David Blonski

The clearest mechanism Blonski describes is ServiceNow's consumption pricing, where the AI unit is a proprietary "assist" whose count the vendor sets rather than actual computing cost.

• SAP is described as blocking outside AI agents from reaching customer data, and Workday's chief executive Aneel Bhusri is quoted saying AI agent pricing has "a lot of upside."
• Blonski's conclusion is not that the AI in these products is useless, but that the pricing and data posture make them poor long-term bets — AI will be the reason customers leave, not the reason they stay.

In a ServiceNow post, Blonski notes subscription increases of roughly 5–10% a year and observes that no CIO has said, "I can't live without Now Assist." They make the same point about Workday leadership's own description of agent pricing.

Blonski reports pricing changes as vendors announce them rather than tracking them systematically. Ask for vendor terms before repeating figures drawn from their statements.

Read original · Original 2 · Original 3

They question Gartner's sourcing, which cites vendor press releases, blog posts and product documentation.

• Blonski accepts the governance argument — probabilistic AI needs deterministic orchestration, auditability and security controls — but disputes the conclusion that only legacy vendors can provide them.
• AI can now reason across systems, workflows and unstructured context, they say, removing the need for a separate copilot, agent framework, workflow engine, interface, security layer and subscription for every function.
• They frame the alternative as one system where rules govern everything deterministic and people stay in the loop for the rest.

A three-way comparison of the CIO's options: a "Mercedes" (legacy SaaS — reliable and familiar, but manual with AI bolted on), a "Tesla" (Palantir — innovative but services-heavy and heavier on hype than delivery), and a "Waymo" (Elementum AI, Harvey, Abridge — full autonomy with humans in the loop, uncomfortable at first).

The three-way analogy turns the same architecture argument into a buyer-facing comparison. It names Blonski's own employer as the Waymo option, making the comparison both an explanation and a positioning exercise.

Read original · Original 2

The reasoning Blonski repeats is that telling people to "use AI" frustrates them when the tool is clumsy and adds work, while redesigning the workflow around what AI does well reduces work and brings people along. Blonski offers this as a blueprint for CIOs struggling to move AI past pilot stage.

Tariff management at Danaher was automated with Elementum AI, freeing 42 employees for other priority projects.

• When tariff rules changed, work that would have taken nine months took a few days.
• A separate post promotes a Sanofi session with Kyle Westwood and Amadeusz Andrzejewski on automating IT, procurement and HR workflows with AI agents, deterministic rules and humans in the loop, co-hosted with Deloitte and Snowflake.

The Danaher account is a partner deployment retold from a conference panel, making it an illustration from an interested party rather than independent proof.

Read original · Original 2

They propose cutting the legacy software stack instead: save the money, keep the jobs and reinvest to move faster.

• Blonski invokes Jevons paradox to argue that efficiency raises rather than lowers demand for the underlying input, so AI should be a growth driver and people are the input.
• The substitutions they name are Salesforce administration so sellers can sell, ServiceNow ticket chasing so engineers can build, Coupa intake work so procurement can negotiate, and Workday workflow bureaucracy so managers can manage.

Sanofi is cited as saving more than €10 million in six months by removing procurement waste, with the savings redirected into hiring scientists and accelerating research. The framing is "fewer toll booths," not fewer people.

Read original

Where AI makes the largest difference — human capital management, talent management, recruiting, onboarding and compensation — they argue companies should start moving workflows out, while leaving payroll and the general ledger in place because they are stable and sensitive.

• The stated payoff is a short-term headache that catalyses modernisation across the rest of the organisation.
• Elsewhere, the same cost claim appears as a 50–70% reduction in total cost of ownership and new workflows delivered in days instead of quarters.

Blonski reads Silver Lake's acquisition of Workday as a white flag for legacy SaaS, predicting that the new owners will cut costs, raise prices and milk customers for as long as they stay. The deal is the occasion for advising CIOs to plan staged exits from three- and five-year renewals.

Ask for a named customer case to substantiate the "few weeks" migration timeline and 50–70% cost figure.

Read original · Original 2

They frame the questions for CIOs as: what are you replacing, who are you replacing, and what are you winning — new contracts, displaced systems, or outsourced and offshored work? Blonski's read is that "better," "faster" and "smarter" sold in 2024 but not in 2026, and that large AI budgets survive only where the outcome plainly justifies the cost.

A single executive's reported second thoughts on Glean, used as evidence that comparative product claims have stopped clearing the buying bar.

