Caitlin Ferguson

They ask what a customer's real alternatives are, including doing nothing, why they pick the firm and which parts of its advantage should never be automated.

• Automating the wrong step can erode what wins business.
• Because implementing AI is digital transformation at speed, the guardrails need to come before the tools.

A founder running a premium grocery delivery service deliberately keeps AI out of customer service because white-glove support retains roughly 500,000 clients against grocery stores, meal subscription services and doing nothing.

Read original · Original 2

Participants mark the steps AI could take, estimate hours per step, and then bolt AI onto those steps instead of rebuilding the workflow as "AI-native".

• Ferguson also sets four filters: where today's models can be trusted, where AI will not erode how the business competes, where the owner is comfortable working that way, and only where it is safe, because you cannot say "AI did it" and be off the hook.
• Two lessons Ferguson draws from the exercise are that spotting where AI fits is much easier with the whole process visible, and that buildable steps should still be validated and hardened before rollout.

Ferguson scrapped the third day of the prepared curriculum mid-class and took the room to a whiteboard. Students mapped the stretch between someone saying "let's talk about working together" and client kickoff, and found 13.5 hours of work AI could reliably own across prep research, CRM entry, team handoffs, go/no-go pre-qualification, and resourcing sheets and hourly rates.

The exercise replaced a day of curriculum Ferguson had already written. They describe the class as mostly non-operations people, including business owners, marketers, university educators, data scientists, nonprofit leaders and writers.

Read original · Original 2

They work from a single map of a professional-services firm covering marketing and sales, client onboarding, internal resourcing, project management and governance, and talent and culture, and cannot find a step that could not be enhanced with AI. Ferguson points to the MIT NANDA study's finding that the AI pilots succeeding most were operations-focused, and sees operations as the natural entry point for services firms.

Ferguson describes watching a group of agents turn a vague landing-page concept into a fully branded, copywritten, interactive website in under five minutes, with the collaboration visible so a human stays in the loop.

Read original · Original 2 · Original 3

They map its parts to concrete tooling: a business overview and persistent instructions to a Claude.md-style file, SOPs and frameworks to reusable "Skills", and reference works to a local folder.

• Students build inside a standard chat interface rather than assembling large stacks of integrations.
• Ferguson frames the result as new capability plus time back, not a technical project.

Students in Ferguson's 2025 sessions built an automation that pulls a client list from HubSpot, scans for news relevant to each client, and drafts a hyperpersonalised newsletter in the brand's voice every two weeks.

• Another runs a list of names through contact enrichment and lead scoring, then into the CRM in minutes.
• Ferguson notes that these were founders whose focus is not marketing systems, working inside ChatGPT's regular interface with no hardcore coding.

Read original · Original 2 · Original 3

They say defining what could be automated is the easy part; deciding which work should be yours is the hard part.

Ferguson's catalogue holds 274 automations with an estimated $91,375 a month in recovered value, costed at $100 an hour.

• Category totals include sales and revenue operations at $11,550 a month, project management at $7,450, and marketing and content at $8,075.
• Student Christina Guerrier cut meal planning and shopping from three hours a week to fifteen minutes, estimating 143 hours a year back and $900 to $1,300 saved from wasting less food.
• Joseph Schommer, who runs a New York Christmas tree stand among four businesses, calls the return "ROTime" — hours and mental energy freed from spreadsheet reconciliation.

Read original · Original 2 · Original 3

They walked through the method: researchers surveyed 152 executives, interviewed 52, examined about 300 announcements, and recorded return on investment as executive perception rather than validating it separately.

• Ferguson summarised the root causes as people using AI for personal productivity instead of organisational transformation, and weak memory recall in most language-model tools.
• They also noted Forrester's prediction that companies would pull back at least a quarter of AI spending.
• Ferguson says the framing changes what a firm should do next, making the study worth reading in full.

The same post points to the parts of the study Ferguson agrees with: the 5 percent of pilots that were succeeding were operations-focused, and professional services was the second most disrupted industry behind media.

Read original · Original 2

They argue that consulting and professional services are trust and relationship businesses, referrals make up around 70 percent of their own business, and they have never heard of a boutique firm being built by high-volume, non-descript outreach. This stance sits alongside their automation teaching: automate internal operational work and protect client trust.

