A $100 TV Campaign, a $5,000 Data Team and the Same Business Model

A $100 connected-TV campaign and a $5,000-a-month data team point to the same change: specialist capabilities are being repackaged for much smaller companies. The useful version of this shift removes a recurring job while leaving judgement, customer relationships and consequential decisions with the owner.

For years, small-business software has mostly meant smaller versions of enterprise software. The owner still had to learn the interface, connect the systems, interpret the output and remember to act on it. This week’s launches suggest a more useful direction: sell the capability, not the machinery behind it. That sounds like a subtle change, but it alters what an owner should expect from AI for small business.

Expertise is being unbundled

ImpactFactory.ai says a small business can now start a connected-TV campaign with a media budget of $100, including ZIP-code targeting and measurement tied to website visits.¹ Amazon is pushing in the same direction with DVA+, which lets advertisers build display, streaming TV, online video and audio campaigns through a four-step workflow.² Those are channels that have traditionally rewarded people who understood media buying, campaign setup and optimisation. The barrier is not gone, but a lot of the specialist machinery has been pulled behind the product.

Kixik is doing something similar at a very different price point. It is offering an outsourced data team to companies with $2 million to $50 million in annual revenue, starting at $5,000 a month, and says recreating that team in-house can cost $1.1 million to $1.6 million a year.³ That is not cheap software aimed at the smallest shop on the high street. It is still a useful signal because the thing being sold is not a dashboard or a model licence. It is access to a capability that previously implied hiring several people.

Meta’s Muse for Small Business pushes the same idea into everyday operations. Muse can connect to Shopify, QuickBooks, Stripe, Canva, Asana, Klaviyo, Notion, Slack, Facebook and Instagram business data, and Meta says it will not publish content, send messages or make purchases without approval.⁴ The product becomes more interesting when it can look across the tools an owner already uses instead of asking the owner to move every problem into a new chat window.

These examples are not really about cheaper access to software. They are about unbundling expertise from headcount. A media buyer, data analyst or operations specialist still knows far more than a general-purpose model, but a growing share of routine setup, retrieval, comparison and first-pass analysis can now be packaged into a product. That makes capabilities available to firms that could never justify the full-time role.

Access is not adoption

Cheaper access does not mean owners automatically know what to do with it. OpenAI and America’s Small Business Development Centers announced a programme that plans to train around 150 SBDC advisers and initially reach at least 1,000 small businesses through hands-on workshops.⁵ That is a revealing design choice. Instead of assuming that a café owner or contractor should keep up with every model release, the programme puts guidance through advisers who already work with small businesses.

The programme is also planning sector-specific playbooks as it expands, which is another useful clue about what adoption requires.⁵ A generic model can answer a generic question, but a business owner needs help connecting the tool to the work they actually do. That is how adoption becomes practical. An owner does not need a lecture on agent architectures to decide that unanswered reviews, inconsistent follow-up or three hours of weekly content admin are worth fixing.

This matters because the gap between trying AI and changing a workflow is still large. Heartland Forward’s survey of 691 small-business owners found that 43% were using AI, and 93% of those users said it saved them time.⁶ More than half reported saving at least five hours a week. Those figures describe real value, but they also imply that the remaining job is not getting owners to type a prompt. It is helping them identify where five hours can disappear from the week without introducing new risk.

A five-person company cannot absorb much implementation overhead. Large organisations can assign people to integration, governance, security, vendor management and training, then call the total programme an efficiency project. A small firm feels every extra step directly because the person configuring the tool is often the same person answering customers and checking the bank account. The adoption test is therefore brutally practical: did work disappear, or did software management increase?

The best products hide the mechanics

BT Business launched an AI receptionist aimed at answering enquiries, helping with appointments and follow-ups, and working with existing business phone numbers.⁷ BT’s research says UK small businesses lose £3.7 billion a year through missed calls, with 67% missing at least one potentially valuable call on an average day. Whether every business needs an AI receptionist is a separate question. The attraction is obvious because the owner can measure the outcome in calls answered and appointments captured, not prompts written.

Bold Penguin and Thimble are applying agents to small-business insurance quoting. Their integration can exchange business information and return a priced quote, but the human agent still reviews coverage selections before a formal quote is requested.⁸ That division of labour is sensible. Moving information between systems and preparing the quote is repetitive work; deciding whether the coverage is appropriate carries responsibility that should remain visible.