Read original

Start here for the central argument behind many of their other posts: AI can reason across systems, making domain-specific software unnecessary.

Blonski challenges Gartner's position that domain-specialist AI agents from Salesforce, SAP, and ServiceNow are the only secure, reliable path to autonomous enterprise operations.

• They call the source list—a ServiceNow press release, a Workday blog post, a Salesforce announcement, and SAP product documentation—big-SaaS propaganda.
• They agree AI should be anchored to deterministic business logic and governed workflows, but argue that legacy SaaS vendors aren't the only ones that can deliver it.
• AI can now reason across systems, workflows, and unstructured context, they say, removing the need for separate copilots, agent frameworks, workflow engines, UIs, security layers, and SaaS subscriptions for every function.
• They present Elementum AI as a single AI-first system of record for ITSM, procurement, CRM, and HRIS, with people in the loop and rules for everything else.

For brands planning a campaign

Use this post as a template for a thought-leadership piece that separates the need for governance from the vendor-lock-in claim. An enterprise AI platform can support Gartner's call for deterministic orchestration while arguing that buyers do not need a separate agent stack from each legacy SaaS vendor.

Start here for the economic case linking legacy SaaS costs to hiring and investment decisions.

Blonski lists layoffs attributed to AI: Oracle 30,000; Amazon 30,000; Intel 24,000; Citigroup 20,000; IBM 8,000—over 100,000.

• They call the attribution mismanagement or misattribution and propose cutting the legacy SaaS stack instead.
• They cite ServiceNow Action Fabric as an AI-agent toll, SAP blocking external AI agents from customer data, Workday CEO Aneel Bhusri saying AI agent pricing has "a lot of upside," and ServiceNow AI EVP Jon Sigler calling agent tolls "a tremendous market opportunity."
• They argue AI should drive growth and invoke Jevons paradox.
• Their replacement examples: Salesforce admin overhead so sellers can sell; ServiceNow ticket chasing so engineers can build; Coupa intake black holes so procurement can negotiate; Workday workflow bureaucracy so managers can manage people.
• They claim 50-70% reduction in TCO, new workflows in days, and management by operators rather than consultant armies—with no data lock-ins, agent tolls, or consumption fees.
• They cite Sanofi saving €10M+ in six months by eliminating procurement waste, with savings going into hiring scientists and R&D, and mention Under Armour and Snowflake.

For brands planning a campaign

This is the strongest cost-and-labor case for an enterprise AI platform.

• It gives a campaign a concrete message: replace software tolls, not headcount.
• Pair the layoff numbers with the replacement categories and Sanofi savings to make the argument tangible.

Start here for a specific example and a clear principle: redesign processes around AI.

Blonski relays Sriram P. at Snowflake Summit describing Danaher's agentic AI implementation.

• Sriram said they didn't just drop AI into the process; they rethought the entire process to optimize it around AI.
• Telling people to use AI without redesigning the process frustrates them because the technology is clunky and creates more work.
• Redesigning around AI's strengths reduces work.
• Danaher uses Elementum AI to automate tariff management, freeing 42 employees for other high-priority projects.
• When tariff rules changed, a task that would have taken 9 months took a few days.
• Blonski calls this a blueprint for CIOs moving AI from POC to production.

For brands planning a campaign

Use this as proof-point content for a process-first AI platform. The operational example—tariff management, 42 employees freed, and 9 months reduced to a few days—gives a campaign a concrete before-and-after, not a generic AI productivity claim.

Start here for a memorable framework that helps newcomers understand the vendor landscape.

Blonski sorts enterprise AI options into three categories: Mercedes (legacy SaaS: reliable, familiar, manual, AI bolt-ons, dated but works); Tesla (Palantir Technologies: innovation at scale, AI capable, human-operated, hype greater than delivery, architecturally limited and services-heavy); and Waymo (Elementum AI, Harvey, Abridge: full autonomy with humans in the loop, a different model, buggy or quirky, uncomfortable but transformative).

• They say CIOs will choose between four 9s of uptime, a name that turns heads, or actual autonomy.
• Like Waymo, once you ride one, the decision gets easier.

For brands planning a campaign

This framework can position a new AI platform against both legacy SaaS and services-heavy AI. Use the Mercedes/Tesla/Waymo lens in an executive webinar or thought-leadership piece to frame the choice: familiar but manual, innovative but human-operated, or autonomous with people in the loop.

Start here for a practical migration sequence and a view of where AI makes the biggest difference.

Blonski calls Silver Lake's acquisition of Workday a white flag from legacy SaaS.

