Last updated 2026-07-10
AAC devices give nonspeaking or minimally verbal children a way to communicate, and ABA therapy can teach a child to use that device with real intent. The two work best together when ABA sessions model the device throughout, rather than holding it back as something to earn. Research consistently shows AAC doesn't get in the way of speech development, and it often helps it along.
AAC stands for augmentative and alternative communication, an umbrella term for anything that helps someone communicate beyond spoken words: picture exchange systems, speech-generating devices, tablet apps with voice output, symbol boards, sign language, all of it counts [1]. ABA, or applied behavior analysis, is a therapy built on the science of learning and behavior, and it's one of the most commonly prescribed therapies for autistic children in the US. It's also often the place where a child first meets an AAC device and first learns to use one.
These two show up together in treatment plans for a simple reason: a lot of children in ABA are nonspeaking or use very little spoken language, so they need a way to communicate, and someone has to teach it. That teaching usually falls to behavior technicians or registered behavior technicians, who often spend more hours per week with a child than anyone else on the team. If those hours don't include steady, natural modeling of the AAC system, a lot of potential practice time simply gets lost.
The history between AAC and ABA isn't entirely clean, though. For years, some ABA programs treated a child's AAC device like a temporary prompt to be phased out rather than a real, lasting voice, or they waited for a child to "show readiness" before introducing one at all. Research doesn't back that up. ASHA has stated plainly that "there is no basis for withholding AAC from individuals who could benefit from it" [1], and the idea of readiness prerequisites for AAC has no real evidence behind it either.
Does using a device stop a child from learning to talk?
No, and this fear is what holds most families back from trying it. A 2008 meta-analysis by Schlosser and Wendt reviewed studies of AAC and speech and found that AAC intervention did not impede speech production, and in many cases came with gains in natural speech [2]. ASHA says the same thing: AAC use does not suppress spoken language [1].
The worry makes sense on its face. If a child can tap a button to get what they want, why bother talking? But the mechanism runs the other way. Communication in any form builds the pathways for intentional expression. A child who learns that pointing at a symbol gets a response is learning that communication works at all, and that's the groundwork spoken words need too.
There's a real caveat worth naming. Some children use speech-generating devices heavily for a few years and go on to develop strong functional speech. Others keep their device as a primary way of communicating for life. Both outcomes are fine. AAC was never meant to be a bridge to speech and only that. Its job is to give a child a reliable voice right now, whatever that ends up looking like later.
How ABA actually uses the device
In a well-run program, the device stays out and in use for the whole session, not wheeled in for certain drills. The behavior technician models language on it throughout the day, not only during structured trials. This is called aided language stimulation, and it mirrors how typically developing children learn to talk: by hearing language used in context over thousands of repetitions before anyone expects them to produce it themselves [3].
A structured session might bring the device into discrete trial training to request preferred items, into natural environment training so the child can comment during play, and into functional communication training to replace challenging behavior with an actual communicative act. That last one has strong evidence behind it. When a child has no reliable way to communicate, problem behavior often fills the gap. Give them a tool that works instantly instead, and that pressure tends to drop [4].
Trouble shows up when the device gets treated as just a prompt, or as a lesser goal than spoken words. You might see an "errorless learning" setup where the device sits next to a preferred item and the child's hand is physically guided to activate it, with no real modeling of language around it. That can teach a child to request. It doesn't build communication that's flexible or carries over beyond that one setup. Watch whether the therapist actually uses the device themselves during the session, talking with it the way you'd talk with your voice. If they're not modeling on it, it's worth asking why.
What the research shows about the two together
Honestly, the two haven't been studied much side by side. Head-to-head comparisons are thin. What we do have is solid evidence for each approach on its own.
On the AAC side, a 2012 systematic review by Ganz et al. looked at speech-generating device interventions with autistic individuals and found positive effects across communication outcomes [5]. On the ABA side, the U.S. Surgeon General's 1999 report on mental health stated that "thirty years of research demonstrated the efficacy of applied behavioral methods in reducing inappropriate behavior and in increasing communication" in autistic children [6].
When it comes to combining the two, the most relevant work involves functional communication training (FCT) within ABA, using AAC as the replacement behavior. A 2019 study in the Journal of Applied Behavior Analysis found that FCT with speech-generating devices produced lasting reductions in challenging behavior and held onto communication gains at follow-up [4]. Nobody has good data comparing AAC-inclusive ABA programs against AAC-absent ones on long-term language outcomes, partly because withholding AAC from a child who needs it is now seen as ethically indefensible.
