AI · What actually changed

The list of unsolvable problems just got a lot shorter.

For decades, a whole category of business problems had no good answer. Not because nobody tried, but because the work required judgment about messy human input, and software could only follow rules. That changed. We build the systems that take advantage of it, and we spend our days testing new models to know which ones can actually be trusted with the job.

// if AI isn't the right tool for your problem, we'll tell you that too.
Short version

If a job in your business requires someone to read something, listen to something, or interpret something before any rule can be applied, that job could not be automated before. It can now. That covers most of the manual work sitting in a small business: invoices and RFQs arriving in a hundred formats, phone calls, emails, handwritten tickets, records that mean the same thing but are written four different ways.

We build those systems, and we do the unglamorous part properly: choosing the right model for each job, testing it against your real data, and keeping the parts that must be exactly right away from a model entirely. See the case files.

01 / WHAT CHANGED

AI didn't make software better. It made different problems solvable.

The constraint for a small business was never what was technically possible somewhere. It was whether the fix cost less than the problem.

BEFORE

Software could only follow rules you wrote down

If you could specify every case in advance, it could be automated. If the work needed judgment about input that arrived in whatever shape a human felt like sending it, you hired a person. There was no third option, no matter how big your budget was.

WHAT MOVED

Models handle input nobody standardized

A model reads a document it has never seen the layout of, follows a phone conversation that goes sideways, and understands that "3/8 hex cap, gr8" and "Grade 8 hex cap screw 0.375" are the same part. None of that reduces to rules, which is exactly why it stayed manual.

WHERE IT LANDED

The build got cheap enough to be worth it

The same shift cut what it costs to build software at all. Problems that needed a contract team and a year of rollout now take weeks. Plenty of work that was never worth automating at a 12-person company suddenly is.

Whole categories of work crossed that line. Most owners haven't noticed yet.
02 / PREVIOUSLY UNSOLVABLE

Six problems that had no good answer before this.

Every one of these is somewhere in a small business right now, being done by hand because there was never another option.

WHAT THEY HAVE IN COMMON

Every one of them needs judgment before a rule can apply

Somebody has to read it, hear it, or interpret it first. That first step was the wall. Once a machine can do it reliably, the rest of the work was always straightforward automation, and often was already automated everywhere except that first step.

THE WALL MOVED
UNSOLVED_01

Documents nobody standardized

Invoices, RFQs, remittances, lab results, statements. The old approach was OCR plus a hand-built template for each layout, which broke the day a vendor redesigned their form. A model reads a layout it has never seen, the way a new hire would.

IN PRODUCTION: EAGLE MEDICAL BILLING
UNSOLVED_02

Understanding a phone call

For thirty years the state of the art was press 1 for sales. Nobody could build a phone system that actually understood a caller, asked a sensible follow-up, and wrote down what was said. That is now a normal thing to deploy.

IN PRODUCTION: INTEGRATED CLINIX
UNSOLVED_03

Making sense of free text

Emails, notes, complaints, work requests. Keyword rules caught the phrasing you thought of and missed everything else, so somebody read every message and sorted it by hand. Sorting, summarizing, and pulling out the actual request is now reliable.

EMAIL · TICKETS · INTAKE
UNSOLVED_04

Records that mean the same thing

The same customer, part, or vendor written four ways across four systems. Matching them was a genuine research problem, so instead a person reconciled spreadsheets every month and the duplicates never fully went away.

DEDUPLICATION · RECONCILIATION
UNSOLVED_05

Handwriting and photos from the field

Job tickets, meter readings, delivery slips, inspection notes, all photographed on a phone in bad light and then typed in by someone at a desk that evening. That transcription step no longer needs a person.

FIELD CAPTURE · JOB TICKETS
UNSOLVED_06

Deciding what happens next

Routing, prioritizing, and flagging the exception that needs a human. Rules engines could do this only if somebody maintained the rules forever, and nobody ever did. Judgment about what matters is exactly what changed.

ROUTING · TRIAGE · ESCALATION
03 / MODEL SELECTION

We spend our days testing models so you never have to think about them.

New models ship constantly. Which one is right for your job changes every few months, and the difference lands on your bill and in your error rate.

Every time a new model lands, the coverage is about benchmark scores. Benchmarks do not tell you whether it can read your vendor's invoice layout, hold a phone conversation that goes off script, or classify your RFQs without quietly inventing a part number. So we test against real workloads instead: the actual documents, the actual calls, and specifically the edge cases that broke whatever we were using before.

That testing is the unglamorous part of this work and it is most of the value. Model choice is not a decision you make once. A system shipped a year ago may be running on a model that now costs several times more than a newer one that does the job better, and the reverse happens just as often: the cheap model everyone recommends may be the one that fails on your hardest 5% of documents. We re-run our evaluations when something new ships and migrate clients when the numbers actually justify the change, not because there was an announcement.

The practical result is that different jobs in the same system run on different models. A small fast model classifies incoming email for a fraction of a cent. A frontier model handles the reasoning that has to be right. Anything touching data that cannot leave your environment runs locally on hardware you control. You should never have to know which is which, and you should never be paying frontier prices for a job a cheap model does better.