• They say innovation is over, there is no AI second act, and the Bending Spoons model is to cut costs, raise prices, and milk customers.
• For CIOs weighing 3- and 5-year renewals, they advise keeping payroll and general ledger in Workday while starting to move workflows where AI makes a big difference: HCM, talent management, recruiting, onboarding, and compensation.
• They claim AI can complete these migrations in a few weeks.
• They call it a short-term headache, but a good catalyst to up-level the organization.

For brands planning a campaign

Use this to shape a migration-focused campaign without asking for a full rip-and-replace. The specific advice—keep payroll and GL, move HCM and talent workflows first—gives an enterprise AI platform a credible, staged entry point and a reason to engage CIOs facing renewal decisions.

Start here for the pricing critique behind the substitution argument.

Blonski says they are surprised ServiceNow adding consumption pricing has not gotten more attention.

• They argue AI makes legacy SaaS cheaper through fewer seats, less software, and more automation.
• ServiceNow is charging more for software, 5-10% per year by most accounts, and now charging for AI on top.
• They say it is not based on actual compute costs but on proprietary "assists," where ServiceNow decides how many assists each AI action consumes.
• They note that no CIO has said, "I can't live without Now Assist."
• They conclude AI should not be another tax on legacy software, but AI will be the reason customers stop paying for it.

For brands planning a campaign

Use this post to build a pricing-and-value argument.

• For an enterprise AI platform, the case isn't just better AI: legacy vendors are adding AI tolls on top of rising license costs.
• The "assists" example gives a campaign a specific pricing mechanism to scrutinize.

This short, buyer-side view explains why being better than Copilot is not enough.

Blonski relays a Global 1000 COO's doubts about a major investment in Glean.

• As focus shifts to ROI and outcomes, being better than Microsoft Copilot is not enough.
• The questions every CIO is asking: What are you replacing?
• Who are you replacing?
• What are you winning?
• “Better,” “Faster” and “Smarter” worked in 2024, but not in 2026.
• They say companies must win new contracts, replace systems, or replace outsourcing or offshoring to justify AI investments.
• AI budgets remain very large where the ends clearly justify the means.

For brands planning a campaign

Use this as a messaging filter for an enterprise AI campaign. Lead with substitution and outcomes—what system, role, or contract the AI replaces—instead of comparative claims that it is smarter or faster than a copilot.

This event connects named companies to a practical automation agenda.

Blonski promotes a session with Kyle Westwood and Amadeusz Andrzejewski on Sanofi's AI transformation.

• They describe it as a practical guide for large enterprises looking to automate operational workflows in IT, procurement and HR with AI agents, deterministic rules and humans in the loop.
• The stated goals are to reduce costs such as SaaS licenses and BPO, and empower people to do their best work with AI.
• Elementum AI and Deloitte are hosting the event, with thanks to Sanofi and Snowflake.
• Date: Tuesday, July 28, 4:00-5:00 PM CEST / 10:00-11:00 AM EDT.

For brands planning a campaign

Blonski turns a customer story into an event and content hook. A campaign can follow the same pattern: take a named enterprise, map its IT, procurement and HR workflows, and present the transformation as a practical guide rather than a product pitch.

Blonski points to ServiceNow adding consumption pricing of roughly 5–10% a year and metering AI through proprietary "assists" that the vendor, not compute cost, defines; to SAP blocking external agents from customer data; and to Workday leadership describing AI agent pricing as having "a lot of upside." The claim is not that AI is useless in these products but that the pricing and data posture make them poor long-term bets, and that AI will be the reason buyers leave rather than the reason they stay.

Names the vendors and executives, quotes their pricing language, and contrasts the claimed saving with the added fee rather than arguing in general terms about modernity.

For brands planning a campaign

Strong fit for a challenger positioning against suite incumbents: the argument names vendors, details pricing mechanics, and quotes executives to support a comparison or cost-of-ownership narrative.

• It focuses on pricing models rather than product capability, making it best suited to messaging on total cost, data access, and contract structure.
• Confirm any borrowed numbers before publication.

Examples: Source 1 · Source 2 · Source 3

The same logic applies across vendors: replace Salesforce admin overhead so sellers can sell, ServiceNow ticket chasing so engineers can build, Coupa intake work so procurement can negotiate, and Workday workflow bureaucracy so managers can manage. The stated payoff is a 50–70% reduction in total cost of ownership and new workflows delivered in days instead of quarters.