Where the field lands: bring in AAC as early as possible, treat ABA as a useful way to deliver AAC instruction when it's done well, and keep the speech-language pathologist in charge of designing the communication goals rather than handing that job entirely to the behavior analyst [1].
| Outcome area | AAC-alone evidence | ABA-alone evidence | Combined evidence |
|---|---|---|---|
| Requesting (manding) | Strong [5] | Strong [6] | Strong [4] |
| Reducing challenging behavior | Moderate | Strong (FCT) [4] | Strong (FCT+SGD) [4] |
| Generalization across settings | Moderate | Variable | Variable |
| Natural speech development | No decrease; often increase [2] | Mixed | Limited data |
Who should be running the AAC goals: the SLP or the BCBA?
Ideally both, but the speech-language pathologist should be the one writing the communication goals. This isn't a turf fight, it's about scope of practice. SLPs train specifically in communication disorders, language development, and AAC systems. Board Certified Behavior Analysts train specifically in learning principles and behavior change. The two skill sets complement each other, but they don't substitute for each other.
ASHA's guidance is clear that AAC assessment and intervention planning fall within the SLP's scope of practice [1]. Where a BCBA earns their keep is in systematically teaching a skill, managing behavior that's blocking learning, or building carryover across settings. A BCBA writing language development goals without SLP collaboration is working outside their training, and the same is true in reverse. The BACB's own ethics code sets these scope boundaries for behavior analysts [11].
Plenty of families don't have access to both professionals. If you're stuck with just one provider, an SLP with ABA training, or a BCBA who consults regularly with an SLP, beats either one working alone. Ask your child's team directly who wrote the communication goals and what AAC training they have. That single question tells you a lot.
There's no single best AAC device for ABA therapy, but a few features matter once you're using one inside sessions. It needs to be sturdy, quick to reach for, and available everywhere the child goes, not something that stays locked in the therapy room.
High-tech speech-generating devices, things like Tobii Dynavox, PRC-Saltillo devices, and tablet apps like Proloquo2Go or TouchChat, tend to be the workhorses. They produce voice output, hold large vocabularies built on a real language framework, and let adults model language back to the child. Cost ranges widely: roughly $100 for an app on a tablet the family already owns, up to $6,000 or more for a dedicated device, though insurance and Medicaid funding can sometimes cover the dedicated units [7].
Low-tech systems like PECS (Picture Exchange Communication System) have a long history in ABA and still make sense when technology isn't an option, or when a child is just starting out with communication. PECS has its own published protocol and a solid evidence base, particularly for early requesting [8]. Between those two ends sit mid-tech options like single-message buttons (Big Mack, GoTalk), useful for beginning communicators still working out that their actions cause a response.
For most children in ABA who've moved past early requesting, the aim is a full vocabulary system on a dedicated device or tablet app, one built around a core vocabulary framework so high-frequency words live on the home page instead of buried inside category folders. Words like "want," "stop," "go," "more," "help," and "no" should take two taps or fewer to reach. A device organized entirely around noun categories like food, toys, and people will cap how far a child's language can grow. Our overview of AAC devices covers the systems available in more detail.
Getting insurance or Medicaid to cover the device
This is one of those places where families lose real time and money if nobody warns them. AAC devices are typically funded through medical insurance or Medicaid as durable medical equipment, not through the ABA benefit. ABA therapy runs through a separate behavioral health benefit entirely. The two funding streams don't overlap, and the justification letters go to different departments, which is exactly why families end up confused.
On the Medicaid side, the EPSDT provision requires states to cover medically necessary assistive technology for children under 21, and that includes speech-generating devices [7]. Every state has to cover this, no exceptions. The catch is the documentation: you'll need a communication needs assessment from an SLP and, often, a letter of medical necessity signed by a physician.
Private insurance works differently. The Affordable Care Act requires coverage of habilitative services in most plans, and many states also have their own autism insurance mandates that specifically name speech-generating devices as covered items. All 50 states plus DC have some form of autism insurance mandate, though what's actually covered varies a lot from state to state [9].
ABA therapy itself is usually covered under those same state mandates for autism diagnoses. As of 2021, an analysis found 44 states had mandates that covered ABA [9]. If your child uses an AAC device during ABA sessions, that device gets billed completely separately from the ABA hours. The ABA program can, and should, write goals around using the communication device, but what they bill for is the therapy time, not the device itself.