Ferguson's complaint targets LinkedIn outreach promising "0 to 20 qualified leads" and asking them to buy another indistinguishable AI SDR service.

Read original

Their stated goal is a tactical result a participant can deploy the same day.

• They have repeatedly rebuilt the design based on participant feedback, cutting content overload and adding time for the Keystone logic and for planning next automations.
• Ferguson measures readiness by validation and hardening, not syllabus completion.

After the 2025 sessions, Ferguson concluded too much had been crammed into one session and moved to cohorts kept intentionally small at three people. They also cut a day of curriculum live in class when it was not working.

The course has changed dates, session count, and cohort size across Ferguson's posts, from a summer multi-cohort schedule to smaller relaunches.

Read original · Original 2 · Original 3 · Original 4 · Original 5

The show is built on the idea that the operator who makes the business run and the finance leader who keeps the capital strategy sound rarely talk openly in public.

• Its focus is service-firm economics rather than automation mechanics: margins, utilisation, and what to pay for finance support by stage.
• The track record belongs to Ripp, not Ferguson: Ferguson says Ripp has helped more than 100 founders toward roughly $200 to $300 million in exit value and has written Finance for Founders.

An episode covers a founder who lost a six-figure bank loan because the books did not balance and the “CFO” was actually a bookkeeper; an $800,000 internal-meeting problem; a four-page profit and loss statement as a red flag; and benchmarks for labour cost as a percentage of revenue, utilisation, revenue per employee, and finance support costs by company stage.

Read original · Original 2 · Original 3

The list is a working method, not an apology: test in market rather than think at the whiteboard, keep marketing alive when client work is busy, reuse material already produced, pitch media regularly, keep the network informed, amplify other people, and start reading benchmarks sooner. They also describe realising they were not following their own advice after a street interview where they said people should not put limits on themselves.

Ferguson says stopping marketing when client work arrived left their pipeline thin, and that getting stuck at the whiteboard instead of testing in market cost roughly six to eight weeks of time to market. They had run a podcast series, Hustle & Humor, with Monique Mills but did not make the most of its clips, and credit the “Heartbeat Newsletter” habit to Charlie O'Donnell.

Read original · Original 2 · Original 3

On the second anniversary of founding COOPilots.io, Ferguson lists eight things they would do differently.

• They fell into the feast-and-famine trap: when client work arrived, they stopped marketing, then found the pipeline thin.
• They got stuck at the whiteboard, thinking about a product or service instead of testing it in market, which they estimate cost six to eight weeks.
• They did not make the most of clips from a podcast series they ran with Monique Mills, pitch media regularly, or keep up a regular "Heartbeat Newsletter" to their network for a year—a habit they credit to Charlie O'Donnell.
• They wish they had spent more time amplifying people in their network.
• They also began reading benchmarks such as the TSIA and Service Performance Insight professional-services maturity models only later, using them to compare client retention and conversion against peers.
• They describe a street interview where they said people put limits on what you can do and you should not limit yourself, then realised they were not following their own advice.

For brands planning a campaign

A prospective client can see Ferguson name their own operational failures in the same terms they use with clients, supporting a frank scoping conversation about where an AI project is likely to stall.

Ferguson scrapped the third day of their curriculum mid-class and ran a whiteboard exercise instead.

• Students mapped the stretch between someone saying "let's talk about working together" and the day a client engagement kicks off, identifying 13.5 hours of work AI could reliably take over, including prep research, CRM entry, team handoffs, go/no-go pre-qualification, and resourcing sheets and hourly rates.
• Before starting, they set four filters for where AI belongs: where today's models can be trusted to perform, where it will not erode how the business competes, where the owner is comfortable with it, and where it is safe—because you cannot say "AI did it" and be off the hook.
• Students reported that mapping a whole process end to end makes it much easier to spot where AI fits, and that even when a step could be built immediately, they would rather validate and harden each step before rolling it out.
• The class included business owners, marketers, university educators, data scientists, nonprofit leaders, and writers.

For brands planning a campaign

The workshop a client buys is visible in this exercise: map the process end to end, estimate the hours, then roll out one validated step at a time.

Ferguson argues that 60 to 80 percent of a consulting business could be automated with tools available today, and says they would not be surprised if that figure is conservative given how fast the tools are moving.