Breaks the decision into functions a buyer already owns, separates systems worth keeping from those worth replacing, and attaches an outcome range instead of making a generic efficiency claim.

The timeline is described as achievable "in a few weeks" and the cost reduction as 50–70%, but no named customer case supports those numbers. Ask for a customer case before treating them as verified results.

For brands planning a campaign

Gives an enterprise-software brand a concrete, function-by-function exit playbook to sponsor or respond to.

• It is a natural hook for migration guides, ROI models, or executive roundtables.
• Ask for risk, compliance, and timeline detail to make the proposed sequencing credible.

Examples: Source 1 · Source 2

They accept the governance argument—probabilistic AI needs deterministic orchestration, auditability, and security controls—but dispute the conclusion that only legacy vendors can supply them.

• Their position is that AI can reason across systems, workflows, and unstructured context, removing the need for a separate copilot, agent framework, workflow engine, interface, security layer, and subscription for each business function.
• The clearest expression is a three-way comparison of the CIO's options: a "Mercedes" (legacy SaaS, reliable but manual), a "Tesla" (Palantir, innovative but services-heavy and hype-prone), and a "Waymo" (Elementum AI, Harvey, Abridge—full autonomy, humans in the loop, initially uncomfortable).

Takes a named analyst position head-on, grants its valid half, then challenges the vendor conclusion and closes with a three-way analogy that turns an architecture argument into a buying decision.

For brands planning a campaign

A ready-made thought-leadership hook for a unified-platform pitch: the analogy is portable, memorable, and easy for executives to reuse.

• Separating what analysts get right from what they conclude also creates a structure for respectful counterargument.
• Because it reads as a competitive attack, decide whether that contrast is the point or whether a more neutral framing of the same architecture argument would serve better.

Examples: Source 1 · Source 2

The paraphrased point: telling people to "use AI" frustrates them when a clunky tool adds work, while redesigning the process reduces work and increases adoption.

• The example is tariff management automated with Elementum AI, freeing 42 employees for other projects and turning a change that would previously have taken nine months into a few days.
• A related post promotes a Sanofi session with Kyle Westwood and Amadeusz Andrzejewski, co-hosted with Deloitte and Snowflake, on automating IT, procurement, and HR workflows.

Tells the story through a named executive's own framing, keeps the quoted language in direct speech, and anchors the method in measurable before-and-after numbers.

The quotes are paraphrased from memory, and the outcome figures come from a customer using the author's product. Ask for source material before relying on the details.

For brands planning a campaign

Useful for vendors whose buyers are stuck between pilot and production: it offers a named customer, quantified outcomes, and a process-first message without overclaiming about the technology.

• The Danaher account is a partner deployment retold from a conference panel, making it suitable as an illustration or for event and webinar material.
• Ask for independent evidence before using it as proof.

Examples: Source 1 · Source 2 · Source 3

The questions Blonski puts to CIOs are: what are you replacing, who are you replacing, and what are you winning? Their read is that "better," "faster," and "smarter" sold in 2024 but not in 2026, and that budgets remain large only when the outcome clearly justifies the cost.

For brands planning a campaign

Good framing for a brand selling into budget scrutiny: it shifts evaluation from feature comparisons to displaced systems and measurable outcomes, supporting messaging about replacements and contract wins. The observation comes from one unnamed executive; ask for the source before treating it as a broader market signal.

Examples: Source 1

They invoke Jevons paradox to argue that efficiency raises demand for the underlying input, making AI a growth driver and people the input. Sanofi's reported €10M+ saving in six months illustrates the pattern: savings redirected into hiring more scientists and accelerating research, framed as fewer toll booths rather than fewer people.

Leads with named companies and figures, reframes the causal claim, then offers an alternative budget decision and a concrete example of where the savings went.

For brands planning a campaign

A socially framed counter-narrative for enterprise-software brands whose buyers face pressure to justify headcount decisions.

• It pairs naturally with cost-substitution messaging.
• Avoid implying that your customers are avoiding layoffs unless you can support that claim: the posts generalize from a single saving figure and a set of announced layoffs.

Examples: Source 1

They describe its approach as combining governed workflows with AI agents and rules. Ask them to confirm their title and formal relationship with the company.

View example

They also promote a webinar about Sanofi’s AI transformation, hosted by Elementum AI and Deloitte.

View example

The post says the work freed 42 employees for other projects and reduced a task from nine months to a few days, results they attribute to the executive’s account at Snowflake Summit.

View example

View example

They challenge its AI consumption pricing and argue that charging for AI on top of legacy software makes the product less attractive to customers.