In practice, start with the SLP for the device evaluation and the funding paperwork, and get the BCBA writing communication objectives that include the device at the same time. Make sure both teams are sharing data with each other along the way.
Red flags to watch for in ABA programs using AAC
Most ABA programs genuinely want to help kids communicate. Still, a few practices are worth questioning if you see them.
One is the device getting put away during sessions. If a therapist removes it or limits access because it's "a distraction" or "we're working on speech right now," that's worth pushing back on. The device is the child's voice, and taking it away during therapy is, in plain terms, asking them to go silent.
Another is no SLP involvement at all. If a BCBA writes the communication goals without consulting a speech-language pathologist, those goals might be behaviorally sound but linguistically thin. Learning to request something isn't the same as developing language.
Watch too for heavy prompting with no modeling. Physically guiding a child's hand to a device button teaches them to respond to prompts, not that communication starts with them. A good sign is a therapist who models spontaneous language on the device during ordinary back-and-forth moments, not just when a request is expected.
If the program only tracks requesting (what ABA calls "manding"), that's another flag. Requesting matters, but real communication also includes commenting, protesting, asking questions, and just talking with someone. A program that only counts mands is raising a narrow communicator.
The most serious flag is punishment for not using the device. Some older ABA protocols used extinction, meaning ignoring the child, to push device use. That can cause real distress, and it doesn't fit with trauma-informed, naturalistic communication practice.
If you notice these patterns, bring them up with the supervising BCBA directly. If nothing changes, it's reasonable to bring in an independent SLP to review the communication goals. And for children whose motor-planning difficulties affect how they use a device, apraxia of speech is worth assessing too.
Supporting AAC at home between sessions
Consistency across settings is everything. A child who uses AAC at therapy but not at home learns that the device belongs at therapy, not to them. You want communication to work everywhere, all the time, not just in the room where the sessions happen.
The single most useful thing you can do is model on the device yourself, without waiting for your child to use it first. Narrate daily life through it: tap "eat" before meals, tap "go" before you leave the house, tap "help" while you're struggling to open a jar in front of them. This is called aided language input, and it teaches your child that the device is a real way to talk, not something that only comes out during therapy time [3].
A reasonable target is modeling 10 to 20 words on the device per day, spread across ordinary moments rather than structured sessions. Bath time, the car, meals, and play all count. Research on aided language input suggests a child's own output typically starts climbing after about 3 to 6 months of steady adult modeling, though this varies a lot from child to child [3].
If you want a way to keep language practice alive between sessions, Little Words is an AI speech companion app built for neurodivergent kids that can run alongside formal therapy. It won't replace an SLP or an ABA program, but it gives families a structured way to fold practice into everyday moments.
One more thing worth saying plainly: keep the device charged, within reach, and in working order. A dead device, or one sitting on a high shelf, isn't a communication system at that point. Treat it the way you'd treat a pair of glasses. It goes everywhere your child goes.
PECS or a speech-generating device?
PECS is a structured protocol developed by Frost and Bondy in the 1990s. Kids exchange physical picture cards with a communication partner, moving through six phases that build from basic requesting up to commenting [8]. It has a solid track record for teaching early requesting to autistic children and shows up constantly in ABA settings.
Speech-generating devices work differently: they produce actual audible voice output when the child activates them. That voice is the real distinction. An SGD speaks to anyone nearby, not just the one person holding the other end of a picture exchange, which matters a lot for a kid trying to communicate independently at school, out in the community, or with peers who've never seen the system before.
Plenty of children start on PECS and move to an SGD as their vocabulary grows and what they want to say gets more complex. Some speech-language pathologists push back on that order and argue for starting with an SGD from day one, so the child never has to relearn communication on a new system later. Honestly, the research doesn't clearly favor one path over the other at the outset. What works best tends to come down to the child's motor skills, cognitive level, and whether the people around them can use the system consistently [5].
There's a practical side too. PECS cards are cheap and hold up to almost anything a child throws at them, sometimes literally. A dedicated SGD can run into the thousands of dollars and take months to get funded. When a child is just starting to communicate, having a low-barrier way in can matter as much as picking the theoretically ideal system. For a fuller picture of how communication develops over time and what tends to slow it down, the early intervention article covers that ground well.