• They point to a Miro map of a professional-services firm end to end, covering marketing and sales, client onboarding, internal resourcing, project management and governance, and talent and culture.
• They say they cannot find a single step in it that could not be enhanced with AI.
• They map Claude's file structure onto the Keystone architecture they teach: a business-overview and persistent-instructions file corresponds to Claude.md, skills correspond to SOPs and frameworks, and a local folder holds reference works.
• They describe watching a group of agents go from a vague landing-page concept to a branded, copywritten, interactive site in under five minutes, with the collaboration visible so a human stays in the loop.

For brands planning a campaign

Their pitch is about capacity and margin, not headcount. A firm can test whether the 60 to 80 percent figure holds for its own delivery model before committing to a larger programme.

Ferguson sets out two opening steps for building AI into operations.

• First, get clear on the competitive edge and set guardrails: what a customer's real alternatives are, including doing nothing; why they choose the firm; and which parts of that edge should never be automated.
• Their example is a founder running a premium grocery delivery service who deliberately keeps AI out of customer service because white-glove support retains his clients against grocery stores, meal subscriptions, and doing nothing.
• Second, identify the operational lever that moves the profit and loss.
• They group the AI projects that show up on the financials into four categories: accelerate time to value, for example turning discovery-call transcripts into proposals in minutes; expand capacity without new hires, for example auto-generating onboarding guides, project plans, and kickoff decks; reduce cost of delivery, for example turning Slack or Teams notes into clean SOPs or using AI for content quality checks; and increase decision velocity, for example auto-summarising client updates so leaders can course-correct in hours.
• They cite MIT NANDA's finding that operations has been the strongest area for AI pilots and say implementing AI is digital transformation at ten times the speed.

For brands planning a campaign

An off-the-shelf workshop agenda, with margin levers tied to specific operational asks rather than tool demonstrations.

Ferguson maintains a database of 274 automations they say could be deployed today, with an estimated return of $91,375 a month.

• They costed it conservatively at $100 an hour for realistic time savings and assumed no revenue upside, counting only recovered capacity.
• Category totals include sales and revenue operations at $11,550 a month, project management at $7,450, marketing and content at $8,075, a founder's personal operating system at $5,250, and customer support at $5,050.
• Their view: defining what could be automated is the easy part; the hard part is finding which ones should be yours.
• The catalogue draws on patterns across more than 100 businesses and is validated against what they consider genuinely automatable with current tools.
• They expect it to pass 1,000 entries.

For brands planning a campaign

A firm can benchmark its candidate list against their categories. Their conservative costing method also gives a workshop a usable template for estimating time savings.

It is a detailed account of one student's build rather than a claim about what is possible.

Joseph Schommer runs four businesses, including a New York Christmas tree stand that opens after Thanksgiving, runs around the clock, and moves inventory on five to seven trucks with bike-based delivery and setup.

• Diesel at $6.30 a gallon and his wish to give his team real raises had to be reconciled on a sidewalk in December while a customer haggled over a one-of-a-kind tree.
• He built two halves.
• The first pulls financial information and website data into one self-updating view: daily sales against prior years, cash in, profit and loss, sales channels, and contractors.
• The second, which he calls his price lab, takes real cost inputs such as freight, fuel, crew, and cost basis; produces a price range for each tree size, anchored to last year's actual averages; sets website prices directly; and gives him a range on the lot instead of a gut call under pressure.
• It replaced spreadsheets holding different information, where fixing a number in one broke the next and he became the reliability engine.
• He called the return ROTime: hours and mental energy freed to spend with customers he has seen come back for 16 years.

For brands planning a campaign

A clear example of why volatile input costs and a short selling window make a live costing model more useful than periodic reconciliation—the same logic applies to professional-services pricing and resourcing.

Across more than 120 people in AI for Operations sessions during 2025, Ferguson describes two student builds.

• The first pulls a client list from HubSpot, scans for relevant news and automatically drafts a hyperpersonalised newsletter in the brand’s voice every two weeks.
• The second enriches a list of names with news updates and lead scores, then moves it into the CRM, ready for action in minutes.
• They note that the founders were not focused on marketing systems or workflow automation, and that everything was built in ChatGPT’s regular interface, without heavy coding or a large stack of integrations.
• Follow-up surveys showed they had crammed too much into one session; students wanted more time on the Keystone logic and planning their next automations.
• Cohorts are therefore kept intentionally small, at three people, so everyone builds something during the sessions.