MeevoIQ, aimed at salons, spas and med spas, follows a similar pattern with advisers for front-desk calls, marketing offers and business performance.⁹ Its marketing adviser shows the rationale and projected revenue before an owner approves an action. Thryv’s Growth Platform for local service businesses scores lead intent, summarises conversations and suggests next steps while connecting marketing activity to revenue.¹⁰ These products are useful examples because they do not ask an owner to admire the intelligence. They ask whether the call was answered, the lead was followed up and the suggested action makes commercial sense.

That should be the standard for Instagram for small business too. Generating twenty captions is easy; keeping the business recognisable while using its own photos, offers, customer context and timing is the work that matters. The same applies to Instagram content planning and AI captions for Instagram business posts. If the owner has to rewrite every generic draft, hunt for the right media and then rebuild the calendar manually, the model may be impressive while the workflow is still poor.

Cheap execution raises the value of judgement

Amazon’s DVA+ workflow lowers the amount of specialist knowledge required to launch sophisticated media. Its Review Requests product lowers another barrier by letting eligible sellers pay to invite verified purchasers to leave ratings or written reviews, with Amazon saying products in the closed beta received around three times as many ratings and reviews per week as before.¹¹ More businesses can now buy capabilities that were previously awkward, specialist or expensive.

That democratisation is good for owners, but it removes some easy sources of differentiation. If everyone can generate acceptable creative, launch multi-format media and inspect how they appear in AI search, simply having access to those functions stops being special. The advantage shifts toward source material, customer knowledge, timing and taste. Cheap execution makes a clear point of view more valuable, not less.

This is especially visible in content. A product photograph, a menu change, a new treatment, a founder’s explanation or a customer question contains information that belongs to the business. A generic prompt does not. The useful role of Instagram AI content is to help turn those raw materials into a consistent stream without flattening them into the same polished phrases every other business is publishing.

That is also why “how to automate Instagram content creation” can be the wrong starting question. The better workflow begins with what should remain human: what the business wants to say, which image is true to the work, what offer is appropriate and what should never be posted automatically. Automation can organise the media, draft variants, schedule the approved version and remove repetitive admin. It should not quietly decide what the brand stands for.

Five hours can become growth

The most interesting number in the Heartland Forward survey was not the 43% adoption figure. Among AI users, 81% said the tools let them do more with the same staff, and one in five said AI-driven growth had helped them afford to hire additional employees.⁶ That complicates the usual assumption that productivity and jobs sit on opposite sides of the ledger. In a small company, saving time can create capacity before it creates cuts.

Five hours returned to an owner can become another sales meeting, faster customer follow-up, product work or simply an evening that no longer disappears into admin. The destination matters because small businesses are constrained by attention more often than by ideas. If AI removes the repeatable work around a valuable activity, the owner can spend more time on the part that produces trust or revenue. That is a much more useful measure than the number of seats, prompts or features activated.

There is a commercial implication for vendors too. Products aimed at smaller firms cannot rely on a six-month implementation followed by training sessions and internal champions. They need to arrive much closer to the finished job. A salon should see fewer missed calls, a broker should prepare quotes faster, and a retailer should get usable campaign options without learning the vocabulary of programmatic advertising first.

The products that understand this will increasingly look less like “AI software” and more like narrowly scoped services with software inside them. Owners will still need to review the decisions that affect money, customers, safety, reputation or brand voice. Everything around those decisions is becoming fair game for automation. The winning proposition is not intelligence on demand. It is less unfinished work left on the owner’s desk.

The owner should feel the absence

The $100 TV campaign and the $5,000 data team sit at opposite ends of the small-company market, but they are selling the same basic promise. Both reduce the amount of specialist infrastructure a business has to build before it can access a capability. That promise gets stronger when the product also absorbs setup, integration and routine execution rather than handing those jobs back to the customer.

This is where the next wave of small-business products will be judged. Owners do not need more evidence that models can write, classify, analyse or call APIs. They need to know whether Tuesday afternoon contains fewer repetitive tasks, whether customers still recognise the business, and whether the person responsible for a consequential decision is still obvious. Those are operational questions, not model questions.

The firms building for this market should treat owner attention as the scarce resource. Remove the task, show the reasoning where it matters, ask for approval at the right boundary and keep the business’s own voice intact. If the owner has to become the full-time supervisor of the automation, the product has not made expertise more accessible. It has simply moved the specialist job onto someone who already had too much to do.

Sources