection queues, local permitting, and power procurement have become first-order risks. That creates immediate demand for software that can find feasible sites, model power availability, and automate the messy coordination between developers, utilities, and regulators.
GridCARE uses AI, grid simulations, and system intelligence to find latent grid capacity so large AI and data center loads can get power faster. ([linkedin.com](https://www.linkedin.com/company/gridcare?utm_source=openai))
Outcome: Raised a $64 million oversubscribed Series A in 2026, and the company says it unlocked 1 GW of hidden capacity for customers in its first year of operation. ([latitudemedia.com](https://www.latitudemedia.com/news/gridcare-raises-oversubscribed-64-million-series-a/?utm_source=openai))
Zendo provides an energy OS for data centers, covering energy procurement, capacity management, and billing automation. ([linkedin.com](https://www.linkedin.com/company/zendoenergy?utm_source=openai))
Outcome: Raised $2.2 million in pre-seed funding, led by Fly Ventures, with participation from Pact, Octopus Ventures, and industry angels. ([datacenterdynamics.com](https://www.datacenterdynamics.com/en/news/zendo-raises-22m-to-build-energy-os-for-data-centers-overhaul-energy-procurement/?utm_source=openai))
GridScout is a site selection and interconnection intelligence platform for data center developers, colocation operators, and industrial landowners. It would combine utility map layers, interconnection queue data, permitting timelines, transmission constraints, local opposition signals, and power pricing into a workflow that ranks sites by actual energizability, not just cheap land. This works because the market is no longer deciding between locations, it is deciding between feasible and impossible, and the teams that can cut weeks off diligence will win deals faster.
PowerQueue is a workflow tool for developers managing utility applications, interconnection milestones, and regulatory filings across multiple jurisdictions. It would act like a deal room for power, with document automation, deadline tracking, status dashboards, and alerts when a utility or local agency changes requirements. This would work because every delayed project is bleeding time and carrying cost, and a better operating system for utility coordination can become sticky very quickly.
Lev turns an idea into the things a company actually runs on: positioning, market and competitive research, a pitch deck, and a go-to-market plan.
Start building for free