View example

In the cited pricing post, they attribute comments about AI-agent pricing to Workday’s CEO; the claim about blocking external agents in that post concerns SAP, not Workday.

View example

Libeau's post on the 17% of leaders seeing clear procurement-technology ROI and their piece on how hypergrowth turns a disruptor into an incumbent reflect the buyer-side skepticism about AI promises that runs through Blonski's work.

Libeau focuses on procurement, finance and spend technology. Their case for software aging centres on product and commercial complexity, unlike Blonski's cross-vendor argument about the cost and architecture problems of legacy suites.

Read original · Original 2

Goumas's Workday posts on agent testing, data quality and mid-market readiness align with Blonski's emphasis on redesigning work and questioning vendor claims.

Goumas works within the Workday and HRIS ecosystem, offering platform teams concrete testing and rollout guidance. Blonski takes a CIO's broader view of the software portfolio and the economics of replacing suite vendors.

Read original · Original 2

Keenan's ServiceNow–Sweep and Salesforce AIforce analyses share Blonski's focus on what vendor strategy means for customers.

Keenan goes deeper into Salesforce and ServiceNow platform mechanics, metadata, DevOps and governance. Blonski focuses on pricing, migration sequencing and total cost of ownership across several suites.

Read original · Original 2

Their posts argue that a company can become legacy ahead of schedule: hypergrowth stacks on more products, services, packaging, specialists, and AI branding than the original product was designed to carry, and the foundation quietly gives out.

• The second post uses a procurement technology survey figure — 17% of procurement, finance, and IT leaders reporting clear ROI from their procurement technology and AI investments — to argue that AI cannot manufacture clean supplier data, adoption, or reconciled workflows, so the foundation decides how far AI gets.
• For an AI platform selling into procurement, finance, or IT spend buyers, that diagnosis-of-failure framing is a stronger opening than Blonski’s vendor-side attack: it lets the audience recognize their own stalled programme before the sponsor names a fix.

Libeau explains why the foundation fails but offers no replacement architecture, migration sequence, or named alternative to the incumbent stack.

• A campaign would therefore carry the full “what to do instead” burden.
• The framing also cuts against the simplest version of the sponsor’s message: if the problem is too much intelligence piled on weak systems, an AI layer that substitutes for the suite stack is exactly what Libeau’s argument questions.

Examples: Source 1 · Source 2

Goumas’s posts give a concrete pilot method rather than a strategy argument: pick one narrow use case, a small tester group, approved sources, verifiable answers, and no automated changes, then trace wrong answers back to outdated policy, missing country exceptions, wrong employee location, or information the system never held.

• The second post explains Workday Build — Extend, Orchestrate, an agent builder, Data Cloud, developer tooling — and contrasts enterprise teams with dedicated platform staff against a mid-market employer with one integrations person, closing with five first steps, including watching consumption-based credit pricing change the budget math.
• That practitioner detail suits a sponsor promising a working pilot plan for a resource-constrained team, not a market argument.

The scope is Workday and the audience is Workday-owning HR technology teams, so a campaign built on Goumas would not reach the CIO and enterprise architecture buyers the AI-platform brief is written for, or carry the cross-suite cost and AI-pricing argument Blonski supplies.

• The advice also points away from replacement: Goumas tells teams to learn the platform they already run, starting with the self-service and deployment agents they have today.
• A sponsor positioning itself against legacy SaaS would be borrowing a message that keeps the incumbent.

Examples: Source 1 · Source 2

Keenan publishes longer pieces with a stated argument, linked sources, and named takeaways — one post reads ServiceNow’s reported $400 million acquisition of Sweep as a bid to govern AI agents inside Salesforce estates without moving customer data, introducing the idea of a dependency-resolved model of the system an agent is about to change.

• The second tracks a 168-day arc from Parker Harris’s public question about logging into Salesforce again to the AIforce announcement, and reads the interface change as a control story, including an unstated price and a governance layer not arriving until at least February 2027.
• A sponsor whose message is about who governs agents, what they are allowed to touch, and how platform changes get delivered can use that frame and the longer-format precedent; a cost-replacement message cannot.

The frame is governance and control inside one vendor ecosystem, not the price-and-substitution argument the campaign needs.

• A sponsor would be buying an audience of ecosystem practitioners rather than CIOs comparing whole suites, and would have to build the cost case from other sources.
• The existing work also raises a question a marketing piece must answer: does the sponsor’s own product have a delivery model for the systems agents change?

Examples: Source 1 · Source 2