WE TEST FOR What actually matters

  • Accuracy on your documents and your edge cases, not a public benchmark
  • Cost per run at your real volume, projected over a year
  • Latency, because a voice agent that thinks for four seconds has already lost the caller
  • Failure behaviour: does it say it doesn't know, or does it confidently make something up
  • Whether it can run locally when the data can't leave your environment

WE IGNORE What doesn't

  • Leaderboard rankings and benchmark headlines
  • Launch-day claims from whoever built the model
  • Whichever model is currently being talked about most
  • Sticking with a vendor because we already integrated it
✦  The right model for each job, re-checked when the job or the models change.  ✦
04 / LIMITS

The parts we won't hand to a model.

AI is a component, not an architecture. Knowing where not to put it is most of the skill.

LIMIT_01

Anything that has to be exactly right

Totals, tax, margin, payroll, quantities. Ordinary deterministic code does the arithmetic and the validation. A model may read the number off a page, but it never decides what the number is, and every extraction is checked against the records you already have.

LIMIT_02

Data that can't leave your environment

We decide where your data is allowed to travel before we design anything else. When information is too sensitive for an external service, parsing runs locally and nothing is sent out. We do this today for HIPAA-regulated medical billing data, with role-based access, encryption, and audit logging.

LIMIT_03

Decisions with consequences and no human nearby

A model can draft, flag, sort, and prepare. Releasing a payment, sending something legal, or committing to a customer goes to a person. The system's job is to make that person's decision take thirty seconds instead of an hour.

A model that's right 95% of the time is a system that's wrong every day.
05 / IN PRODUCTION

The AI voice agent we built for Integrated Clinix.

A problem with no answer before this: a phone system that actually understands the person calling.

CASE_03 Integrated Clinix logo
AI voice agentCall handlingSpam screening
CASE_03 · INTEGRATED CLINIX
Integrated Clinix website, specialty therapy from the comfort of home
FIG. 03 · AI voice agent

Every billing call answered, including the ones at 9 PM.

The problem

Billing questions all landed in a single voicemail box. The team could not answer every call, and a phone tree was never an option, because the callers were patients with real questions rather than people who knew which extension they needed.

What we built

A voice agent that picks up every time, has an actual conversation, captures the details that matter, and screens out the spam calls that made up a large share of the volume. Callbacks now start with the full picture already written down.

What changed
Every callanswered, day or night, instead of a voicemail box
Spam screenedbefore it reaches anyone on the team
Hundreds of hoursof call handling returned to the staff
Also in production

For Eagle Medical Billing we built local document parsing so claim data never reaches an external AI service, with validation against existing records on every extraction. Same category of problem, opposite privacy constraint. Read the Eagle case file.

06 / QUESTIONS

Straight answers

Q_01

Is this just ChatGPT with your logo on it?

No. A chat window is a product you can already buy for $20 a month, and if that solves your problem we will tell you to go do that. What we build is a system that sits inside your process: it receives the document or the call, does something specific with it, checks the result against your records, writes it to the place your team actually works, and escalates when it isn't sure. The model is one component inside that, usually a small one.

Q_02

Which models do you use?

Whichever one wins on your workload, and usually several within one system. We test the current frontier models from the major providers plus the open models that can run on hardware you control, and we pick per job rather than per client. Classification, extraction, conversation, and reasoning have genuinely different winners, and they change.

Q_03

What happens when a better model comes out?

We re-run our evaluations against your real workload and tell you what we found. If a change is worth making, we make it, and in a well-built system that is a configuration change rather than a rebuild. If the new model is worse for your job, which happens more often than the coverage suggests, we leave it alone and say so.

Q_04

Will it make things up?

A model on its own will, occasionally, which is exactly why the architecture around it matters more than the model. Extractions get validated against records you already have. Numbers that must be right are computed in ordinary code, not read out of a model's answer. Anything the system isn't confident about gets flagged to a person rather than guessed at. We would rather build something that escalates too often at first and gets tuned down than something that quietly gets a claim wrong.

Q_05

Is my data used to train anyone's model?

Not on the configurations we deploy. Business API tiers from the major providers do not train on submitted data by default, and we confirm that per provider rather than assuming it. When the data is sensitive enough that the question shouldn't depend on a vendor's policy at all, we run the processing locally so nothing leaves your environment. That is what we did for HIPAA-regulated claim data.

Q_06

Do I need to understand any of this to work with you?

No, and that's the point of the page. You need to know which parts of your week are eaten by someone reading, typing, or interpreting something. We work out whether a machine can do it reliably, what it costs to run, and whether the savings justify the build. If they don't, we'll tell you that on the first call.

ONE CALL. BRING THE PROBLEM.

Tell us what your team reads, types, or listens to.

One 30-minute call. Describe the work that needs a person to interpret something before anything else can happen, and we'll tell you whether AI can handle it reliably, roughly what it costs to run, and whether it's worth building.

// no retainer required to talk. no junior team on the other end.

Or email us directly: hello@hellospark.ai

Book your free strategy call
Tell us a little about your business. We reply personally within one business day.

30 minutes, free. You leave with a plan either way.

Thanks. Your request is in.

We'll take a look at what you have live today and get back to you personally within one business day.