For brands planning a campaign

Evidence that a short, small-cohort format can produce working automations from non-technical staff, supporting a pilot group to validate adoption before a firm-wide rollout.

It is the shortest entry point in the set.

Ferguson argues that the hardest part of starting with AI is often not the tool but the blank page: wondering where to use it.

• Their answer: sometimes you don’t need a strategy; you need a list to react to.
• They wrote down eight roles they use AI for in their own business: chief of staff, executive assistant, marketer, research assistant, curriculum designer, web designer, executive coach and fractional chief financial officer.
• Each comes with a starter prompt.
• They tell readers to scan the list, notice which role prompts a reaction and start there.

For brands planning a campaign

A low-cost way to open a workshop or pilot: ask a team to review a list of roles and pick one, surfacing genuine demand before committing budget to a larger programme.

They work from a single map of an entire professional-services firm — marketing and sales, client onboarding, internal resourcing, project management and governance, talent, and culture.

• The framing is consistent: implementing AI is digital transformation at speed, not a tool purchase.
• They point to the MIT NANDA study finding that the pilots succeeding most were operations-focused, and treat that as the natural entry point for services firms.

For brands planning a campaign

A firm buying a workshop should expect a map of the whole business first, not a tool list. Agree up front that the deliverable will be a ranked set of operational steps to automate, each tied to a line-of-business process the firm already runs.

Examples: Source 1 · Source 2 · Source 3

Its components map to concrete tooling: a business overview and persistent instructions to a Claude.md-style file, SOPs and frameworks to reusable "Skills", and reference material to a local folder. Participants build inside a standard chat interface instead of assembling large stacks of integrations.

Diagnose the operational lever that moves the P&L, then encode how the business actually works — guardrails, process knowledge, examples — so the model can act inside those boundaries. Ferguson describes Claude Cowork as a close fit for this architecture and reports watching agents produce a branded, written, interactive page in minutes with a human in the loop.

For brands planning a campaign

For a margins brief, ask for the context layer as a deliverable in its own right, not just finished automations. That layer makes a build reusable after the workshop and gives a client-facing team something it can maintain without the facilitator in the room.

Examples: Source 1 · Source 2 · Source 3

The same discipline shows up in student results: a meal-planning system that cut three hours a week to fifteen minutes, with an estimated 143 hours a year back and $900–$1,300 saved from less wasted food. The recurring promise is time and cognitive capacity, not growth.

For brands planning a campaign

A campaign requiring measurable workflow time savings matches how Ferguson already quantifies work. Ask for a before-and-after hours estimate per workflow, the hourly rate assumed, and the baseline so the figure can be audited rather than taken at face value.

Examples: Source 1 · Source 2

Students mapped the stretch between "let's talk about working together" and client kickoff, then found 13.5 hours of work AI could reliably own across prep research, CRM entry, team handoffs, pre-qualification, and resourcing sheets. Ferguson drew out two lessons: it is easier to spot where AI fits when the whole process is visible, and even buildable steps should be rolled out slowly.

Give participants a real end-to-end workflow, five minutes, and sticky notes; have them mark the steps AI could take and estimate hours per step. Bolt AI onto individual steps rather than rebuilding the workflow as "AI-native".

For brands planning a campaign

Ferguson teaches teams to harden each step before scaling it: pilot a step, verify it, then widen. For a campaign that needs to verify changes across client-facing teams, ask how Ferguson would measure the pilot before expanding it.

Examples: Source 1 · Source 2

The worked example is a premium grocery delivery founder who deliberately kept AI out of customer service to protect white-glove support and assessed real alternatives, including "do nothing". Ferguson pairs this with three questions: what are the alternatives, why do customers choose the firm, and which parts of its edge should never be automated?

For brands planning a campaign

For a professional-services firm, this is the argument for keeping senior client relationships human while automating the preparation and administration around them. Raise it early with client-facing teams: it addresses the main objection to adoption.

Examples: Source 1

They have repeatedly rebuilt the format from participant feedback, including cutting content overload, adding build time, and relaunching with new dates. Sessions run live and remote, with webinars and recordings offered as lower-commitment entry points.

For brands planning a campaign

A scoped workshop with Ferguson should be hands-on, involve a small group, and produce a working artifact by the end — not a presentation. If the firm needs several teams trained, plan for sequential cohorts rather than one large session.