Finding an ABA program that actually supports AAC
Before you enroll anywhere, ask pointed questions. "We support communication" is a slogan, not an answer, so push for specifics. Is there an SLP on staff, or at least a regular SLP consultant writing the communication goals? There should be. Is the child's AAC device present and actually in use during every session, not just pulled out occasionally? It should be. Are behavior technicians trained in aided language modeling? They should be. Ask how staff handle a situation where a child's challenging behavior seems to be a form of communication: a good answer describes functional communication training built around the child's own AAC system, not just behavior reduction. And ask how they share data with your SLP, because you want regular contact, not a once-a-year check-in. You can look up qualified SLPs through ASHA's ProFind directory at asha.org, and BCBAs through the Behavior Analyst Certification Board at bacb.com. Both keep public directories, so this is easy to check yourself. Some children qualify for programs that combine AAC and behavior support under one roof, worth asking about if you'd rather not coordinate two separate teams; these fall under autism spectrum speech therapy services. And if in-person options are thin where you live, online speech therapy has grown a great deal since 2020 and can supply the SLP consultation your in-person ABA team needs, even from a distance.
What good AAC modeling looks like in a session
Video would make this easier to show, but here's what you should be seeing in the room.
The therapist uses the device themselves, all through the session, without stopping to make the child watch or respond. They tap "go" while walking toward a toy, tap "stop" when something ends, tap "want" before holding up a choice, tap "more" partway through an activity. This happens over and over before anyone expects the child to try it themselves. Research suggests most children need to see a word modeled roughly 100 to 200 times before they'll produce it on their own, though that number varies a lot from child to child [3].
The therapist doesn't push the device on the child. Instead they set up small moments of temptation, offering a bite of a favorite snack, pausing a favorite activity, and then they wait. If the child responds at all, a reach, a sound, a glance, the therapist treats that as a real communication attempt and models the matching word on the device right after.
They build on what the child does rather than correcting it. If the child taps "more," the therapist models "more + [item name]" and hands over the item. That's expansion, and it's the same thing parents of typically developing kids do all the time without thinking about it.
If your child's sessions don't look like this, it's worth raising with the supervising BCBA. Plenty of programs will adjust their approach once a parent asks clearly for something different.
Frequently asked questions
Can ABA therapy help a nonspeaking child learn to use an AAC device?
Yes. ABA's structured teaching, particularly functional communication training, works well for teaching a child to initiate communication on an AAC device. What matters is that the program includes consistent aided language modeling from the therapist rather than just prompting button presses, and that an SLP is involved in setting the communication goals.
Will my child stop trying to talk if they have an AAC device?
No, and the research backs this up clearly. A 2008 meta-analysis by Schlosser and Wendt found AAC use did not reduce speech production and was often linked to gains in natural speech. ASHA holds that there's no basis for withholding AAC from children who could benefit from it. Plenty of children use AAC and develop functional speech side by side.
At what age should an AAC device be introduced?
There's no minimum age, and no readiness test to pass first. Children as young as 12 to 18 months have used basic AAC systems successfully, and a child doesn't need to show symbolic understanding or reach a certain cognitive milestone before starting. Earlier tends to be better for communication development overall. ASHA recommends considering AAC whenever speech alone isn't meeting a child's daily communication needs.
Does insurance cover an AAC device for a child in ABA therapy?
The device is usually funded separately from ABA therapy itself, either through Medicaid (which must cover medically necessary assistive technology for children under 21 under EPSDT) or private insurance. You'll need a communication needs assessment from an SLP along with a letter of medical necessity. The ABA program typically doesn't bill for the device, only for the therapy time.
What is the difference between a BCBA and an SLP in AAC treatment?
SLPs are trained specifically in communication disorders, language development, and AAC systems, while BCBAs are trained in learning principles and behavior change. ASHA places AAC assessment and goal-writing within the SLP's scope of practice. The strongest programs use both professionals: the SLP sets the communication goals, the BCBA builds the teaching procedures, and the two share data regularly.
What is aided language stimulation and how does it work?
It means the communication partner models language on the AAC device throughout everyday activities, without expecting the child to respond each time. The adult taps words on the device while talking, so the child gets a visual and auditory model together. Research shows that this, repeated across hundreds of interactions over months, builds spontaneous device use more reliably than drill-based prompting does.
Can a child use both PECS and a speech-generating device?
Yes, and many children do, especially during transition periods. PECS is a structured protocol built around physical picture cards, while SGDs produce voice output. Some kids start with PECS and shift to an SGD as their vocabulary grows. There's no strong evidence that using both at once causes confusion, though it does take consistent follow-through from everyone involved so the child doesn't just default to whichever system is easier at the moment.
What should I do if an ABA therapist takes away my child's AAC device during sessions?
Bring it up directly with the supervising BCBA and ask them to explain the clinical reasoning. If the answer is that the device is "distracting" or that they want to focus on speech instead, that doesn't line up with current evidence or ASHA guidance. The device is the child's communication system and shouldn't be taken away. If the program won't change course, it's worth consulting an independent SLP.
How long does it take for a child to start using an AAC device independently?
There's no set timeline. Some children communicate independently within weeks of consistent modeling; others take six months to a year. Research on aided language input suggests most kids start increasing their device use after three to six months of steady adult modeling, though individual variation is wide. Progress depends on the child, how consistently the program is run, and how much modeling happens across all the settings the child moves through.
Do children with apraxia of speech benefit from AAC in ABA therapy?
Yes. Childhood apraxia of speech affects motor planning for speech, so a child may have the language but not be able to produce it reliably out loud. AAC gives these children a way to communicate while they continue working on motor-speech skills. ABA can support AAC use here, but the motor-speech treatment for apraxia itself needs to come from an SLP, not a behavior analyst.
Is there a core vocabulary approach to AAC and why does it matter?
Core vocabulary is the small set of high-frequency words, things like "want," "go," "stop," "more," "help," "like," "no", that make up roughly 80% of what people say day to day. A system built around core vocabulary puts these words on the home page, one or two taps away. Systems organized only by categories like food or toys tend to limit a child to nouns and requests. ASHA and most AAC specialists recommend core vocabulary as the organizing framework for any full communication system.
How do ABA programs handle challenging behavior related to communication?
Functional communication training (FCT) is the gold standard here. The therapist first figures out what the behavior is communicating (escape, attention, access to something wanted, or a sensory need), then teaches a more efficient way to communicate that same thing, using the AAC device or another mode. Research shows FCT paired with AAC produces behavior reductions that hold up over time, along with lasting communication gains.
Sources
- American Speech-Language-Hearing Association (ASHA), Augmentative and Alternative Communication: ASHA states there is no basis for withholding AAC from individuals who could benefit, and that AAC falls within the SLP's scope of practice
- Schlosser & Wendt (2008), American Journal of Speech-Language Pathology, meta-analysis on AAC and natural speech: Meta-analysis found AAC intervention did not impede speech production and was often associated with increases in natural speech
- Drager et al. (2006), Augmentative and Alternative Communication, aided language modeling research: Aided language stimulation research supporting adult modeling on AAC devices to increase child device use
- Rooker et al. (2019), Journal of Applied Behavior Analysis, FCT with speech-generating devices: FCT with SGDs produced durable reductions in challenging behavior and maintained communication gains at follow-up
- Ganz et al. (2012), Research in Autism Spectrum Disorders, systematic review of SGD interventions: Systematic review found positive effects of SGD interventions on communication outcomes for autistic individuals
- U.S. Department of Health and Human Services, Mental Health: A Report of the Surgeon General (1999): Thirty years of research demonstrated the efficacy of applied behavioral methods in reducing inappropriate behavior and increasing communication in autistic children
- Medicaid.gov, Early and Periodic Screening, Diagnostic, and Treatment (EPSDT): EPSDT requires states to cover medically necessary assistive technology including speech-generating devices for children under 21
- Frost & Bondy, Picture Exchange Communication System (PECS) overview; Pyramid Educational Consultants: PECS is a structured six-phase protocol using physical picture cards developed in the 1990s with published evidence base for early requesting
- Autism Speaks, Autism Insurance Resource Center: All 50 states plus DC have autism insurance mandates; 44 states had mandates covering ABA as of 2021
- American Academy of Pediatrics (AAP), Identifying Infants and Young Children With Developmental Disorders in the Medical Home: AAP guidance on developmental surveillance and early intervention referral supporting early communication supports
- Behavior Analyst Certification Board (BACB), Ethics Code for Behavior Analysts: BACB ethics code defines scope of practice boundaries for BCBAs, relevant to collaboration with SLPs on communication goals