Examples: Source 1 · Source 2 · Source 3 · Source 4

Joseph Schommer runs four businesses, including a seasonal NYC Christmas tree stand, and built a live financial view plus a "price lab" that takes freight, fuel, crew, and cost inputs and produces a price range for each tree size, set against last year's averages.

• Christina Guerrier built a meal-planning and shopping agent in three class sessions.
• Both write-ups include concrete hours, dollar figures, and screenshots shared with permission.

The meal-planning build is a household task, not a professional-services workflow. Treat it as evidence of the method, not proof of results inside a B2B firm; ask for a reference from a firm that has run a comparable build.

For brands planning a campaign

Ask Ferguson to connect a prospective buyer with participants who ran comparable builds. Their existing case-study habit makes a reference request reasonable rather than unusual.

Examples: Source 1 · Source 2

Professional services and consulting are trust- and relationship-based businesses, they argue, where referrals make up roughly 70% of their own work. Their position is consistent with the rest of their content: automate internal operational work and protect client trust.

For brands planning a campaign

A firm selling outreach automation to this audience would clash with the stated view. Frame any campaign around margin and delivery capacity instead, and avoid promising lead volume.

Examples: Source 1

The show is framed as the conversation between the COO and CFO that is rarely held openly.

• Episodes cover six-figure loans lost because the books did not balance, an $800K internal-meeting problem, a four-page P&L as a red flag, and benchmarks for labor cost as a percentage of revenue, utilization, revenue per employee, and what to pay for bookkeeping, accounting, and CFO support by stage.
• They cite TSIA and SPI professional-services maturity models as comparison tools.

For brands planning a campaign

The margin-lift brief connects directly to this material. Ferguson can speak to service-firm economics as well as automation mechanics, making a workshop tied to P&L levers more relevant than one focused on tool adoption alone.

Examples: Source 1 · Source 2 · Source 3

They walked through its methodology: 152 executives surveyed, 52 interviewed, about 300 announcements examined, and ROI recorded as executive perception rather than validated separately. They summarised the root causes as using AI for personal productivity instead of organisational transformation, and weak memory recall in most tools.

For brands planning a campaign

A sceptical buying committee may apply the same standard to the firm's baseline data. Bring defensible numbers for current workflow hours and delivery cost.

Examples: Source 1

They tell readers to begin with whichever role made them react, and also reference a one-week sprint format for firms that want to kickstart AI in their work.

For brands planning a campaign

This is low-friction entry content for a firm's first step and a template a workshop can reuse. For a campaign requiring measurable savings, use it to get teams naming candidate workflows before scoping.

Examples: Source 1 · Source 2

They have delivered conference workshops, including “AI 201: Beyond Prompting to Operating” with Michelle J. at Operations Nation's London event, aimed at participants leaving with something deployable the same day.

For brands planning a campaign

Format options span a single webinar through a multi-week build. For a scoped workshop with measurable savings, the cohort format fits best; a webinar can introduce the topic to a wider client-facing group first.

Examples: Source 1 · Source 2 · Source 3

They cut a day of curriculum live when it was not working, concluded after a cohort that too much had been crammed into one session, moved to smaller cohorts, and relaunched on new dates. The founding retrospective records the same instinct applied to COOPilots.io itself, including falling into feast-and-famine marketing and spending weeks at the whiteboard instead of testing in market.

Treat each cohort as a test: run the exercise, collect participant feedback, and rebuild the affected module for the next run. Ferguson measures whether a build is ready by its validation and hardening, not by completion of a fixed syllabus.

For brands planning a campaign

A pilot engagement is likely to be adjusted between sessions, which suits a firm validating adoption team by team. Set expectations that the first cohort produces a hardened step, not a finished firm-wide rollout.

Examples: Source 1 · Source 2 · Source 3

They use it to host the AI for Operations class, teach the Keystone architecture, and enroll workshop participants.

View example

View example

They present it as past collaborative content.

View example

View example

View example

They use the research to support their teaching.

View example

View example

They use the platform in their teaching.

View example

The example illustrates the work taught in their class.

View example

The post identifies it as a class member’s company.

Original disclosure

Screenshots of his demo shared with permission.

View example

Brubaker examines trust-accounting errors and effective hourly rates by case type for law firms; Ferguson examines workflow hours and recovered capacity for consulting and services firms.

Brubaker covers law-firm accounting, compliance and case-selection economics through client advisory examples, not live cohort courses or automation builds. Brubaker does not write about AI.

Read original · Original 2 · Original 3 · Original 4

Stambovsky covers CRM hygiene, data silos, and how LLMs and MCPs affect consolidation decisions; Ferguson covers process maps, guardrails, and encoding how a business works so AI can act within it.

Stambovsky covers marketing, sales and revenue-operations tooling in CRM and HubSpot stacks, often through commentary on the software ecosystem. Ferguson focuses on operational workflow scoping and cohort teaching for services firms, with an emphasis on workshops and builder case studies.

Read original · Original 2 · Original 3 · Original 4 · Original 5

Parker brings supply-chain and tournament-merchandise operations into a university classroom and a podcast, including discussions of how digital product passports and AI fit into manufacturing. Ferguson brings services-firm workflows into live classes and publishes participant builds.

Parker covers supply chain, manufacturing and product data, primarily in academic and industry-media settings. Ferguson covers AI operations inside professional-services firms, teaching paid small cohorts around a named automation architecture.

Read original · Original 2 · Original 3 · Original 4

Her published work focuses on catching errors before automation is built on top of them: bookkeeping by someone unfamiliar with the rules governing an account, errors that compound month after month until the client balance no longer matches the bank statement, and cleanup that costs more than hiring the right specialist from the start.

• She also shows the diagnostic a services firm needs before deciding whether to grow.
• For a solo attorney weighing whether to staff up or take fewer cases, she asked for the effective hourly rate by case type — hours from intake to close divided into the net fee — which surfaced two case types worth pursuing harder.
• If the firm cannot yet say which work is profitable, that calculation is a more useful first engagement than a workshop: it shows which workflows are worth automating.

Caitlin Ferguson's posts do not teach AI implementation: they do not cover workflow automation, model reliability, guardrails, or a class format.

• They therefore cannot support a scoped operations workshop or the hours-saved figures a campaign measuring time savings needs.
• The posts focus on law-firm finance, including trust-accounting rules that do not transfer to most other industries, so firms outside legal services face real translation risk.
• Use Ferguson for the financial baseline, then bring in an AI operations teacher to build the workflows.

Examples: Source 1 · Source 2

He argues that the consolidation era may be moot: businesses that stayed on many specialized tools, each holding clean data, can now let LLMs work across those tools through connectors, while LLMs cannot unify data that actively disagrees with itself.

• Firms that rushed to consolidate and made a mess may be worse off than those that stayed siloed and tidy.
• His second post makes the same point at the account level: "who owns this lead?" should be a property lookup, not a meeting.
• The important work is agreeing what a lifecycle stage means, passing context in a handoff, and identifying which workflow keeps changing the owner.
• For a professional-services firm whose CRM is the actual bottleneck, that diagnosis is the prerequisite for a productive AI workshop.

His published material covers diagnosis and architecture, plus a survey report on HubSpot practice, rather than a taught curriculum.

• His posts provide no workshop format, cohort logistics, or per-workflow hours baseline.
• For a campaign seeking measurable time savings and adoption validation, ask for those specifics or bring in a different supplier.
• He frames problems at the systems and data layer, so the engagement may produce a cleaner diagram rather than a working automation or a trained client-facing team.

Examples: Source 1 · Source 2

The published posts cover merchandise and inventory flow at a major golf tournament — front-of-house selling and back-of-house inventory management during a short, intense window with a lot of product moving — and a conversation on digital product passports as a way to bring sustainability, sourcing, manufacturing, and the product record into one place, alongside AI and automation in manufacturing. A firm whose margin depends on sourcing, stock, and traceability, or that needs a single product record spanning suppliers and end of life, gets more from that perspective than from a services-delivery automation teacher.

This is the least direct substitute for a campaign seeking an AI workshop for professional-services firms.

• The published posts show no professional-services delivery work or AI training offering: the teaching context is a university classroom, and the AI commentary focuses on manufacturing rather than consulting margins.
• A services firm would be borrowing general operations discipline, not a workshop, automation catalog, or time-savings method.
• Treat him as a fit for a different need, not a like-for-like alternative.

Examples: Source 1 · Source 2